How Would You Refocus Linux Development?

buddyglass writes "The majority of Slashdot readers are no doubt appreciative of Linux in the general sense, but I suspect we all have some application or aspect of the platform that we wish were more stable, performant, feature-rich, etc. So my question is: if you were able to devote a 'significant' number of resources (read: high-quality developers) to a particular app or area of the kernel, and were able to set the focus for those resources (stability, performance, new features, etc.), what application or kernel area would you attempt to improve, and what would aspect you focus on improving?"

Comments

Three things.

[ArcherB]

Better hardware support
Better performance
Maintain excellent reliability.

What else could you need?

Re:Three things.

[__aawkdb2598]

• Visual coherency and a refined GUI. Taste in UI's vary between people, but most linux GUIs that aren't very minimalist tend to suffer from wasted space.
• In interests of making linux more accessible, more configuration utilities that don't require specific knowledge and in-errant editing of configuration text files.

Re:Three things.

[pcnetworx1]

If I could refocus Linux Development... I would try to pool all the development into 1 distro to reduce duplication of so much effort.

Re:Three things.

[The Mad Debugger]

1. Better GNOME usability for Ubuntu (with delivery of Bulletproof X and the GTK Xconfig ASAP, please)

Seriously, the desktop lacks stuff that has been in Windoze since '95. The kernel works pretty good. We have pluggable storage okay.. but there's still basic holes in the usability (like changing the res on the fly when I move my laptop in and out of my office) that just need to get fixed.

2. Spend whatever time is left over to make OOo faster and easier to use.

The MS Office import filters are so *almost* there, but this app really needs to close the usability gap with Office. I have a semi-decent machine running Ubuntu, and even with Java disabled, it still takes what seems like forever to open a simple document that someone emails to me.

I know these aren't *really* linux-specific, especially OOo, but it's what needs to happen to make linux a real, legitimate desktop force. I'm an easy sell, I love open source, but right now there are too many excuses for why this stuff isn't gettin' fixed, and not enough fixin' it, and right now I'm not telling my computer-illiterate friends that they should go order a Dell machine with Ubuntu preloaded.. I'm telling them to buy a Mac, so I don't have to tell them how to fix basic stuff.

Re:Three things.

[ageoffri]

And those 3 reasons are a huge part of why desktop Linux has been on life support for years, way too much effort has been devoted to them. Nerds concentrate on those features and don't get me wrong they are important. What constantly gets left out is usability, installation needs to be simple enough that I can give my parents a CD/DVD and let them install it. There needs to be a consistent UI between applications and components. Installing software must not require editing config files and if additional components are needed then it should just be a click Yes to install additional components.

Re:Three things.

[timeOday]

You know what I'd love more than further improvement in any of those areas? Comprehensive, well-written documentation.

Re:Three things.

[larry bagina]

There not Linux specific (they're not even "Linux"), but you're right. Microsoft or Apple has someone that can make sure the parts all work together to provide a consistent experience. "Linux" doesn't have that (remember a few months back when gnome developers rejected Linus' patches?) Distros package stuff together, but they're primarily cat herders.

Re:Three things.

[ArcherB]

Visual coherency and a refined GUI. Taste in UI's vary between people, but most linux GUIs that aren't very minimalist tend to suffer from wasted space.

Granted, this is important to the Linux community, but when I hear Linux development, I think kernel, modules, and organization (like what goes in /etc, what goes in /bin, what goes in /usr/bin, and so on). Things like KDE, Gnome and other window managers are merely applications as far as I'm concerned and should be considered no more Linux development than, say, Open Office. Of course, I don't mean to say your view is wrong in the least. I just considered the question more narrow than you did and wanted to explain why I didn't consider any X development as part of the question.
Also, since X relies on video hardware, I'd consider X and XGL/Compiz-Fusion/Beryl to be categorized under hardware support.

In interests of making linux more accessible, more configuration utilities that don't require specific knowledge and in-errant editing of configuration text files.

Good point, or better yet, make these files standard across distros so the same configuration utilities works as well on Gentoo as Ubuntu.

A standard method for installing applications across distros would be nice too. I forgot to mention that!

Re:Three things.

[ArcherB]

And those 3 reasons are a huge part of why desktop Linux has been on life support for years, way too much effort has been devoted to them. Nerds concentrate on those features and don't get me wrong they are important. What constantly gets left out is usability, installation needs to be simple enough that I can give my parents a CD/DVD and let them install it. There needs to be a consistent UI between applications and components. Installing software must not require editing config files and if additional components are needed then it should just be a click Yes to install additional components.

I wholeheartedly agree, but I just considered that type of thing more Distro development than Linux dev. On that same train of thought though, I would like to see a easy to use, consistent software installation manager across distros.

Re:Three things.

[omeomi]

when I hear Linux development, I think kernel, modules, and organization (like what goes in /etc, what goes in /bin, what goes in /usr/bin, and so on). Things like KDE, Gnome and other window managers are merely applications as far as I'm concerned and should be considered no more Linux development than, say, Open Office.

How is a window manager less a part of Linux development than basically anything other than the kernel? I mean, just about anything else *could* be considered an "application"--even something as basic as 'ls'--and could potentially be left out of a distro. Like you, I'm not saying you're wrong, but to the Linux-desktop community, things like KDE and Gnome are pretty important.

Re:Three things.

[TW Atwater]

"installation needs to be simple enough that I can give my parents a CD/DVD and let them install it."

Why? Could you give them a naked box and a Vista DVD and have them get it up and running?

Re:Three things.

[Tom9729]

Not everyone uses Gnome/KDE or even X. Everyone who uses Linux, uses a kernel though. I think that was the point he was trying to make.

Re:Three things.

[erroneus]

Having some sort of GUI authority would be pretty cool. I'd like to see something like a single GUI committee that could even go so far as to unify GNOME and KDE. (I know that's asking a lot...)

But I'd also like to see more program and data compatibility. Wine has a LONG way to go and it's an constantly moving and morphing target. But hey, now that MacOSX is on Intel, I'd also like to see some project work on getting Mac stuff to run in Linux too.

There is definitely one drawback to the spirit of the open source community and that would be because the developers are often emotionally involved with their work. Making suggestions or attempting to steer these impassioned developers is like adding salt to a master chef's dish. It'll be tough to work through that. And these same developers are less likely to want to develop based on some model decided upon by a committee. Yes, there are exceptions. Yes there are paid developers who do the bidding of those who pay them too. Frankly, what I'm saying is perhaps we need more of exactly that.

Re:Three things.

[omeomi]

True, but he said Kernel, Modules, and Organization. I'm asking how Gnome/KDE is substantially different from "modules", not "kernel"

Re:Three things.

[YrWrstNtmr]

Why? Could you give them a naked box and a Vista DVD and have them get it up and running?

A) The box they bought already comes with Vista/XP. To change, the process needs to be easy. (I've given a bright 11 year old an XP CD and said 'go for it'. No problem)
B) Why isn't making the change easier an admirable goal?

Re:Three things.

[piojo]

In interests of making linux more accessible, more configuration utilities that don't require specific knowledge and in-errant editing of configuration text files. Good point, or better yet, make these files standard across distros so the same configuration utilities works as well on Gentoo as Ubuntu. No way! A big part of the reason that influences someone to pick a given distro is the configuration style. Slackware and Gentoo are known for how much they teach you about a general Linux system. SuSE is known for YaST, the configuration utility that does everything. Ubuntu is known for being configured very well by default (and it is, compared to some other distros). There is no one-size-fits-all configurator that everyone would like. Windows has a style that seems good enough for everybody, but who actually likes Control Panel? I preferred YaST.

Re:Three things.

[EvanED]

Granted, this is important to the Linux community, but when I hear Linux development, I think kernel, modules, and organization

Then you didn't read the summary very carefully:
if you were able to devote a 'significant' number of resources (read: high-quality developers) to a particular app or area of the kernel

In other words, something that improves KDE, Gnome, X, etc. is a perfectly fine answer to this question.

Re:Three things.

[piojo]

If I could refocus Linux Development... I would try to pool all the development into 1 distro to reduce duplication of so much effort. Linux is about choice. If Linux distros were combined into a one size fits all environment, guess what would happen? It would fork, because there would be people that didn't like it. Combining the distros is an idea that would never get off the ground. The users just wouldn't stand for it.

Re:Three things.

[Chandon Seldon]

Would you do the same for Windows? Would you lump Windows Vista, Windows Server 2003, Windows CE, Windows Cluster Edition, and Windows Mobile all together into one product?

There aren't really more significant Linux distros than that, and they aren't really much more redundant by purpose than that either. Sure, there are a lot of hobbyist distros and LiveCD distros, but those aren't duplicated effort - those are hobbyists playing around.

Re:Three things.

[Blakey Rat]

Apple and Microsoft made customers happy with computers that frankly weren't stable, didn't perform well, but were easy to use. Mac OS had co-operative multitasking, and zero memory protection, and it was proclaimed the best GUI for years. Windows 95 crashed constantly, and yet it sold tons more computers than anything that came before.

Linux is already stable enough, work on other stuff. When it starts to get unstable, then shelve the GUI stuff and go back to the stability.

Re:Three things.

[Yaztromo]

Why? Could you give them a naked box and a Vista DVD and have them get it up and running?

This is exactly the sort of attitude that I would first change about Linux: the idea that Linux needs to replicate Windows, and that it doesn't need to excel past Windows to win the hearts and minds of users.

Linux distros need to strive to be better than Windows. They shouldn't be attempting to duplicate the Windows desktop (something early versions of both KDE and Gnome were, IMO, quite guilty of); they shouldn't be attempting to simply improve upon Windows; they should be attempting to create a unique, best-of-breed solution that users will actually be excited about running (like what Apple is doing with OS X).

So if you can't give the relatives a naked box and a Vista DVD and expect them to get through the install, then we should be aiming to allow them to do this with Linux. The goal of a non-server Linux distro shouldn't be to wind up with the same mundacity that Windows has, or a feature-to-feature parity while ignoring those features Windows lacks; it should attempt simply to be the best operating system it can be -- and if this means forging new ground, then that's what should be done. Linux will never win by trying to be Windows, or by having people say "Well, Windows isn't good at that, so we don't have to be good at that either" -- which is, in effect, the attitude you've just displayed.

Yaz.

Re:Three things.

[Anonymous Coward]

I still have Windows on my laptop for one reason... ...to use a microphone.

Simple things like mic inputs should not be a hassle at this point. Hardware support needs more effort even if it takes putting more pressure on companies to release more hardware info.

Re:Three things.

[timeOday]

I do think most of the different flavors of Windows are bogus. OK, not CE/Mobile - it sucks (at least on my PocketPC), but it is genuinely different from real Windows. The rest are just Windows Ultra Super Cool Edition with various things removed (e.g. remote desktop) and artificial restrictions (e.g. number of simultaneous connections).

Anyways, Windows is nothing like a Linux distro. Windows doesn't come with thousands of apps. Linux itself (the kernel) usually does not differ between kernels in ways a user would notice.

Re:Three things.

[boaworm]

He probably talks about "kernel modules", which are very far from KDE/Gnome stuff. They are runtime loadable parts of the kernel frequently used for device driver kernel code etc. Your kernel does not need to contain code allowing it to handle all NICs (network cards), only the ones you have physically connected at the moment.

Sorry for being a bit too obvious, but this thread already seen enough confusion :)

Re:Three things.

[sohp]

Comprehensive, well-written documentation. ... that would be out-of-date the moment it's published, if not before?

Re:Three things.

[try_anything]

What constantly gets left out is usability, installation needs to be simple enough that I can give my parents a CD/DVD and let them install it.... Installing software must not require editing config files and if additional components are needed then it should just be a click Yes to install additional components.

Yep. I'm a professional developer, been running Linux since the 20th century, and I just utterly failed to install Kubuntu (of all things!) on a box I built about a year ago out of cheap, standard-issue hardware. A number of things didn't work, but most notable were the monitor (I ended up stuck in 800x640 or something like that) and the network device configuration (utterly baffling.)

There needs to be a consistent UI between applications and components.

I think this is overrated. Complaints about UI inconsistency are a symptom of a general (and completely justified) lack of confidence with Linux. People aren't convinced they're savvy enough to handle Linux, so they're easily discouraged. People who aren't confident they can survive and cope with Windows -- there aren't many left anymore, but they exist -- have exactly the same complaints about inconsistencies in Windows. If you take an old geezer who's scared of computers and plop him in front of XP Professional with Office 2007, he'll talk your ear off about the confusing and (to him) arbitrary differences. To him, they're crippling. Of course, he's the exception -- virtually everybody handles the inconsistencies of Windows just fine, and in general they shrug off the disorientation of Windows application upgrades pretty quickly.

The difference between Windows and Linux is that people assume they can scrape by with Windows. By contrast, they know that someday, probably pretty quickly, they will run into something they simply cannot manage to do on Linux: printing, playing a movie, sending a fax, writing a double-layer DVD, who knows what. That knowledge sits in the back of their head, and every time they encounter a difficulty, they're tortured with self-doubt. Is this the stopping point? Is this the point where I'm too dumb or techno-illiterate to go any further? It's exactly the same torture that most non-techy adults had to suffer through in the eighties and nineties.

The way to cure it is to keep polishing the basic functions that challenge even somewhat savvy people today: printing, writing double-layer DVDs, Samba, etc. Grow the user base downward. Even technically minded people like myself have to make cost-benefit decisions about figuring out certain unpolished functionality. Make it usable for us before worrying about "Grandma." My parents and my sister are not going to put up with Linux until they know I can bail them out when they hit a difficulty. I can't even bail myself out sometimes. The last time I tried to set up printing under Linux was 2004. I'm not sure I'm ready to try again.

Installing software must not require editing config files and if additional components are needed then it should just be a click Yes to install additional components.

This is technically a solved problem. Ubuntu kicks ass at this. All that remains is manpower to package the applications, test the packages, and keep the repositories up to date.

Re:Three things.

[Chandon Seldon]

The rest are just Windows Ultra Super Cool Edition with various things removed

I'm pretty sure that Windows Vista and Windows Server 2003 are innately different OS distributions, and that having them on different CDs is actually beneficial to the end user.

Linux itself (the kernel) usually does not differ between kernels in ways a user would notice.

Well, unless it's compiled for a different architecture. But generally the kernel isn't what differs between distributions - usually different distributions are created for different use cases - just like the editions of Windows.

Unless you really think that Mobilinux, MiniMyth, and Ubuntu all need to be installable off the same CD.

Re:Three things.

[bjorniac]

Exactly. I'd like: 1) To be able install some form of linux right off the bat, have it autodetect all my hardware and just WORK. 2) To install programs simply by going to some kind of "Install New Software" select the program (possibly from some ordered list of say, Word Processors, Graphics Packages, Games etc etc) and have it download and install, with a password required for install. 3) Never see the command line unless I want to. I don't want to have to compile things from a text interface. Hell, I don't want to compile things at all. Let me configure things through the GUI, let the techy people configure by editing text files. 4) Some kind of plug-and-play like utility for new hardware. I shouldn't have to mount a USB drive from the command line, it should just be recognised. 5) I know this is a licensing thing, but I'd really like to get standard formats to work (mp3, jpeg, wmv, etc) without having to poke around.

Re:Three things.

[Anonymous Coward]

"when I hear Linux development, I think kernel, modules, and organization (like what goes in /etc, what goes in /bin, what goes in /usr/bin, and so on). Things like KDE, Gnome and other window managers are merely applications as far as I'm concerned"

So they are cat o init. But then, if I could expend fair quantity and quality efforts I'd go for horizontal integration. Linux is quite good on single machines, but lacks a lot when you think about managing the "standard" company on an holistic manner: desktops, laptops, departamental servers, boundary access... of course you have tools that more or less fullfil any given niche, but you've to go through too much pains when you try to integrate them. And since there's no "holistic Best Current Practices" so to say, you have even more problems because all those apps don't accord on how they expect their surrounding environment to be organized and managed making integration even more difficult.

From things as "simple" as single sign-on or proper security models from the kernel up to K3B (since you mentioned KDE) to integrated deployment of new boxes, software, etc. everything works on an "onion model" patch over patch, more dirty hacks than proper integrated solutions.

Re:Three things.

[thePowerOfGrayskull]

What you said:

If I could refocus Linux Development... I would try to pool all the development into 1 distro to reduce duplication of so much effort.

Linux is about choice. If Linux distros were combined into a one size fits all environment, guess what would happen? It would fork, because there would be people that didn't like it. Combining the distros is an idea that would never get off the ground. The users just wouldn't stand for it.

What I heard: blah blah duplicated effort. blah blah choice.

Both sides of this tired old discussion have merit or else it wouldn't still be around after so many years. So please find something that's not a waste of time and energy to debate -- unless either of you honestly thinks you're going to change somebody's mind...

I know, I know, I'm gonna take a flamebait hit for this. I just wonder if people didn't spend all this time having the same pointless circular arguments, maybe there really /could/ be a refocus on common ground. In the words of Somebody Famous, "Can't we all just get along?"

Re:Three things.

[Robocoastie]

Hi Archer, I've read recently people suggest a separate kernel development be done for desktop use which may speed up their multimedia apps, boot times and other things desktop users like. I don't know enough about kernel dev though to know if that suggestion has actual merit or not.

Re:Three things.

[Solra Bizna]

Ubuntu wins on all of those points except 5. (and "proprietary" 3D drivers...)

-:sigma.SB

Re:Three things.

[number11]

Ubuntu wins on all of those points except 5.

Unfortunately, it still has a way to go on #1 (have it autodetect all my hardware and just WORK). I handed out a few Ubuntu CDs.

On the first (a cheap Acer laptop) it complained about the wireless chip. No real options for the user, other than to get a new laptop. But it did work.

On the second, it attempted to make a W2K machine dual-boot, and (going with all the defaults) instead made it no-boot (Win BSOD). The user then attempted to reinstall Windows, which blew away everything. Now, I'll grant that the user made things worse, but what would you expect them to do? And how would they know?

On the third, I attempted to test what had happened on the second, by just grabbing a random used hard drive (containing some version of Windows) and installing Ubuntu on it, accepting all default choices. Ubuntu choked. (Possible causes: multiple partitions, fragmentation, Win swap file in the middle of the area Ubuntu wanted.)

So I reformatted the drive and did a clean W2K install. Ubuntu installed over that very nicely.

So, my impression is that Ubuntu is not really ready for the general public to install, not on a Win machine that's been in use. At the very least, not until it knows how to deal with fragmentation, Win swap files in inconvenient places, and the like. I'd even settle for a message like "Fool, this HD is hopelessly fragmented, fix that and move the swap file, before Ubuntu can help you." Bonus points for instructions on how to do those things. Double bonus points for creating a batch file that will run on reboot and do those things for you.

Don't get me wrong, I think Ubuntu is the best of the lot for the naive user. But the install isn't ready for prime time yet.

Re:Three things.

[timeOday]

I'm pretty sure that Windows Vista and Windows Server 2003 are innately different OS distributions

Of course, because they're different versions from years apart.

I'd say most of (all?) the mainstream Linux distros ARE like all the "types" of windows put together - all the Linux distros can do everything, such as running both desktop apps and server / router functions, limited only by what your hardware can handle. You can select a "desktop" package list, but it doesn't gimp the OS or prevent you from installing servers later.

Re:Three things.

[ozmanjusri]

I would try to pool all the development into 1 distro to reduce duplication of so much effort.

I intend to release my own distro in a couple of months.

How do you plan on stopping me?

Re:Three things.

[piojo]

Looking back, it may have been naive of me to respond to what may have been a troll. It was a pretty good troll, if so, because it just didn't seem like a troll. However, of you, Grayskull, I'm curious, are you also a prick to people that are naive or that bore you, in real life?

Re:Three things.

[Gerzel]

As far as what to add I have to say three things as well

Better Hardware Support, especially in the music and movie hardware field, IE graphics sound and drivers for the HD/BlueRay stuff.

Better Sound, Music, and Movie presentation and support

Better Desktop User Interface

Note none of these are to the exclusion of the classic things such as reliability etc.

The reason I posit these three things in particular is with all the DRM, and Copy Right protection racketeering and general trend of pervasive spyware I feel there may be a niche opening up for users who don't want other people mucking with their systems and want to be able to play the latest tunes and watch the latest movies etc.

A fourth thing is support for mainstream games, through wine or otherwise. A lot of users would switch even if not all games are ported but "the good ones are." Working with game companies to if not develop a native version to make it easier to run under emulation.

Re:Three things.

[vtcodger]

***You know what I'd love more than further improvement in any of those areas? Comprehensive, well-written documentation.***

I'd settle for comprehensible documentation.

I think it is fair to say that documentation is an area where both OSS and Microsoft suck. It's hard to say which is worse.

Re:Three things.

[Daengbo]

I don't want to see KDE, Gnome, and XFCE merged into a single desktop. Having all this diversity is great.

What I would like to see is either a framework similar to wxwindows which allows a developer to write an application that can be run natively in any of those three (or even more) DEs, or a project which allows compilation of one DE's program in another DE, similar to what the Wine developer tools seek to do with Windows apps, allowing them to be easily ported to Unix.

Then I wouldn't need silly themes and widgets which mask the surface differences between DE's applications but which leave glaring usability problems because of differences in architecture. I think there would be more of a level playing field, and Krita could defeat Gimp (reverse if you like) without causing any disruption to someone's DE of choice.

WAAAAY pie in the sky, but the question wasn't about creating a realistic scenario.

Re:Three things.

[Daengbo]

You're not the only one who thinks so.

Re:Three things.

[arivanov]

All of these are not that bad already. Granted they can be better, but it is a case of diminishing returns. To make them better you have to invest a lot for a little return. Also, as Linux is used more and more in embedded devices serving this purpose many of them will improve anyway without throwing effort at it.

Also, IMO Linux actually does the job for an average desktop more or less all right already. I am judging by the number of complaints I get from my significant half which is a Mac user when using any of the machines at home. They have been diminishing towards 0 lately.

Where Linux/OSS sucks royally is business use. There are plenty of areas there which will deliver Linux to a bigger audience with much less investment than any of the UI improvements and support for fancy hardware which linux supports better then windows anyway. The last 3 pieces of really obscure hardware I threw it at it just worked.

• Linux/OSS needs good diagram drawing program support including basic visio import/export (without the executable extensions and VB integration). Dia has significantly improved over the years and it currently approaches Visio as far as features. Some of course are fundamentally different like python integration instead of VB integration, etc. It is still way behind visio on ease of use for the basic UI (when it comes to advanced stuff they both suck bricks through a thin straw sidewise). Also, dia integration into openoffice is inexistent. What made visio the de-facto corporate diagram standard is the integration into office. Dia has none. It does not even have a suitable export which can be imported into openoffice as vector graphics. While at it - openoffice support for internal graphics drawn using impress and/or writer. These leave a lot to be desired as well.
• OSS needs good client-server project manager. There is no need to go for a standalone UI project management software (it is not the route everyone else in the market is is going anyway). Server-client does best for things like this. There is a number of hacked together serverside projects which are desperately asking for a non-Web or Ajax UI. Here OSS can also jump straight ahead of the MSFT in the game if it integrates it to issue tracking and CRM where OSS is way ahead of the game (MSFT still does not have it and is least likely to).
• Contact management integration and contact management in all PIM apps must improve before Joe Average can use it. At the moment most of linux PIM has finally reached some form of useable state and linux itself can sync to most devices like phones and PDAs (if everything else fails Bluetooth usually works). Still, setting it up and getting it going is often outright painful. There are also annoying glitches all over the place.
• Correct timezone processing in calendaring applications must improve. Both Evolution and Sunbird suck at this royally. One forward or import export and the TZ of your calendar gets completely and utterly screwed. Granted MSFT has this f*** up as well so OSS apps are not alone here (dunno when will the developers learn to use universal time for all time storage).

Any one of these will

Re:Three things.

[EvanED]

[Warning: rant ahead. I think I got too carried away with this post... I apologize for the length.]

I think this is overrated. Complaints about UI inconsistency are a symptom of a general (and completely justified) lack of confidence with Linux.

I suspect it's overstated too, but I still think there's something to it.

People aren't convinced they're savvy enough to handle Linux, so they're easily discouraged.

While I suspect this is part of the problem, it's not the whole story.

There's also just an aesthetics issue, which is something that was neglected for a long time on Linux. There have been some recent strides made in this area -- there is now XGL and Beryl and such for some pretty neat window effects, OpenOffice 2.0 was a tremendous improvement over 1.1 in this department -- but there are still some issues, and one of them is the inconsistency between especially Gnome and KDE apps.

For instance, take me. I use Windows as my primary OS. (I seem to be one of the rare breed of people who actually likes Windows.) I have Gentoo installed on this computer too, but it doesn't get much use. Why? I have no good reason to switch, because I don't find the technical arguments in favor of Linux compelling, and I'm not really a subscriber to the Free software mentality. (On the latter point, there's also the "it's not going to affect anything if I stay with Windows" issue that is such a problem with so many issues; I fully admit guilt in that area.)

I mean, what are the major technical arguments in favor of Linux? Security and reliability come to mind as ones that are commonly brought up. Well, I haven't had a problem with either of these on Windows really. I've had a virus once, and that was many years ago, from me doing something stupid (PSA: don't run warez without running it through a virus scanner first -- but I was young and stupid then), and had the whole episode been translated to Linux, would have caused almost as much of a problem then. As for reliability, about the only times I reboot are when I need to turn off the computer anyway, or if I'm installing an update. In the latter case, I can wait 'till I'm going to sleep or something, so the downtime doesn't really affect anything. I recently had to reinstall Windows after the defragger went haywire and corrupted Windows and some other things. But I've also "had" to reinstall Gentoo after Portage started going haywire too. (I could have posted a message on the Gentoo forums or something, but I didn't really want to wait for an answer so I could try out half a dozen things. I did a search for existing problems that were similar, but didn't find anything helpful.) Basically, in short, in my experience, Windows doesn't have a problem in this area. (This isn't to say that someone who is less careful about browsing the internet and stuff wouldn't benefit.)

Linux is often brought up as a great platform for development -- you get GCC, grep, Emacs and Vi, make, etc. But I too don't like this argument, because I don't like those tools. Emacs is a great editor (pretend I said "Vi" if that's your thing) and has some things that are nice for coding, but I'll take a true IDE in a second. I like having better autocompletion ("Intellisense"), a "go to definition" action (tags doesn't cut it), refactoring support, etc. So if I were to work on a personal project where I had a choice, I would use something like Eclipse for development anyway. And hey, that works on Windows too! And on Windows, you get Visual Studio (which I can get for free). And for the things that Linux does better even after that, you can replicate most of it with Cygwin/SFU. So I don't find those arguments compelling.

I have seen people be super-productive with Linux -- using the command line everywhere, set up the window manager to do exactly what they want (I know someone who configured his WM so that windows would display without the title bar or borders so that he didn't waste space on it), etc., but rising to that level would go

Re:Three things.

[WNight]

Why should we strive to beat MS at some ridiculous game someone else thought of?

Either there's a simple installer (no buttons for advanced mode) and it wipes partitions without asking, or there's a complex one that asks too many questions. This is why there are multiple distros - not every one needs to be installed onto a bare metal machine by a newb without instructions.

Besides, Linux has *far* exceeded every MS windows installer. Boot Knoppix and install Debian or RedHat on a partition, while browsing the web, SSHed into your servers, watching a movie, using the included development tools.

Windows on the other hand, if you don't get it pre-installed on the machine requires you to reboot, answer a bunch of questions without help available, makes your machine unusable until finished, requires you to go find the patches, requires multiple reboots or an admin to make a slipstreamed install CD, etc...

Besides, if your parents need a machine they can drool over, buy a Mac, or at least pre-installed Ubuntu. They've turned off the fancy stuff to keep people from hurting themselves.

Re:Three things.

[Proud like a god]

Mandriva/Mandrake's been doing these for years, and if you buy the PowerPack you've got licensed players, drivers, libraries and codecs. Otherwise just go to EasyURPMI and find a PLF (Penguin Liberation Front). mirror.

Re:Three things.

[Zwets]

How do you plan on stopping me?

You'll find out if I ever these frickin' lasers to stay on!

Re:Three things.

[theuedimaster]

You've got good points here, but you're missing something.

Linux development HAS to look at Windows. Sure, it should have a spirit of its own, but in this competitive world that we live in the best products compete and win. I'm sorry, but Windows is the biggest os competitor out there, and Linux development, in order to take advantage of all the momentum and motivation that goes behind competition, must face Windows. That's how the Linux movement can truly succeed - it's pretty clear that having a goal and a competitor can make people work harder and faster.

Re:Three things.

[bateleur]

Emacs is a great editor (pretend I said "Vi" if that's your thing) and has some things that are nice for coding, but I'll take a true IDE in a second.

Not only that, but speaking as someone who loves Emacs, *nix shells and so on... I still run Windows as my desktop OS of choice because Emacs runs fine on Windows, Cygwin deals with my command line needs and (crucially) all the proprietary apps and weird hardware I own just work.

Back when I was a student I ran a dual boot machine and would try running Windows content using Wine and the like. Now? I just don't have the time to wrestle with all the problems that arise. I still run Linux on my server and it's great. But on the desktop I haven't even bothered to install Linux on my newer machine.

Re:Three things.

[b4upoo]

I often fear that Linux might get forked to death. We have talented people that branch out and create an ever ending assortment of distros. I don't have a clue on how to balance that against the Linux communities need to advance as a unit. People have different needs and the Linux community is filling quite a few needs in quite a few directions. And we are seeing some apparent fallout from the process. For example Kanotix and Knoppix have been too quiet for quite some time now and they both helped get Linux over that certain toggle point of popularity. Just maybe the various distros could concentrate on making sure that some sort of run on any platform programs like programs written in Java are sure to run as is without any assistance across the board on Linux platforms.But I fear that run on any platform programs would be aimed mostly at the windows market and as such be all commercial offerings.

Re:Three things.

[Anonymous Coward]

Linux is about choice.

Well, Linux isn't "about" anything, it's a kernel. Having all the choice is a consequence of the evolution history of Linux the kernel and Linux distributions. Linus (thankfully) didn't burden his project with ideology, it's been the FSF who brought that bottle.

Of course choice is good, in a great many if not most situations. (It's perhaps not so terribly desirable when a newbie end-user is just trying to get the damn thing to work and facing a myriad of cryptic options and approaches -- Apple early on understood what choice to eliminate and when... Ubuntu is learning fast too.)

Re:Three things.

[Anonymous Coward]

Better hardware support
Better performance
Maintain excellent reliability.
Every application 'Single Click Install'
Improved GUI

To allow more people to switch:
WINE (run win32)

So my list is I guess like what Lindows once was...

Re:Three things.

[RAMMS EIN]

``Linux distros need to strive to be better than Windows. They shouldn't be attempting to duplicate the Windows desktop (something early versions of both KDE and Gnome were, IMO, quite guilty of); they shouldn't be attempting to simply improve upon Windows; they should be attempting to create a unique, best-of-breed solution that users will actually be excited about running (like what Apple is doing with OS X).''

I think it does that, and I think _that_ is the real strength of Linux (and, actually, other free *nixes). As far as duplicating the commercial vendors goes, of course the open source world will always be behind. This is necessarily true. Still, people will be looking for the same functionality in the same places, and will complain when they don't find it. What they are missing is that the Linux world has so much *more* to offer besides duplicating the functionality of proprietary software in Free software.

Forget about MS Office for a second. Forget about the latest edition of your favorite Windows-only game for a second. What do you really want to do with your computer? Do you want to run these programs, or do you want to obtain the results that these programs allow you to obtain (i.e. getting your work done, or having fun, or both)? If it is the former, your _only_ option is to actually use these exact programs. But I think that isn't actually what you want. What you want is not actually using specific software, but the results of using the software. And this, Linux can give you in so many more ways than Windows can.

Want to write a paper? You can do that in MS Word (even on Linux). Or you could use OpenOffice.org Writer. And that's where most people will stop. But there is so much more! For example, you could write your paper in LaTeX using a text editor. This is a completely different approach, and, if you only know MS Word, it will require a significant effort to master it. However, that doesn't mean it's not a good idea. I use LaTeX, and I always get complimented on how good my papers look. And I didn't have to do anything to make it that way: I just typed the text, LaTeX made it beautiful. And got all the numbering and references right (something MS Word still gets wrong...wtf?) I use LaTeX because it makes my work _easier_.

Want to install software? You could download the source code and compile it. Or you could go to RPMSeek or whatever and search until you find the right package for your distribution, all its dependencies, and install them all. Or you could use a distro like Debian where you just use apt-get (or Synaptic, if you like GUIs) and it will fetch the package and all its dependencies _and_ provide you with upgrades together with the rest of your system. Ever update _all_ the software on your Windows machine with a single click or command?

Wonder why Linux users always seem to use the command line? Perhaps it is because there aren't any GUIs on Linux. Or perhaps there are GUIs, but not for what these users happen to be doing. Or it could just be that they use the command line because they find it convenient. Ever renamed a whole bunch of files? Thrown out old files, but not new ones, from a directory? Looked up the IP address associated with a certain hostname? Explained to someone how to navigate, through several screens, to a certain field where they had to make a change to some setting? Sometimes, the command line actually is quicker, more powerful, or simply easier.

Ever had the same operating system you work with on your desktop run on an embedded device (say, a wireless router)? Ever upgraded from one Windows release to another, updating all your installed applications in the process, with a single command? Ever experimented with different windowing systems, for example, ones that start up in a blink of the eye, that can be controlled solely using the keyboard, or that didn't have any overlapping windows? Ever trimmed down the options a program gave you to just the ones you used, by commenting them out in the source code and recompiling? Ever ad

Re:Three things.

[Anonymous Coward]

How do you plan on stopping me? You'll find out if I ever these frickin' lasers to stay on! It appears you omitted the word "masturbate".

Re:Three things.

[ddrichardson]

"In interests of making linux more accessible, more configuration utilities that don't require specific knowledge and in-errant editing of configuration text files."

This is actually a very good point, one thing inherent too open source development and one I'm guilty of - if I'm writing an application to do x then I know where all the options are and how to set them and it frequently doesn't occur to me that others wont find it so straight forward.

The same can be said for documentation as well, which truly shows how self centric I can be being as I write documentation for others...

Re:Three things.

[David Gerard]

freedesktop.org is unifying what can be unified, and KDE and GNOME work with them enthusiastically. That's why apps from either systray properly in the other and why adding menu items just works, and why installing stuff under Wine puts stuff in the menu (though it frequently can't find the right place). Expect this to only get better.

Re:Three things.

[David Gerard]

The winning aspect of Ubuntu is: if something's too hard for a newbie, that's considered a reportable bug that should be fixed. This attitude makes all the difference.

Re:Three things.

[Lars Clausen]

Linux/OSS needs good diagram drawing program support including basic visio import/export (without the executable extensions and VB integration). Dia has significantly improved over the years and it currently approaches Visio as far as features. Some of course are fundamentally different like python integration instead of VB integration, etc. It is still way behind visio on ease of use for the basic UI (when it comes to advanced stuff they both suck bricks through a thin straw sidewise). Also, dia integration into openoffice is inexistent. What made visio the de-facto corporate diagram standard is the integration into office. Dia has none. It does not even have a suitable export which can be imported into openoffice as vector graphics. While at it - openoffice support for internal graphics drawn using impress and/or writer. These leave a lot to be desired as well.

As the current (fourth) maintainer of Dia, I'm all ears about what could be improved, but alas permanently short on time. I wish SVG was a better output format, but I have yet to see two renderers render text the same way. Writing an exporter for Dia is not hard, but real integration is a different beast. I had hopes that Bonobo would be the glue that could get diagrams to "live" in other documents, but it seems to have died with no replacement. As for ease of use of basic UI, there are several points we are working on as quickly as time with a full-time job allows.

-Lars

Re:Three things.

[jd]

Maintain excellent reliability

I would argue that we need far better reliability. Not just in the kernel, but in the apps as well. Yes, Linux has one of the lowest bugs per thousand lines of code of any popular OS, by Coverity's metrics, but it is still higher than comfort should allow. Some of the recent 2.6.x kernels have not been too impressive, Firefox is getting sloppy and if KDE's QA had been more thorough, they'd be in a position to put more developers into the 4.x cycle rather than maintenance releases of 3.x.

This is not to say that new stuff shouldn't be integrated into the kernel or Firefox. Quite the opposite. I think it's time to be aggressive and add far far more. Provided additions are truly modular, bloat doesn't occur. Bloat only happens when you are compelled to include something. However, I feel that there also needs to be far better QA on everything. This is the area geeks tend to fall down on. Testing isn't exciting. In fact, it's mind-numbing at times. And once you get to the real, real low-level hard testing, it is painfully slow. Running a syscall or loading a web page is trivial, if painful after the first few thousand times, but that's nothing compared to validating the logic of code where potential side-effects can arise, or devising tests against slowly accumulating errors that may only arise on certain permutations of arc through the code.

Logic (and the eventual heat-death of the Universe) dictate that you cannot get code bug-free by testing alone, which means that thorough QA must also include mathematical techniques. And precisely how many of the hundreds of thousands of part- and full-time kernel hackers have mathematical proofs in their background? Of those, how many have the time, patience, skill, money, resources or sanity to actually apply those techniques to eliminate bugs where case-by-case testing won't work? Mathematical techniques can easily take two to three days for every hundred lines of code, if you're working on it full-time. It is the ONLY way some bugs will be found and fixed, but you'd need close to a million mathematical experts full-time for a year to just find the intra-function bugs in the kernel as it stands. If you wanted to test X, KDE, Gnome and a few standard applications as well, you'd best be prepared to convert a small nation into a gigantic QA factory.

For those arguing that the desktop applications are important, I agree. However, and this is important, nothing can be better than the foundations you build on. If the foundations are weak, what you build will be weaker. That is the lesson we should learn from a certain other OS. We need the foundations in Linux to be far stronger than they already are. The GUI is especially in need of a revamp, but there is no layer in Linux that I would consider to be sufficiently strong that it could endure the kind of abuse it WILL get when it has 99% of the users on it.

Re:Three things.

[michrech]

Actually, that's *exactly* what Linux could use for "the desktop". Having to recompile your kernel every time you change a tiny piece of hardware sucks. Yes, on these fast and modern machines, that isn't a *huge* issue, but do you really thing the average user is going to (want to) do that?

At the least, parts, that can be, should be set as M, so they can get loaded as they are detected during boot. Just have something that does a scan on first boot, stores the hardware it finds in some location, then each boot after that it can scan for changes (shouldn't take long -- it doesn't in Windows land...) and compare the findings to the already stored config from first boot.

He probably talks about "kernel modules", which are very far from KDE/Gnome stuff. They are runtime loadable parts of the kernel frequently used for device driver kernel code etc. Your kernel does not need to contain code allowing it to handle all NICs (network cards), only the ones you have physically connected at the moment.

Re:Three things.

[Xabraxas]

Actually, that's *exactly* what Linux could use for "the desktop". Having to recompile your kernel every time you change a tiny piece of hardware sucks. Yes, on these fast and modern machines, that isn't a *huge* issue, but do you really thing the average user is going to (want to) do that?

That's not true at all though. Most distros ship with kernels that have modules built for almost all hardware and the module simply loads when a piece of hardware is present. If you have a custom compiled kernel like I do all you have to do is build the module. You don't need to recompile the whole kernel and haven't had to do things that way for many years now.

At the least, parts, that can be, should be set as M, so they can get loaded as they are detected during boot. Just have something that does a scan on first boot, stores the hardware it finds in some location, then each boot after that it can scan for changes (shouldn't take long -- it doesn't in Windows land...) and compare the findings to the already stored config from first boot.

Hal and udev work quite well for me recognizing hardware and then loading support for it.

Re:Three things.

[NickCatal]

This brings up a good point: ONE, Standard, Package, Distribution, System ..

Re:Three things.

[xophos]

That sounds like you never tried any of the mor popular Linux distributions. SuSE, Fedora, Mepis, PCLinuxOS, Ubuntu, With none of those the average User does ever need to compile their own kernel. Since i use Ubuntu exclusively i never had to do anything at all to configure my Laptop correctly other than clicking "Yes i want to use the proprietary graphics driver" since that was my only option for 3d acceleration. Even - and that's what impressed me the most - suspend to ram works flawless.
None of the above can be said about my installation of Windows XP (which is for gaming exclusively since i never bothered to go hunting for all the Applications needed to make it a remotely useable working environment).

Re:Three things.

[lordtoran]

There is this thing called manpages, and the Gentoo wiki (a very exhaustive source of general Linux knowledge, not only Gentoo).

Re:Three things.

[jon_anderson_ca]

Better hardware support

Specifically, open-source drivers for everything.

I've been super-frustrated over the past few months with ATI's FGLRX driver (I used to be able to suspend and resume, but not any more), and my card (X1400) doesn't work with the open-source Radeon driver.

One major success story: the Avivo driver for X1xxx cards. The newest driver in Ubuntu (and Debian) is only a few months old, but its 2D performance is already better than FGLRX! Once they get Suspend/Resume and some rudimentary 3D (just enough to run Compiz), I will be a VHP (Very Happy Penguin).

Re:Three things.

[Xabraxas]

Seriously, the desktop lacks stuff that has been in Windoze since '95. The kernel works pretty good. We have pluggable storage okay.. but there's still basic holes in the usability (like changing the res on the fly when I move my laptop in and out of my office) that just need to get fixed.

I am using the the latest Intel display driver on my laptop and it fixes a lot of old problems with the driver. No more 915resolution, no more mode lines in xorg.conf, automatic detection of external monitors, etc. So what exactly does Linux lack that Windows 95 had? MS BOB?

The MS Office import filters are so *almost* there, but this app really needs to close the usability gap with Office. I have a semi-decent machine running Ubuntu, and even with Java disabled, it still takes what seems like forever to open a simple document that someone emails to me.

On a modern machine with an up-to-date version of java OpenOffice opens quick enough for me. About two seconds when cached, maybe 5-7 seconds on a cold boot. As for the filters I guess it all depends on what you need to view. I am not opening many very complicated Office documents so I never have problems but I wouldn't doubt that there are still issues with complex documents.

Re:Three things.

[Xabraxas]

There not Linux specific (they're not even "Linux"), but you're right. Microsoft or Apple has someone that can make sure the parts all work together to provide a consistent experience. "Linux" doesn't have that (remember a few months back when gnome developers rejected Linus' patches?) Distros package stuff together, but they're primarily cat herders.

That's not what happened. Linus' bug fixes were accepted but GNOME was already in feature freeze at the time so no new features went in but some of his feature patches are kicking around in 2.19.x.

Re:Three things.

[squiggleslash]

I can tell you the last thing I want to happen is for the kernel developers to start working on GUIs. Not only is this getting the wrong people to do the job, but it leads to the frightening possibility of the GUI actually being more integrated with the kernel.

That's not to say there aren't GUI-focussed processes that the kernel developers shouldn't be looking at, but you're asking them to work on making front ends that are minimalist and make the system more accessible, all of which are high level issues, whereas the work the kernel developers should be focusing on concerns performance, responsiveness, making good use of the hardware, and other lower level aspects that the kernel can directly help with.

Leave the look and feel to the experts on look and feel. The kernel is not the right place to be discussing look and feel.

Re:Three things.

[xophos]

Well usually the best Marketing wins. But you are right. To displace an established Market Leader you have to beat them in every Aspect.

Re:Three things.

[siride]

Hmm, this is very insightful. I hope somebody figures this out. Oh wait...Linux distros have been doing this for years. How long have you been using Gentoo or LFS? If you had touched a modern Linux distro, you'd find that this is exactly what they do and have done for a number of years.

Re:Three things.

[SiChemist]

Have you actually used a Linux distribution in the last 5 years? In modern distributions, the kernel detects hardware and loads drivers during the boot up process. That means that you can take a hard drive out of one machine and put it in a non-identical machine and it will boot and use all the hardware (as long as that hardware is supported by Linux). I think that maybe I've been trolled.

As for what I would improve about Linux, there isn't anything that I think can be directly improved by the developers. I would like to see more commercial software support, but that's really an increased adoption issue.

Re:Three things.

[l3mr]

Gentoo is what? Here I thought that Gentoo was known for letting you mindlessly recompile everything and giving non-developers the chance to feel 1337 by watching hours of those very informative and educative gcc compile messages ...

Re:Three things.

[LinuxDon]

The installation of Linux is not the issue.
Where distros fail (Both SUSE and Ubuntu) is the following:
- GNOME is the default, and is NOT user friendly and looks different in both distro's
- Mplayer and ogle are not installed on default, so you can't just throw in a DVD and start playing
- Mplayerplugin is not installed on default

The average user (not me) just wants to do the following:
- use a wireless card
- use a printer
- write a document
- use MSN
- use a photoshop kind of application
- use bittorent
- use the s-video port on a laptop
- burn some CD's
- view movies
- listen to music
- not being bothered with inserting a CD/DVD etc. when something needs to be installed

All the apps and possibilities are there, but still it doesnt "just work".
SUSE requires other installation sources to be configured to install mplayer, but then ultimately the auto-update just breaks. Also you'll have to actively search for those other installation sources.

Regarding hardware: They should just make an HCL and make sure it works out-of-the-box flawlessly. I will happily purchase more expensive hardware if I can be sure it works flawlessly.

And..
when will ALSA just fix the 'sound card in use' issues, so it always works on default?
when will they make it possible to make 'universal install packages' so an app can just be installed on ANY Linux distro?

I know there are legal issues surrounding the codecs, but a solution/workaround should be found to make it possible anyway. Last time I tried to view a DVD under SUSE, it took me a lot of time to get the apps installed, even though I have 10 years of Linux experience.
IMHO, Gentoo does a much better job, but compiling from source takes too much time if you just want to install it on an old laptop.

IMHO, if they'd just fix the issues mentioned above then Linux would do MUCH BETTER on desktop then it does now.
If enough money would be thrown into fixing the issues, it all could have been resolved within a year.

BTW, I love Linux and use it on my desktop at home and all servers at work.

Re:Three things.

[Anonymous Coward]

> Comprehensive, well-written documentation.

I'm looking at your comments solely from an end-user (i.e., non-developer) point of view where I would more or less equate "documentation" to an application's built-in "Help" system.

My general response is: Amen, brother! :)

My specific response to the excerpt above is: Why limit ourselves to "written"? Couldn't "documentation" also take other forms (e.g., multimedia-based tutorials)?

If the answer is "Yes", then to what extent is written end-user "documentation" just a poor substitute for end-user "training". Or, is written documentation simply a sub-category under training?

In my experience, many users will hesitate/refuse to use an application's "Help" system but will gladly accept having someone more knowledgeable and experienced sit down with them and patiently help them learn (show them and explain to them) how to use the application.

The bottom line is: Perhaps we need to seriously rethink how end-users learn to use applications and how they can quickly and easily obtain helpful answers to their application-related questions.

Re:Three things.

[Rakshasa Taisab]

With a squad of ninjas.

Re:Three things.

[Provocateur]

bright 11 year old an XP CD and said 'go for it'

The PC option, at least with this crowd, would be to also give him an Ubuntu CD, whilst saying, BTW take off the desktop background on this one, it's got naked chix on it.

'Excuse me, what was that?'

'Nothing. Hey look, is that a roomba vacuum cleaner? Make it go in the closet!'

Be sure to tune in again next week when we visit the 11 year old in his basement to see which OS led him to the downward spiral.

Re:Three things.

[tr1907]

"On the third, I attempted to test what had happened on the second, by just grabbing a random used hard drive (containing some version of Windows) and installing Ubuntu on it, accepting all default choices. Ubuntu choked. (Possible causes: multiple partitions, fragmentation, Win swap file in the middle of the area Ubuntu wanted.)" On the other hand Windows Installation has no effort to determine any non-MS Operating Systems on the machine. No detection of partitions other than FAT,NTFS. Replaces the MBR after the installation. That makes Windows "not really ready for the general public to install, not on a NON-Win machine that's been in use".

Re:Three things.

[Pollardito]

you don't think Debian, Xandros, SUSE, Redhat, Linspire, Gentoo, and Ubuntu are more redundant on purpose than the 5 Windows versions listed? it's not like any of these is a great deal more optimized than the other for mobile, server, or cluster use (though at least SUSE and Redhat do have versions for server, i'm just counting their regular distribution which is completely redundant with several others here).

Re:Three things.

[Provocateur]

IIRC Mint is an Ubuntu derived distro that can do #5 quite well. Of course if it detects the HW/audio/sound properly that is...

Re:Three things.

[doshell]

On the second, it attempted to make a W2K machine dual-boot, and (going with all the defaults) instead made it no-boot (Win BSOD). The user then attempted to reinstall Windows, which blew away everything. Now, I'll grant that the user made things worse, but what would you expect them to do? And how would they know?

They would know if Microsoft played nice and displayed a big fucking warning during the Windows installation procedure, "Your MBR will be replaced, denying boot menu access to any other operating system you might have installed."

Or if Microsoft software actually gave the user a choice, instead of assuming all they'll ever want to run on their machine is Windows.

Re:Three things.

[Jack9]

Linux is about choice. ...the users just wouldn't stand for it.

You meant to say SOME users wouldn't stand for it. Specifically the user who have uncompromising egos or specific needs, while the rest of humanity might actually CHOOSE Linux for day-to-day use. People are mostly the same in what they want. I'll give you time to count how many companies are wildly successful, based on this premise.

Re:Three things.

[doshell]

The problem isn't just that there isn't enough documentation. It's also that the average user doesn't care about it (i.e. expects everything to work by clicking their way through menus and dialogs, and blames the system when their data is eaten up or similar).

Re:Three things.

[number11]

Windows "not really ready for the general public to install, not on a NON-Win machine that's been in use".

True. If Linux was the dominant desktop OS and MS was trying to convert users over to Win, users who wanted to add dual-boot would experience a lot of problems and be unhappy. In that alternate universe, we'd all drive flying cars, "Katrina" would be famous only as the name of a Bollywood actress, and George Bush would be just another rich guy with a losing sports team.

Re:Three things.

[mazarin5]

In the words of Somebody Famous, "Can't we all just get along?"

That's Rodney King, baby.

Re:Three things.

[nine-times]

I think you're right in the sense that Linux is actually pretty good already, even on the Desktop. For most purposes I deal with in my job, the the big issue would be to get pixel-perfect clones/ports of programs like Soundforge and Photoshop. Admittedly, I've been working for media companies for a while, so my vision is a little skewed. There are media programs that do a lot of the same things as Windows/OSX counterparts, but they really aren't usually as refined and mature. I'm sorry, I know people will jump down my throat on that one, but it's been my experience.

I think there are an awful lot of things that could continue to be refined, matured, fixed, or whatever you want to say. There's still work to do, but I don't think there are gaping holes in Linux anymore.

If you had access to good-quality programmers and really wanted Linux to excel and take more market-share, my advice would be the following: Involve a variety of usability experts and experienced users from a variety of fields in your process. In my opinion, this is the area where lots of developers fail (not just FOSS developers, but Microsoft seems pretty guilty of this). Sometimes programmers just don't have the right perspective on things because they aren't really power-users of their own software.

As someone who's done some graphic design, I'll tell you that I don't think Photoshop/Illustrator are perfectly intuitive or sensible. As someone who's been a network admin, I don't think anyone has yet solved issues with data-storage involving backups and archives. Imaging isn't as easy as it should be. E-mail/Calendar/Contact management is still pretty retarded. As a user thinking about usability, even OSX has its problems.

When I've talked to my peers in each of these fields, we tend to have the same problems, and though there's no shortage of developers working on the related software, it sometimes seems like developers just don't know what the problems are.

So I guess I'm saying that Linux developers (and Windows developers and OSX developers) should really be communicating with their users better about what problems actually need to be solved.

Re:Three things.

[pivo]

I've used Linux on laptops for many years, but recently I haven't been able to get current Linux distributions (Fedora or Ubuntu) to work correctly on my IBM T40. By work correctly, I mean suspend/resume and Wifi with WPA. After weeks of trying to get these things working I gave up and bought a MacBook Pro (though I still use Linux on non laptops.) Now I have the added benefit of being able to attach an external monitor and have it work immediately, without restarting X or whatever (which as far as I know is a feature not supported by any distribution of Linux.)

I'd still rather use Linux on my laptop, for openness reasons and because OSX isn't great for working in Unix shells (e.g. no decent terminal), but the general ease of use of OSX has spoiled me and I don't expect that to change in the near future.

Re:Three things.

[m2943]

In addition to Microsoft's variations, there are the dozens of wildly different and incompatible varieties of Windows that the PC vendors create: different themes, different desktops, different applications, different administrative tools, different startup sequences, different drivers, etc.

Re:Three things.

[Anonymous Coward]

The Tango! Project by Freedesktop.org should solve it, but I agree with you, visual coherence is important. I keep seeing KDE apps with weird icon sets, they don't look anything like the GTK+ ones which have been set to use Tango, or any other apps being run.

Re:Three things.

[bipolarpinguino]

Next version of X.Org (7.9) is supposed to have hotplug support for monitors, release date next wednesday. I'm hoping itll work, i just got a new 20" monitor that I haven't bothered trying to set up in linux yet since I'm hoping xorg'll do it automatically. I'm just using OSX until then.

Re:Three things.

[Anonymous Coward]

At first "Usability" popped into my mind but then it hit me, again. Some fucking documentation would be nice. Its not an afterthought people, Microsoft products are documented to no end, Linux is a grope in the dark. I have no problem using a command line for everything if I have docs. The docs that I find are ALWAYS outdated and skimpy as hell. Just the facts would be nice. The nitty gritty bare bones workings of the damn system would thrill me to death. I call this "the plague". For instance when I read docs it refers me to a path and low and behold some asshole decided to put everything under a completely different directory structure! EVERYTHING needs to stay in the same directory for every version and distro. This is the main factor in preventing me from using any linux distro, I just got sick of having to relearn everything. We are creatures of habit!!!! Anyore I don't give a fuck about any OS, whatever it is it just has to work and be intuitive. I just want to concentrate on developing applications and have them work everywhere. I don't give a crap about Windows, OS X, Linux, Unix; if it doesn't facilitate a smooth workflow I don't want anything to do with it. The average dumb user is the MOST important thing to consider. I go through great pains to develope apps for these idiots the least we could ask for is an OS based on the same priciples. Hacking kernels is EASY that is why people do it, designing a product an idiot could use takes talent, focus and hard work. This includes docs even if you think no one reads them, they do!

Re:Three things.

[thePowerOfGrayskull]

Looking back, it may have been naive of me to respond to what may have been a troll. It was a pretty good troll, if so, because it just didn't seem like a troll. However, of you, Grayskull, I'm curious, are you also a prick to people that are naive or that bore you, in real life? Honesty often has the effect of making people think you're a prick. I only tend to be that blunt about it when I reach a tipping point of frustrating from reading the tired arguments all the time.

Re:Three things.

[The Anarchist Avenge]

Mod parent funny

Re:Three things.

[DarkMantle]

In interests of making linux more accessible, more configuration utilities that don't require specific knowledge and in-errant editing of configuration text files.

I couldn't agree more. I hate having the find out how to edit a damn text file to share something on SMB (samba). But there's more it needs.

• Standard install method so all linux packages don't need to be repacked. .deb .rpm etc... pick one please
• Pick a WM so all apps work/look the same on all distros.
• "By developers, for developers" must stop. Even Ubuntu isn't ready for the average Joe.
• Focus on usability to make it usable by Joe.
• Hardware support. It still doesn't work with my Intel DG965WH and that's an old board.
• WINE, or Crossover office integrated for compatabilty for windows apps.

If you can get my motherboard and games running on Linux, I might finally switch.

Re:Three things.

[Annymouse Cowherd]

A standard method for installing applications across distros would be nice too. I forgot to mention that! ./configure && make && sudo make install

Re:Three things.

[DarkMantle]

Debian, Xandros, SUSE, Redhat, Linspire, Gentoo, and Ubuntu

You forgot Mandriva, PCLinuxOS, MEPIS, and Slackware. As listed on DistroWatch as other "Major" distros.

Re:Three things.

[watchingeyes]

Have you tried downloading iTerm, or any of the other terminals available for OS X? All of the terminals available in Linux are also available on OS X through fink.

Re:Three things.

[Verity_Crux]

Better hardware support

It's frustrating to me that we still lack fakeraid support in the installers, in the kernel, and in the boot loaders. What gives? I know that fakeraid is considered a crappy hack, but that's no excuse to upset the numerous people using it.

In an only-semi-related comment, I wish Adaptec would provide Linux drivers for all their raid controllers. And I wish Linux distros made it easy to use from the installer.

Re:Three things.

[watchingeyes]

Could we not keep the "style" of different distros the same without having basic text files that control settings be completely different between distros?

Re:Three things.

[pivo]

Thanks, but yeah, I think I've tried them all. The only one that I can get to work like a normal Unix terminal WRT key bindings and emacs, is xterm from Apple's X server package (or rxvt I think.). Even then, it took some tweaking to have Backspace/Delete and the Apple key-as-Meta work properly locally and on remote Linux hosts.

Re:Three things.

[Goaway]

Downward spiral into what, exactly? From all your hints, I was thinking bisexuality, but I'd like to be sure.

Re:Three things.

[Enderandrew]

Portage, portage and portage. Or maybe a second-generation portage-replacement. Pkgcore and Pauldis are awesome.

USE flags are by far the greatest strength of Gentoo, and the packages within a portage tree can be source, .deb, .rpm or whatever. It doesn't matter. Seriously, the entire Linux world should consider a portage-style universal method of package delivery.

Re:Three things.

[Anonymous Coward]

Better Applications.

Re:Three things.

[man_of_mr_e]

What else do you need?

Tons.

The two most successful (in terms of end-user acceptance, not market share in particular) are OSX and Windows. Both of these platforms share at least some common traits, and since both of those companies have invested millions, if not billions of dollars into usability studies, one should try to learn from what they do right.

This is not to say everything they do is correct, but areas that are common between the two suggest that two independant entities thought it was a good idea, and thus should be giving serious consideration.

Areas that are similar between them:

Common clipboard formats
Consistent user interface (whether or not you agree with that particular interface)
Unified installation and packaging systems.
Consistent configuration systems
* The appearance of monolithic applications (explained below)
Fast UI

* Now, what do I mean by the appearance of monolithic applications? By this, I mean that each application is an island, even though they have interoperability between them. Each app appears to stand alone, even if part of a "suite". Each app, even if built upon other components, doesn't expose those components to the user. It's just the app.

Linux gives the appearance of being piecemeal. Of being a disparte group of objects, applications (even within an application), and systems. Even apps in the same group (kde, gnome, etc..) often have significantly different UI's, configuration engines, and general feel. Skins aren't enough to change that.

I know many people like to prattle on about choice, but you can still have choice and consistency. Some will argue that Apple and Microsoft aren't all that consistent either, but there's a difference. When apple and ms make a change, they move forward with that change so that new apps start to conform to that change. In linux, people just do things the way they want, and screw what anyone else thinks.

It's a tough nut to crack, and, unfortunately, probably not possible in a community development model simply because there is nobody with any authority to set standards. Even the LSB is often disregarded, and the LSB doesn't go anywhere near far enough because it has to appease dozens of different distro makers.

Simply put, what needs to be done, can't be done without destroying what makes Linux so great to so many people. As such, there will always be too much resistance to do what is necessary to make Linux compete with Apple and Microsoft (or any other proprietary vendor, such as Adobe). You'll get bits and pieces, but it will never be enough...

Re:Three things.

[perlchild]

No, but they'd accept it, and try, and ask their windows-saavy friends to do it for them...

The fact that windows is more widespread than linux means that to be on parity with windows, linux cannot just be as good, it has to be significantly, perceivably better... That's why MacOS X(despite the hardware barrier) is on the rise, it's perceived as better, as long as you can afford it.

And I know, it's one of those comparisons where I can't give them a Mac OS X cd and let them get it up and running either... But when they buy a new pc, a Mac is an easier sell than linux... Despite the fact that Windows is more expensive, it's also an easier sell than linux.

Something to keep in mind when you want to expand linux's market... I won't comment on any "why"s.

Re:Three things.

[watchingeyes]

I don't understand. The ones available through Fink (Debian's apt for Mac OS X) are merely the Linux/Unix terminals recompiled on OS X. I don't think they've been modified at all. They even run through X11, not Aqua. Which terminal do you normally use on Linux?

Re:Three things.

[turgid]

You know what I'd love more than further improvement in any of those areas? Comprehensive, well-written documentation.

It's called source code.

Re:Three things.

[turgid]

How do you plan on stopping me?

Easy - public humiliation. You are an anit-American, dirty, smelly hippy communist. You can't even stand up for your subversive beliefs by sticking with your unwashed comrades.

Now, eat the biscuit, install Windows and forget about it. You will be forgiven.

Re:Three things.

[try_anything]

Linux is often brought up as a great platform for development -- you get GCC, grep, Emacs and Vi, make, etc. But I too don't like this argument, because I don't like those tools. Emacs is a great editor (pretend I said "Vi" if that's your thing) and has some things that are nice for coding, but I'll take a true IDE in a second.

Well, I'm halfway on this point, since the two applications I spend the most time in (besides my browser) are Eclipse and Emacs. No IDE that I know of has a decent text editor and no text editor that I know of has decent IDE tools. It's a real problem that only exists for historical reasons. Personally, I prefer Emacs for everything except Java. C++ is evidently too grody a language to have decent tool support, and the dynamic languages of course are just text anyway :-)

So basically this comes down to the following: I don't have a good reason to switch.

Hmmm, if I told you you were a pawn in a global OS popularity contest, would that sway you? That's how I feel about much of the Linux advocacy stuff: A bunch of pimply-faced kids trying to force their friends and family to use Linux because it's important for their ego.

On the other hand, Microsoft is a corporation; their interests are narrowly focused on making money. Unix is a world that programmers built for themselves and each other. Unix assumes that every user is a guru in training, which is a wonderful thing, even for those who will never get very far. In eight years of using Linux, I've dealt with about a dozen different kinds of config files. Most of the time, I never have to even know they exist, but when the GUI facade proves insufficient, the mechanism behind it usually turns out to be clean, presentable, and acceptably documented in various places on the web. I hate working with config files -- in almost every case, I consider it a major UI failure when I have to edit a config file by hand -- but it is not as bad as having no fallback, as in Windows. When you work with any kind of Linux config files or command-line tools, you're dealing with a user interface that is designed to be as simple and easy as possible for the people who have to use it.

As far as Windows is concerned, if the GUI facade doesn't work, then game over, go home, wait for Microsoft to fix it. The stuff behind the facade in a Microsoft product is *supposed* to be shit, because it isn't part of the supported product interface, and Microsoft doesn't waste resources on something that isn't part of the supported interface. In fact, it actually helps Microsoft if the stuff behind the facade is inaccessible and/or undocumented, becaues it discourages users from going beyond the single official interface. I had to debug a problem between a C# SOAP client and gSoap server at work a few months ago, which meant running Visual Studio and using its "magic" web services code generation. It was torture, because there was exactly one right path for using Visual Studio, and it meant handing over all control to Visual Studio's automated tools. Every attempt I made to take control of one part of the process or another, or simply delve into the workings and understand them, was utterly thwarted. This would be fine if everything worked, but something was wrong, and the maddening opaqueness that Microsoft *intentionally* programmed into Visual Studio turned what should have been merely a laborious and boring debugging task into an epic struggle against the "Microsoft Way." Documentation on how Visual Studio worked internally was scarce, even using all my google-fu. All I managed to do was find and read the C# code generated by Visual Studio. Eventually I tracked the problem down to a known bug in Visual Studio 2005. Will not fix, wait for Orcas (currently scheduled to ship in February as Visual Studio 2008.)

Why in the world did Microsoft make this ordeal so unpleasant? Well, they're a for-profit corporation. That really covers it. Everything

Re:Three things.

[Bluesman]

I think Microsoft's documentation is excellent, but the tools and API's they document are generally things I'd never want to use.

Open source stuff is usually exactly the opposite.

Re:Three things.

[Anonymous Coward]

This brings up an interesting idea for improvement: Marketing. I know it's expensive and unlikely to happen, but wouldn't it be nice to see a Dell commercial showing a slick Ubuntu box with a few impressive 3D effects. Mention it is safe from viruses and spyware. Show some common tasks like web browsing, email and word processing.

Preloaded Linux is ready for the masses and it surpassed Windows long ago. Those Wow! Vista commercials would seem pretty flat along side a spinning cube and a demonstration of Beryl's scaling and alt-tab effects. Linux is ready. Now we need to get the word out there.

Sorry to post anonymously, but the mod-gods have graced to me today and you don't want to loose that +1 Insightful I gave your post, do you?

Re:Three things.

[David Gerard]

The Microsoft stuff is bad documentation that looks good. As the Wine and ReactOS projects perennially discover, it tends to lie a lot.

Re:Three things.

[Eli Gottlieb]

You just switched from Linux PC -> Macbook Pro too, huh? Check out VirtualBox an x86 virtualization app that seems to support enough to run a Linux spare-desktop or development environment (meaning no heavy multimedia) under OSX with very good performance. It comes in a closed-source edition that allows a personal/educational use freeware license and an open-source edition.

In case you wondered, yes: after getting a Macbook Pro I didn't want the hassle of running Linux on top of BootCamp (since BootCamp was designed explicitly for Windoze, damnit), and I was too cheap to shell out $80 for Parallels or VMWare Fusion. Also, I wanted to leave an open-source option open should licensing become too restrictive for me.

Re:Three things.

[Eli Gottlieb]

While we're at it, how about a configuration and build system that doesn't require me to learn a medium-sized macro language just to distribute a package that will compile and run on other people's machines? Maybe we can be really nice to ourselves and even invent such a system that doesn't care what language I coded in!

Re:Three things.

[Eli Gottlieb]

I dunno. Things like scheduler design, the VFS layer, and I/O latency do affect user experience.

Re:Three things.

[arivanov]

My biggest gripe with Dia itself nowdays is the lack of support for text labels on objects (I am looking at .95 currently shipping with Debian which is one version back from the leading edge). As a result something that should be a more or less point-n-click operation ends up being done in 3-4 steps: object, text, group, etc). Even that does not work for arrows and relations where the label has to be clearly associated with the arrow and render with it. As a result making a good looking network diagram or data flow diagram is all but impossible (labels do not stick nicely to the links).

As far as lack of integration it is openoffice fault for not caring for vector graphics (dunno how is it as far as API is concerned).

Re:Three things.

[Eli Gottlieb]

Moreover, Microsoft and Apple (really mostly Apple, but M$ has done it) can break compatibility when they want to. Quartz and Aqua honestly provide my Mac with a better user experience than I've ever had or even seen under Linux (including: Red Hat, Mandrake, Linux From Scratch compiled with KDE, Gentoo with KDE and then Fluxbox, Ubuntu and Kubuntu), but the insistence in the Linux/FOSS community on using nothing but "open standards" and Unix technology mean that nobody has started writing anything comparable to Quartz for pure *nix, even though most Linux/Solaris workstations, desktops and laptops now include the necessary hardware power.

Re:Three things.

[pivo]

It's the key mappings for Meta that are the big problem. I want my Meta key in the same place on both Linux and OSX, and that place is the keys to the immediate left and right of the spacebar. For some reason, I can only get xterm and rxvt to do that for me. That's ok, but it'd be nicer if I could just use the standard OSX terminal with the nicer font rendering. (I normally use Konsole on Linux.)

Re:Three things.

[Anonymous Coward]

On the first (a cheap Acer laptop) it complained about the wireless chip. No real options for the user, other than to get a new laptop. But it did work.

Is that really the case? What about downloading a driver or firmware?

Let's also consider what happens when you install Windows on a machine that has a chipset for which Windows doesn't come with a driver. I have had this happen many times. You install Windows, and a piece of hardware doesn't work. Sometimes it's a network card. In that case, it can be a pain to get working.

This is especially bad with Vista, by the way. But I've had it happen on all sorts of versions, from 98 to XP.

Re:Three things.

[wilymage]

Dear troll,

http://www.gentoo.org/main/en/philosophy.xml

Regards,

Re:Three things.

[cabalamat3]

You know what I'd love more than further improvement in any of those areas? Comprehensive, well-written documentation.

I agree. Both for newbies and for experienced users / programmers.

How to achieve this? Perjhaps there could be a Wiki covering all versions of Linux (and other free software Unix-like operating systems). Anyone would be able to add content to it (as with Wikipedia). And there would be automatic updates from the package repositories of various distros, so the wiki would correctly report that version 7.04 of Ubuntu uses version 12.34.56 of some package Foo (and this would be automatically updated). People would be encouraged to post problems that they have on a special area (maybe like a bulletin board) and post fixes when they find them. The fix reports would be specific, and would givew full details of what distro and package versions they used, and (for example) the exact command line commands that were used to fix the problem.

Over time, this would grow to a useful resource.

Re:Three things.

[l3mr]

Instead of calling me a troll, why not answer on-topic instead of just providing some link with PR blahblah? Using Gentoo will not give you a deeper understanding of a linux system than using Debian or SuSE, or most of the other 100 linux distros available. It might even give you less of an understanding, since more time is wasted on recompiling stuff than actually doing some work and getting hands-on experience in system administration.

Re:Three things.

[Chandon Seldon]

Are those really major Linux distributions, or are they just something you found on a list somewhere? More relevantly, how many distributions are worth considering as serious options for a given application? I don't think anyone's going to seriously consider Slackware, Gentoo, or MEPIS for a corporate workstation.

Re:Three things.

[bram]

What are you missing in "Terminal.app"?

Re:Three things.

[wilymage]

Instead of calling me a troll, why not answer on-topic instead of just providing some link with PR blahblah?

You are evidently not familiar with Gentoo in the slightest, and as such thought I may aid in the abation of your misconceptions by pointing you at a nice starting point.

Using Gentoo will not give you a deeper understanding of a linux system than using Debian or SuSE, or most of the other 100 linux distros available.

Really? That's an interesting generalisation. I primarily used BSD (mostly Free- and Net-) prior to Gentoo. I have installed and used Debian, SuSE, Slackware, RH/Fedora/CentOS, even Ubuntu. It's odd, then, that -- contrary to your generalisation -- I have gained a 'deeper understanding' from a few years with Gentoo.

I rarely watch compiles, except with new systems on unfamiliar architectures. Gentoo's package management requires less interaction that the other Linux distributions I have used, and borks much less often.

It might even give you less of an understanding, since more time is wasted on recompiling stuff than actually doing some work and getting hands-on experience in system administration.

That's like saying "I can install a Nagios virual machine in VMWare, it requires no compiling or setting up, I can do real hands-on system administration; thus, I know more about Linux! QED"

I'd seriously consider researching -- or perhaps even using -- Gentoo prior to generalising (this has already been done by uncyclopedia.org in a manner that was actually funny). Until then, I will continue to call troll.

Re:Three things.

[geekboy642]

FUD.
I haven't compiled a kernel since I gave up using Gentoo about 2 years ago. My distro has modules for most hardware supported by the Linux kernel. Every time I boot the machine, the kernel scans the hardware space, loads the required modules and (after some other stuff) passes off to the GUI.

xorg.conf is the only part of my system that needs to be manually altered when I change hardware. And THAT isn't a kernel recompile.

Re:Three things.

[jackspenn]

Why? I love Linux, but when it comes to documentation it isn't even close, MS has technet, MSDN, Q articles, MS Press, 3rd party documentation, books and on and on.

Re:Three things.

[blackest_k]

do you mean something like checkinstall

http://asic-linux.com.mx/~izto/checkinstall/

now that is easy easier than building an install for windows.

Re:Three things.

[Bonobo_Unknown]

I'd spend all that money making Linux native versions of applications like Photoshop, Office and the like.

Re:Three things.

[mwvdlee]

Right, because every newbie knows how and where to file a bug report.

Re:Three things.

[Eli Gottlieb]

No, I mean a replacement for the antiquated Unix build system of M4 (the macro language I spoke of), autoconf, automake, and make itself. Once you look into how one actually uses them, they basically only support Fortran and the various dialects of C. Other languages have to supply their own tools for generating makefiles, and then of course makefiles themselves are now so inhumanly complicated that few can manipulate them without some kind of automated tool any longer.

I'd like a build system that:

1) Programmers can comprehend and use as an incidental tool without starting a "learn autotools" project at work.
2) Supports both compilation of source files and the installation of interpreted files.
3) Part and parcel with #2, supports any language I can supply a compiler or interpreter for.

Note that build systems perform separate functions and solve separate problems from package managers. I have fond memories of working under Delphi 5 Standard: I'd activate a menu selection (maybe by its hotkey) and have my entire project update itself or even fully rebuild itself. I understand that not everyone should have a copy of my IDE, but I'd like to think that we can make the building of distributed source code that easy and language-transparent. It would save application developers a lot of hassle.

Re:Three things.

[EEBaum]

If the average user actually read all the documentation, they'd still be reading it now, and would never get to experience the wonderful world of Linux at all! What the documentation needs is an informative "terse" section ("You push the button and it makes it go") followed by the currently-existing "verbose" doc.

Re:Three things.

[l3mr]

Instead of calling me a troll, why not answer on-topic instead of just providing some link with PR blahblah? You are evidently not familiar with Gentoo in the slightest, and as such thought I may aid in the abation of your misconceptions by pointing you at a nice starting point. Yeah, right. I don't agree with you, so obviously I have no idea what I am talking about. Your arrogance is funny...not.

Using Gentoo will not give you a deeper understanding of a linux system than using Debian or SuSE, or most of the other 100 linux distros available.

Really? That's an interesting generalisation. I primarily used BSD (mostly Free- and Net-) prior to Gentoo. I have installed and used Debian, SuSE, Slackware, RH/Fedora/CentOS, even Ubuntu. It's odd, then, that -- contrary to your generalisation -- I have gained a 'deeper understanding' from a few years with Gentoo. I rarely watch compiles, except with new systems on unfamiliar architectures. Gentoo's package management requires less interaction that the other Linux distributions I have used, and borks much less often. You didn't understand what I said. Of course you can and probably will get a deeper understanding of linux from using it a long time - but this is true with any distro, not only with Gentoo. Gentoo's package management requires more interaction than Aptitude or apt-get. It's less painful that RPM, of course, but what isn't...

It might even give you less of an understanding, since more time is wasted on recompiling stuff than actually doing some work and getting hands-on experience in system administration.

That's like saying "I can install a Nagios virual machine in VMWare, it requires no compiling or setting up, I can do real hands-on system administration; thus, I know more about Linux! QED" I'd seriously consider researching -- or perhaps even using -- Gentoo prior to generalising (this has already been done by uncyclopedia.org in a manner that was actually funny). Until then, I will continue to call troll. Your arrogance again - just because I don't think Gentoo is great doesn't mean I have no experience in it. But when I use Linux I usually have to get some programming done, and I need the processing power of my PC for compiling my own programs, not the base system. And I really don't want to wait for minutes or even hours just to install some dependency I might need. But of course, if you have a lot of time and nothing to do, Gentoo might be optimal for you...

Re:Three things.

[jimicus]

I am using the the latest Intel display driver on my laptop and it fixes a lot of old problems with the driver. No more 915resolution, no more mode lines in xorg.conf, automatic detection of external monitors, etc. So what exactly does Linux lack that Windows 95 had?

Yes, but that's specific to the driver (and quite often the vendor provides a tool which writes an appropriate xorg.conf).

AFAIK (please correct me if I'm wrong) there is no generic way to reconfigure for a second monitor, change resolution/colour depth on the fly or any one of the other things that Windows and Mac users have been able to do for years short of editing xorg.conf by hand. (Well, I guess there is with resolution - ctrl-shift-+/- - but I'm going to discount that because it's not widely documented and outside of circles of people who actually know X pretty well isn't terribly well known. It's certainly not a right-click on the desktop job.)

Re:Three things.

[MajorCatastrophe]

"I would try to pool all the development into 1 distro to reduce duplication of so much effort."

I intend to release my own distro in a couple of months.
How do you plan on stopping me?

I wouldn't stop you, you go ahead.

Having one distro would act as a focus for the community and for the wider industry. Single relevant documentation sources, consistent locations for files, single package management system, single target for devs to test their apps/drivers against, and many other things could be standardised on. This would make Linux a more consistent target to develop for, test against, and provide technical support for, and to learn to administer, and use.

If you want to release your own distro and use that instead, the onus is on *you* to work around the complications caused by all the little differences. As it stands, that ball is in the user's court whether they like it or not.

Re:Three things.

[MajorCatastrophe]

SuSE is known for YaST,

And if I really like YaST, can I run it on my Ubuntu machine? No? Why not? After all, it's just an app front end to all the same settings I have on my Ubuntu box as you do on your SuSE box, or your Gentoo box or whatever else.

Re:Three things.

[gnatters]

more pr0n

Re:Three things.

[David Gerard]

What usually happens is that someone who does know how to file a bug report notices the n00b complaint and files one.

Re:Three things.

[Anonymous Coward]

Not everyone uses Gnome/KDE or even X.

And, conversely, not everyone who uses Gnome/KDE uses Linux - for example, Gnome is the standard desktop on Solaris these days. So it really doesn't make sense to consider either to be a "part of Linux".

Re:Three things.

[andrewd18]

On the first (a cheap Acer laptop) it complained about the wireless chip. No real options for the user, other than to get a new laptop. But it did work.

Did you try ndiswrapper? There's even a GUI available for installing your drivers, so you never have to touch the command line.

Re:Three things.

[Haeleth]

On the third, I attempted to test what had happened on the second, by just grabbing a random used hard drive (containing some version of Windows) and installing Ubuntu on it, accepting all default choices. Ubuntu choked. (Possible causes: multiple partitions, fragmentation, Win swap file in the middle of the area Ubuntu wanted.)

So I reformatted the drive and did a clean W2K install. Ubuntu installed over that very nicely.

Amusingly enough, I had a very similar problem when I installed a new OS on my computer recently. I had a 100-gb block of empty space that I decided to put a second OS on, just for fun, so I stuck the OS install DVD in the drive, rebooted, and clicked through endless configuration windows until I finally reached the partitioner. Here I created a new partition in the empty space and tried to continue.

Oops - the installer claimed that, for some unspecified reason, it couldn't install in that partition.

Double oops - when I rebooted to look the problem up on the web in my primary OS, it turned out that while the new OS refused to install in a lovely large empty partition, it HAD nonetheless blown away my MBR so my computer wouldn't boot at all! Good thing I had a Linux LiveCD to hand so I could fix the damage.

I eventually managed to persuade it into installing via a long and tedious process of trial and error, marking partitions active and inactive and all that jazz.

The OS in question? None other than the super-user-friendly, easiest-ever-to-install masterpiece that is Microsoft Windows Vista.

So, my impression is that Ubuntu is not really ready for the general public to install, not on a Win machine that's been in use. At the very least, not until it knows how to deal with fragmentation, Win swap files in inconvenient places, and the like. I'd even settle for a message like "Fool, this HD is hopelessly fragmented, fix that and move the swap file, before Ubuntu can help you." Bonus points for instructions on how to do those things. Double bonus points for creating a batch file that will run on reboot and do those things for you.

And I can only conclude that Windows isn't ready for the general public to install either, not on a machine that's been in use. At the very least, not until it knows how to deal with an existing OS being installed and the like. I'd even settle for a message like "Fool, I'm an arrogant OS that insists on being installed into the active partition, disable your other OS before you try to install me." Triple bonus points for not actually rendering the entire computer inoperable when the installation fails.

Linux ain't perfect, but compared to the competition, it's a hell of a lot better than Windows. It's just that most end users don't have to install Windows, so they don't know just how bad its installer is.

Re:Three things.

[Haeleth]

True. If Linux was the dominant desktop OS and MS was trying to convert users over to Win, users who wanted to add dual-boot would experience a lot of problems and be unhappy.

No, because Linux - as a result of the hacker culture that produced it - would still assume that users might have other OSes installed (one of the BSDs, maybe, or Minix, or the user's own toy OS), so it wouldn't arbitrarily blow away the boot loader without asking first.

Re:Three things.

[number11]

Did you try ndiswrapper? There's even a GUI available for installing your drivers, so you never have to touch the command line.

Oh, Ubuntu quite properly resorted to ndiswrapper, and it worked. After telling me what a Bad Thing it all was. So it handled it well enough, except for giving a warning message that there wasn't anything the user could do about anyhow.

Re:Three things.

[number11]

most end users don't have to install Windows, so they don't know just how bad its installer is.

That's true. I never said making an installer that will just pop in, no muss no fuss, was easy. Or even possible. Ubuntu is the best that I've tried, of that sample of 4 installs it did almost half of them perfectly. I'm just thinking that to be handing out CDs on the street corner, it needs to yield a higher "no problem" rate than 50%.

Re:Three things.

[mhall119]

Now I have the added benefit of being able to attach an external monitor and have it work immediately, without restarting X or whatever (which as far as I know is a feature not supported by any distribution of Linux.) Support for this and more has been added to X.org 7.3, which will be included in Ubuntu 7.10 (betas are already available, grab a LiveCD and try it out).

Re:Three things.

[mhall119]

So you want only one installation file format, only one window manager, and the ability to run windows apps? What would be the point then of running Linux at all?

Re:Three things.

[mhall119]

Honesty often has the effect of making people think you're a prick. So does being a prick.

Re:Three things.

[Reziac]

The other thing about testing, is that you've got to be open to user complaints, no matter how stupid they sound to you. If a user does something "stupid", ask yourself how the user got to that point in the first place. I've found plenty of apps (under every OS) that I can crash just by doing something out of the prescribed order, or by clicking in an unexpected place. You can't prevent users from doing unexpected things. You CAN try to cope sensibly with the effects. But way too often, the response is "Well then, don't DO that" or "PEBKAC" or even removing the feature entirely, rather than fixing the problem.

Re:Three things.

[Reziac]

"Sometimes programmers just don't have the right perspective on things because they aren't really power-users of their own software."

And sometimes they're not even *ordinary everyday users* of their own software. If they were, they couldn't help but trip over some of the pitfalls they've left for regular users. But in my observation, developers are often far more interested in the solving the creative puzzle than in viewing the resulting picture. (This isn't uncommon among creative endeavours -- frex, many actors never watch the shows they star in.)

"developers ... should really be communicating with their users better about what problems actually need to be solved."

Exactly. This is a great deal of why M$ has become so widely accepted -- they DO sit ordinary everyday users down in front of raw new software and watch how these users cope. Opensource doesn't generally have the financial resources to do that, but it DOES have a sufficienly wide userbase -- if only developers are willing to recognise that both power users =and= ordinary everyday users may well know more about *how the software behaves in the Real World* than the developer does.

Re:Three things.

[thePowerOfGrayskull]

Honesty often has the effect of making people think you're a prick. So does being a prick. Indeed, you are demonstrating this point quite well.

Re:Three things.

[Reziac]

I agree with you completely. As a user, I don't mind a UI being different from other UIs; I =do= mind when its *own body parts* apparently never met.

"It's a tough nut to crack, and, unfortunately, probably not possible in a community development model simply because there is nobody with any authority to set standards. Even the LSB is often disregarded, and the LSB doesn't go anywhere near far enough because it has to appease dozens of different distro makers."

Maybe one major distro needs to step forward and say "This is how it's going to be from now on. If you want to stay with the market, get in step," and ENFORCE it with respect to everything that goes into their distro. There will be screams of protest at first, but if the result is better UI consistency in ways that improve linux for Real Everyday Users (not just for geeks or just for n00bs) it will be worth it.

And of course if someone doesn't like it, they're still perfectly free to make their own distro, and set their own standards. Whether they'll attract any significant userbase is then entirely up to them.

Re:Three things.

[LWATCDR]

Well for one KDE and or GNOME also run on BSD, Solaris, and a few other POSIX like OSs.
They are not limited to Linux.
I do agree that they need work but they are not what I consider part of Linux.

Re:Three things.

[piojo]

Rudeness isn't helpful on either side. If someone told me that slackware didn't do a good job of teaching people about the internals of Linux, I would laugh at them and go elsewhere. But gentoo? It has a somewhat bad reputation, and it's propagated by people that misguidedly attempt to use gentoo for a speed benefit (not its main strength).

Here is what gentoo will teach you:
-you will learn about the interdependencies of Linux programs and functionality that they provide.
-you will see (at a somewhat abstract level) that when building programs, you can turn certain functionalities on or off.
-it will teach you to be comfortable with the command line, because most (all?) of the system administration tools are command line based, unlike Debian or Ubuntu, you cannot easily search for and install the software you need without using the command line.
I miss stage3 installations though, because they really walked you through the setup of a basic Linux system in a way that the new installers don't.

But it's not slackware (and sorry, I'm prejudiced, Slackware was the first distro I used for more than a month). If you want to learn a lot of what you would need to become a professional sysadmin (except multi-machine networking), you might try slackware for a while. Try installing software that doesn't come with the distribution. Try compiling a new kernel with support for more hardware. Try editing many of the core system files in /etc by hand. Learn to use vi.

If you don't agree but haven't tried gentoo or slackware for more than a week, maybe you should.

Re:Three things.

[d0rp]

Using Gentoo will not give you a deeper understanding of a linux system than using Debian or SuSE, or most of the other 100 linux distros available. Actually, it does, or at least it used to. Though the last time I checked, you could only install from stage 3 unless you really did some digging.

I used to use Gentoo, but have since moved to Ubuntu because as much as I liked Gentoo, Ubuntu is far easier to setup and maintain (though some might take issue with the latter), and I like to spend my time using my computer, not building, compiling, rebuilding.

I've installed Gentoo from a stage 1 on several occasions, and that gave me a great understanding of how the system actually works from the ground up. When you have to configure your kernel, modules, use flags, boot loader, partitions, and desktop environment by hand, it really gives you a lot of knowledge. Certainly a lot more than the Ubuntu installer does; you could install Ubuntu without knowing anything about partitions or what a kernel and modules are. All of that is fine, great actually, but if you do want to know how the system works, its going to be a lot harder to figure that out.

Re:Three things.

[d0rp]

Standard install method so all linux packages don't need to be repacked. .deb .rpm etc... pick one please I agree with this. This would be a HUGE step forward, but I suspect that this wouldn't work too well because of differences in the directory structures (among other things) of different distros, so will probably never happen

Pick a WM so all apps work/look the same on all distros. While this would be a good thing (I use gnome and hate how KDE apps like Amarok look in it), it really kind of defeats the entire point of Linux. Its all about choice. Now if they could come up with some sort of common way to describe menus, buttons, graphics, etc so that things like themes would work with all of them, that would be something I would love.

WINE, or Crossover office integrated for compatabilty for windows apps. For most apps (or at least things I've tried), this works great. VMware also works fantastically for pretty much anything other than games. Of course there is still a lot of room for improvement.

If you can get my motherboard and games running on Linux, I might finally switch. Just this weekend I installed Windows on my gaming machine after using Linux only on it for 6+ months. All the games I tried either worked (with varying degree) with either WINE or Cedega, but whenever a patch came out, it was a gamble whether or not it would make the game unplayable. So, you can do it, but its nowhere near 100%.

However, I think the burden for this doesn't lie on the Linux developers as much as it does the games developers. Why should software designed specifically for one operating system necessarily be required to run on another? If you went out and bought a PS3 game, you wouldn't expect to be able to put it in your xbox360 and have it work. What we need is the companies making the games to release versions of their games designed specifically for Linux.

Re:Three things.

[d0rp]

Well, you can change the resolution in Ubuntu if it detects your screen resolutions properly, unfortunately my monitors are 1440x900 and for some reason it put me in 1024x768 mode (so everything is stretched out) when I first installed, and put "1440x1440" in my xorg.conf file...

As for reconfiguring for a second monitor, changing resolution/color depth on the fly, you're right, there is no generic way, but if you use the binary nVidia drivers, you can do all of that (and more!) in a nifty little GUI they provide you with. Of course this is the doing of nVidia themselves.

Re:Three things.

[Xabraxas]

Yes, but that's specific to the driver (and quite often the vendor provides a tool which writes an appropriate xorg.conf).

If other vendors had open source drivers this wouldn't be a problem. The intel drivers are built to work with with the existing X framework for display drivers. This isn't a vendor specifice issue as much as it is an open vs closed source issue. Closed source drivers do things in their own way and you can only blame the vendors for that.

Re:Three things.

[mhall119]

I'm just saying, sometimes it's not honesty's fault.

Re:Three things.

[jimicus]

Well, you can change the resolution in Ubuntu if it detects your screen resolutions properly, unfortunately my monitors are 1440x900

Sounds like you're using widescreen displays. Not all graphics cards officially support them (even though if the chipset supports it, it shouldn't be too difficult to implement in a driver) so it's possible that your graphics card wasn't reporting accurate resolutions back to X.
a
The vexing thing is, X has supported multiple monitors for some time, and can happily span desktops across monitors with xinerama - but it can be a nightmare to set up. I configured it correctly once about 6 years ago and I've been rehashing and tweaking the same config file ever since. Unless something has changed drastically (and I'm prepared to accept it may have - I got a Mac bout 18 months ago because I was sick of digging around in text files to configure stuff and not getting paid for it), I have never yet seen a half-decent generic graphical tool to do it properly.

Re:Three things.

[Anonymous Coward]

LinuxDon,
With all the respect I have for you, somehow, I doubt that you have been using linux for 10 years on a regular basis. I dropped SUSE quite some years aqo but even then the yast tool provided an inteface for packageinstall, and usually got the dependences right.

With distros like a Debian, and so Ubuntu and its derivative for your DVD issue, I can only suggest to use VLC, which is the best player around, both window, linux and mac. It is maybe not the best looking one, but it is by far, IMHO, the best performer in the field.

I would be something like that
aptitude search vlc
will give you the right name for the package
aptitude install vlc-whatever (and it might be just vlc)
installs the bloody thing.
Then ... well, it works.

On most other OSes, you would need to find the package or the installer, install the thing, pray for it not destroying your filetype associations and so on and so forth. So I guess the DVD issue is just an issue of your inability to use a simple search engine.
I understand that these kind of packet management systems are now ported to OS X.

This point is actualy interesting, as it points out that using linux on a non networked box is a bit of a pain, you only have old package and cannot search for walktrough if you need advice on which app to choose.

Maybe the best thing to do would be to offer some kind of public advice website offering a pick of the most useful application, and links to their documentations. And for people who do not have network resources fast enough, maybe the posibility to get that trough a physical medium for a realtively small fees.
Or maybe offering some meta packages for some type of istallation. It prooved to be a very bad solution for computer enthousiasts, but for a lambda user, i guess having something like
aptitude install dvdplayer looks like something acceptable to type.
dvd player would be a meta package which would have for dependency x system, at least one gui, sound subystem meta packages ...

For me, I do not care about this kind of things, I am happy to tweak the install at application level, but I guess some people would like to have that. It is only possible to do that in realistic manner doing 2 kind of things:
either you pay for the use of the applications, possibly in a limited time license service (what microsoft will I think push in the future)
each of the apps is free (like in free beer), and this is a totally painless action done by a packet manager

For non gaming purposes, unixes and specialy linux outperform windows without problem on the x86 desktop market. For other architecture, it is just foolish to even think about the expensive hardware hog that is windows.

The linux community should focus on the really important things: protecting their system and model from the vicious attacks from content provider and OS makers.

The war on DRM is alreay lost, for linux, as DRMs are going to be there for the next 5-10 years at least, IMHO. This will in pratice forbid the use of linux for quite a few users that will be interested by the new services like legal unlimited music streaming offered by ISP (this is happening right now).
The new generation of media format with (useles and irealistic) content protection is either gonna make standard the use of windows as media center or push the media player use out of the computers(unlikely).
DRM on linux, I do not see that happening, unless we are talking of a decss kind of solution.

It is not some high level developer time that is needed for that, even though high level developes would proably be fit fo some of the tasks.

Done feeding the troll,
Ben :)

Re:Three things.

[thePowerOfGrayskull]

I'm just saying, sometimes it's not honesty's fault. True, sometimes it's not.

Re:Three things.

[d0rp]

The vexing thing is, X has supported multiple monitors for some time, and can happily span desktops across monitors with xinerama - but it can be a nightmare to set up. I configured it correctly once about 6 years ago and I've been rehashing and tweaking the same config file ever since. Unless something has changed drastically (and I'm prepared to accept it may have There are several ways to get multiple monitors working, none of which are terribly hard to get working, but which one(s) you can get working depends on your chipset (i.e. ATI, nVidia, Intel). The nVidia Twinview is probably the easiest to set up with the binary drivers, at least from my experience.

Re:Three things.

[NateTech]

And when did I need the Gentoo wiki the worst? During the installation process.

You either have two computers, or you end up with a screwed up install. This is not a good way to introduce average people to "Linux". Not that Gentoo ever was intended for that.

Re:Three things.

[MrCrassic]

So far, I have been able to solve most of my business application needs through the use of either Crossover Office (sucks for Outlook) or VirtualBox running Windows Server 2003 (great for Outlook). However, in my opinion, I think there are some simple user options left out in Linux that are beyond reason at this point.

One example is an equalizer. Why is it impossible to get a decent equalizer for any media application in Linux? There is a HUGE variety of media players for Linux, but NONE (except VLC) offer some kind of reasonable equalizer support! Maybe Amarok, but under GNOME it has run pretty terribly and is kludgey anyway.

Also, why is it so difficult to set up surround sound for my system? So far, it has been quite the mission to get any surround sound support for my sound card (external) under any distribution. Then again, the fault may not be necessary aimed at Linux...

Re:Three things.

[Ciaran_H]

(psssst... you can use loadable modules on Gentoo and LFS too.)

Re:Three things.

[Anonymous Coward]

> ...and how they can quickly and easily obtain helpful answers to their application-related questions.

Add "cost effectively" to "quickly and easily".

I'd use a combination of convex

[Jeremiah Cornelius]

and concave lenses, with a relatively low refractive index and arranged in an optimum series for magnification of subtle surface-details, at quite a close range - say between 200 and 400 mm.

Thanks. I'll be here all week.

Re:I'd use a combination of convex

[Anonymous Coward]

Thanks. I'll be here all week. That is assuming that you aren't murdered for making horrible jokes.

Re:I'd use a combination of convex

[Glowing Fish]

Tip your waitress?

Re:I'd use a combination of convex

[Anonymous Coward]

Why not? If she can carry the burden.

Re:I'd use a combination of convex

[Memroid]

Now, if we can just suspend fine particles in a gas, while increasing the reflectivity of these lenses, we'll be set...

Re:I'd use a combination of convex

[Anonymous Coward]

Since this is Linux we are talking about you must distribute the source, better put some mirrors in your lens distribution or you will get Slashdoted.

Re:I'd use a combination of convex

[naddington]

Try the veal.

Re:I'd use a combination of convex

[Anonymous Coward]

... Mmm, donut-shaped bokeh! *Drool*.

common ui standard

[Anonymous Coward]

Take a page from the MSFT playbook and work on a 100% common UI toolset, one that works with KDE, GNOME, and the rest.

Obvious examples: Make cut and paste work the same everywhere, same for file management.

Basically, if I open a GNOME app in KDE, it should be have the same as every other app on my screen. And vice-versa.

Re:common ui standard

[irc.goatse.cx troll]

To extend on that idea, I'd also add common keybindings and a central place to register global keybindings. Maybe even mix it with some kind of app api structure (i.e "this keybinding calls this function, or this menu entry, or this button")-- The kind of stuff that Leopard's new scripting is promising, and Microsoft's new console (MONAD?) bragged about but didnt yet deliver.

The other thing I'd personally spend some developers on is picking up XMLTerm, the abandoned mozilla Xterm project that was capable of inline rendering whatever Gecko could render. For example letting you have a shell script spit out a html table, or images.

Re:common ui standard

[Verte]

Nice suggestions!

I've got some thoughts on usability options too. I'd really like to have Emacs-y editing in text boxes in Firefox. Also, I think anywhere you are finding a file, you should be able to press down to have possible completions. Bash completion is ok, but confusing, especially when you're doing pathnames in quotes because they have spaces. And then, that doesn't work from the Gtk saveas or open boxes.

Be nice if the Xterm could inline render LaTeX, ps, pdf etc too :) I think that's asking too much though.

Re:common ui standard

[Haeleth]

Take a page from the MSFT playbook and work on a 100% common UI toolset, one that works with KDE, GNOME, and the rest.

Obvious examples: Make cut and paste work the same everywhere, same for file management.

Wait, when did Microsoft do anything like this?

Last time I checked, for example, Microsoft Excel still did cut-and-paste totally differently from every other Windows application.

Unification

[Anonymous Coward]

I know a lot of people do NOT want to hear this, but what Linux needs is an official "desktop distro" that regular people can use (in plain english: don't even think about "compilation", "scripts", "command-line" and all that computer mombo-jumbo). One GUI, one packaging system (or whatever you call that), etc.

And I don't mean "copy whatever that huge lazy corporation in Redmond is doing", either. If you can only copy, you will never lead.

Re:Unification

[Paradigm_Complex]

It's called Ubuntu. "Out of the box" it doesn't require "compilation" "scripts" "command-line" etc. There's still a few wholes, but they're being plugged quickly (a GUI to manage X, for example, is scheduled for the next major release). It's quite easy to use, I've gotten a few people to dual-boot (although I haven't seen them boot into 'doze in quite some time). Hell, the motto is "Linux for humans," humans being regular people. Compilation, scripts, etc still exist for those who aren't scared off by them, but "regular people" can get around fine with Ubuntu without them. I agree Ubuntu needs continued work, but it really doesn't need as much focus as other things in the Linux world - many of the "regular people" I know are quite content with it.

I know that this stuff is not in the kernel

[Anonymous Coward]

But I would focus on package management, so that software install is uniform accross the top distros. I know that is a lot to ask for, but one thing I think is necessary for the increasingly new user.

Make everything "Just Work"

[vonFinkelstien]

Find out all the things at take too many clicks, or require editing text files and make them "Just Work" in a simple and easy way.

Re:Make everything "Just Work"

[Orthuberra]

Thank god somebody is with me on the command line cruft. So tired of the "but the command line is a powerful tool!" Yeah it's a powerful tool to the 1 percent of the population that doesn't mind learning or using it. To the other 99% of desktops out there, Linux will just be a 'worse' alternative. Listen up slashdotters please, (almost) nobody wants to type in rpm/apt-get/emerge somepackage -u -r /dev etc etc in the command line to install a program. They want something that just works (i.e. clicking to install, maybe a login and password, maybe a pop-up menu with options of types of installs that can be done with explainations). Also better out of the box support from the wi-fi folks and graphics card folks (although this relies on said companies and not ourselves). Don't worry about the games situation, the gamer community is small anyways, don't believe me, the Wii is outselling the 360 and PS3 combined and it's not trying to appeal to them at all. Things like synaptic are a good start, but their is still a lot of work that can be done to make it just work.

Re:Make everything "Just Work"

[morgan_greywolf]

Find out all the things at take too many clicks

But, how would you fix that? It's not like finding out things in Windows takes any fewer clicks, but then, we shouldn't be using Windows as our benchmark. The idea is not to copy Redmond, but do better than Redmond.

require editing text files

Very few things require text file editing anymore. Most common system administration tasks have a GNOME GUI now in most distros. The only egregious example I can think of is in switching video cards. Get a new video card that requires a different driver, and you'll find yourself hand-hacking /etc/X11/xorg.conf with vi[m] or nano.

The worst thing is permissions editing in Nautilus. Despite the fact that it's SUPPOSED to work, I still can't change permissions recursively on all files and directories from Nautilus without resorting to the command line. I'm about this close to writing a Python-based GTK GUI script and adding it to Nautilus' action menus in Ubuntu Feisty.

The other egregious example of forced command-line usage for me was cleaning my Epson printer, aligning its cartridges, etc. This is now handled for me by my GUI-frontend to escputil, Stylus Toolbox

Re:Make everything "Just Work"

[644bd346996]

Keeping the command-line tools around is a great way of keeping the underlying system architecture clean, which makes it much easier to develop GUI front-ends. Unfortunately, it also reduces the incentive for programmers to develop those frontends: they don't need them. The solution is to pay for the GUI development if you can't afford to learn how to do without.

Re:Make everything "Just Work"

[greenguy]

Bingo. This is what keeps me from recommending Linux (more enthusiastically) to my friends and co-workers. I find myself saying things like "They've come a long way on wireless... but they still have a ways to go." Same thing for hibernation. And don't get me started on installing -- I know what "make" does, but I've been at it for several years now. God forbid I try to get a Python app going. (Yes. I do know about the install front-ends on Ubuntu, SuSe, Fedora, etc.]

You want to be 31337? Great, more power to you. Some people have work to do, and aren't interested in matching skillz with you.

I'm aware this is boring shit to focus on. But that's the stuff I want to see.

Re:Make everything "Just Work"

[Harmonious Botch]

Listen up slashdotters please, (almost) nobody wants to type in rpm/apt-get/emerge somepackage -u -r /dev etc etc in the command line to install a program. They want something that just works (i.e. clicking to install, maybe a login and password, maybe a pop-up menu with options of types of installs that can be done with explainations). Well said.

In addition, I want a teaching utility that generates a command line when I use the graphic interface, ie: "if you want to do what you just did on a command line, this is how to do it". That way, when there really is something that is better done by command line, I'l have learned enough to do it.

Re:Make everything "Just Work"

[Athaulf]

I'm with you 100% on the text-based configuration. The first thing I found that really aggravated me was how hard it was to set up a simple FTP server (the whole reason I started using Linux). I still don't even know what the last two 0's on /etc/fstab lines are supposed to mean. Thankfully though, I also ditched windows for the easier to learn GUI programming... This little annoyance is at the top of my list. I'm thinking maybe a repository of XML based configuration frontends loaded up by a single dynamic configuration program would be a good start...

P.S., I've already written my own XML parser and half of the GUI functionality ;)

Re:Make everything "Just Work"

[SniperClops]

That is why I use Mandriva on my desktop, everything just works.

Re:Make everything "Just Work"

[Anonymous Coward]

"nobody wants to type in rpm/apt-get/emerge somepackage -u -r /dev etc etc in the command line"

*I* do.

And it happens I'm codeveloper of a package management command-line based tool.

You don't want to use the command line. That's OK with me. But don't expect *I* will develop something I neither want not need out of my spare time.

Of course you are free to either hack such a killer up yourself or *pay* somebody (i.e. me) for it... or even expend some of *your* spare time instead of mine and learn to use current tools (heck, as if synaptic were so much "commandliner").

It is not as if you had no choice.

Re:Make everything "Just Work"

[Anonymous Coward]

"God forbid I try to get a Python app going"

God Forbid!

Start synaptic (you know, a GUI-based tool); look for Trac (you know, a python based tool) double click on it and there you go.

To be true I didn't use synaptic and typed `aptitude install trac`, but only because I find much more productive the command line for those tasks, not because my outlined procedure wouldn't work as told.

Any case, the single sentence `aptitude install trac` or the double click on synaptic installed a properly running Python environment, Trac itself and all its Python dependencies. How can it be much less difficult than this?

Or is it that you are trying to install apps *not* packaged for your distribution? Then the problem is *you* being fool. Or do you try installing Mac OSX apps on Vista and then say installing apps on Vista is terrible because the outcome of that?

Remember: you don't use "Linux"; you use "Fedora 7" (or whatever). So, if you lack the knowledge *stick* with whatever is packaged for your version of your distribution, just like you only install on Vista properly packaged for Vista apps. The fact that you *might* have *more* options on Linux (like compiling sources or adapting packages meant for other distributions/versions) doesn't mean those options are for you.

Re:Make everything "Just Work"

[Anonymous Coward]

Dude, don't write your own xml parser. That's retarded. Use something like libxml.

The beauty of unix style config is it's simplicity, something you clearly haven't grasped if you're writing your own xml parser. I believe that the complexity of the xml spec was intentional - to dissuade exactly what you're doing.

Re:Make everything "Just Work"

[Anonymous Coward]

You (and most of the people that replied to you, it seems) aren't looking for a better linux. You're looking for a free OSX or Windows clone.

But neither will ever exist. For one reason, anything that's "open" enough lacks a central organism that ensures consistency and conformity to a single set of guidelines, a single paradigm that's followed for all organisational and setting-related stuff. Some things will always require manual editing of the Holy etc files, while other projects will regard that as the devil's work.

Yes, I know some people will think I'm trolling, but I'm not. If it can be a consolation, the same problem exists in a lot of third-party software for windows too. And because most open source software for linux can be made to compile for OSX, it exists there too almost by definition.

Re:Make everything "Just Work"

[heinousjay]

Haha, awesome. I was waiting for this response somewhere along the way. I love the attitude, it's so great. It goes a long way towards explaining why Linux is taking so damn long to gain any kind of popularity outside of arrogant geeks.

Re:Make everything "Just Work"

[Jessta]

Now all you have to do is contact the 1000's of projects and get them all to agree that your XML format is the best and that they should put in the effort to support it.

This idea has been around for going on 10 years and it hasn't even come close to getting off the ground.

If I had developer resources I would spend them on the kernel, which is currently going fine, I'd spend them on reducing complexity(from a developers perspective) across the base system, libraries, applications etc.
Less complexity means that most developers can contribute and more users can become developers and gain real freedom in their computer usage.

Dumbing down the user interface doesn't help the free software community nor does it help the user gain freedom and GNU\Linux is about giving users freedom.

Re:Make everything "Just Work"

[belmolis]

This isn't unheard of. To toot my own horn a little, my msort sort utility has an optional GUI with a "Show Command Line" command on the File menu.

Re:Make everything "Just Work"

[Anonymous Coward]

You sir, are a troll. You had to use make because you were installing from a source distribution. Ever tried using a source tarball on Windows?

Ever noticed Windows doesn't even come with a compiler? Ever tried getting the Platform SDK ( http://www.microsoft.com/downloads/details.aspx?Fa milyId=A55B6B43-E24F-4EA3-A93E-40C0EC4F68E5&displa ylang=en ) to compile *anything*?

Next time, compare apples to apples please. And don't forget to tell those friends and coworkers about gnome-app-install (Install/Remove under Applications).

Re:Make everything "Just Work"

[WNight]

Do you know what all the values in the Windows registry mean?

Not all settings are exposed to the GUI, this is hard, slow, and usually unneeded. So when you dig deeper to find them do you want them to be

1) in a system-wide binary database you must read through a specific application
2) proprietary interfaces like Firefox's about:config
3) program specific text files

If the GUI works, all are identical. But, if it doesn't you can either have a straight forward file you can checkpoint with CVS tools, or you'll have a binary blob you can't find your changes in, let alone roll back from.

Re:Make everything "Just Work"

[Anonymous Coward]

(Yes. I do know about the install front-ends on Ubuntu, SuSe, Fedora, etc.] So why are you talking about "make" then?

Re:Make everything "Just Work"

[dbIII]

so damn long to gain any kind of popularity outside of arrogant geeks.

After a point you learn that you can do more if you read and write instead of pointing at pictures :)

Re:Make everything "Just Work"

[Error27]

You get annoyed by make and I get annoyed by people who suggest that make is a valid way to install applications under linux.

In fact if you use "make" on a production system, you're probably doing something wrong. Package management is not just a "front end" to "make." Most of it is very definitely back end.

Packagemanagement is a complete record of what you have installed. It track library dependencies and conflicts. It allows for easy security updates. Once you start using make install to go around the back of your packagemagement you lose any kind of maitainability.

Don't try be clever folks. Use the package management and you'll make all you co-workers' lives easier.

Re:Make everything "Just Work"

[Athaulf]

Yes, I have to admit, it probably wasn't the best idea... I knew about libxml when I started writing the program, but writing my own xml parser sounded pretty fun, so I went for it. I probably will switch it over to libxml if I ever make a release version because I don't want anyone who looks at my code to hate me, but for testing at least, it's usable for me. I'm still learning C and this helped me learn how to make inheritance with structures (similar to GTK's system), so despite being a waste of time, it was pretty fun.

Re:Make everything "Just Work"

[Athaulf]

I'm hoping I won't have to contact the developers because I'm hoping to create a user supported repository, so anyone could make a configuration utility for their favorite programs and put them up for everyone else to use. I wondered why there didn't seem to be one out there already, but maybe I should have looked harder. Also, what do you mean by "dumb down" the user interface? Considering the fact that there isn't a user interface for configuring many server programs like samba, apache, and vsftpd, I'm not sure this is possible.

Re:Make everything "Just Work"

[Athaulf]

I'm hoping all three, but no one has made a system wide database (which has its strong points and its weaknesses) and I don't intend to. The proprietary interfaces are great because they're a part of the program, and who wants to open up another program just to disable a button in firefox? If your making a reference to the many plugins available in firefox, you're right, I/the user who wants to create the script probably won't be able to write a script for each of the plugins, but for the main program, and any larger extensions, there needs to be some way to configure them if they lack one. I don't much enjoy searching through man pages for some obscure variable to set in a text file somewhere in /etc. Also, it's worth pointing out that I don't plan on centralizing the many configuration files, I plan on editing them in place. The idea of the program is based on the idea that the target programs are usually configured using a text editor, and keeping you away from it.

Re:Make everything "Just Work"

[WNight]

What I mean to say is that, because they are text files, you can do this [your plan]. If these settings were stored compressed in the registry you wouldn't be able to do this except via MS's API which requires the system to be booted and untrojaned.

And I still don't see why you have to dig through a man page to tweak a config file, but don't have to read the docs to change registry settings...

Really, it's just a choice between putting "foo = 7" into a file with a standard format (text) instead of a proprietary format (registry). You still need the name, the value, and the correct location. Occasionally syntax is an issue, but you can usually copy that from another line more easily than copying a registry tree.

Re:Make everything "Just Work"

[Athaulf]

What I mean is that you have to search for the names and valid values. It takes me forever just to find the documentation, let alone the variable that makes the changes I need. Half the time I'm also missing out on some of the functionality of the program because I haven't found some of the options. This is where configuration GUI's really shine, because you don't have to know the names. It also leads you into finding a legal value through check boxes, entries, browse buttons, and so forth.

Re:Make everything "Just Work"

[WNight]

Sure, a GUI with those options is great. But even in GUI programs most settings aren't user-visible. Firefox for example has thousands of settings in about:config, only 100 or so of those are in the GUI.

Re:Make everything "Just Work"

[Jessta]

So, you idea is to convert current configuration files in to xml files, use them for configuration and then have them converted back in to the current configuration files formats for the individual applications. This has been thought of before many times and implemented many times. It sounds great is theory, but when you actually look at it, it's a lot of work for little gain. Have you seen the apache config file? Do you really want to create an xml version of that?...and a GUI interface for that XML file? and then do you want to do the same thing for thousands of other programs? Maintain and update these for config format changes?

Re:Make everything "Just Work"

[Anonymous Coward]

It "just works" just fine when you understand Unix. If you want a Windows clone, go play with ReactOS.
Who the fuck cares if joe user doesn't want to learn a new OS? Why is Linux so obsessed with the grandma market?

Re:Make everything "Just Work"

[Athaulf]

I guess I didn't explain my idea well enough. I don't want to convert the config files themselves to xml, I want to create an xml script that describes the GUI. The program itself will load up the xml and use it like HTML. There is a tag that creates an text entry, one for creating a file browser, one for creating tabs, one for creating check boxes, etc. It will then follow patterns which the script writer can specify in the XML code to output what the user selects to an external configuration file (like httpd.conf in apache) easily. And no, I don't intend to include every single feature apache has to offer in the GUI. If you're technical enough to realize the purpose of many of those features, then you're probably going to prefer the text editor over my any XML GUI I can create anyway. For example, the main things I want for the apache XML script is easy CGI configuration and the ability to activate SSL. I was also thinking about adding in shell scripting so that you could, say, restart apache whenever the user saved their configuration file. However, when talking about user-submitted scripts, this seems to be a bit too large of a security hole. Eventually, I'd also like to see this either integrated into a GUI package installer like synaptic, or have one built around it, that way you could install and instantly be given the option to configure whatever you've just installed.

Re:Make everything "Just Work"

[sowth]

The first thing I found that really aggravated me was how hard it was to set up a simple FTP server

Problem is: to make a easy ftp setup, the project needs to make it easy to set up the server securely, otherwise people who use the easy tool will just be screwing themselves. It won't be very easy if your box is rooted.

I still don't even know what the last two 0's on /etc/fstab lines are supposed to mean.

They are both optional. The first one indicates if you want to backup with dump. The second one indicates whether you want the filesystem fscked at boot (0=no check), the fs is the root filesystem to be checked (a 1), or it is a non root filesystem to be checked (listed as 2). It's all in the man page for fstab. Problem is, the GUI tool to look up man pages (xman) sucks ass--you can't type in the command/file to look up and you have to select it from a HUGE list. Maybe there is a better one out there, but I haven't seen it...

I'm thinking maybe a repository of XML based configuration frontends loaded up by a single dynamic configuration program would be a good start...

If you want to get rid of the somewhat human readable text files, then at least you should go for a more machine readable format. Yeah, you can't edit binary files with a text editor, but it would save programming effort, complexity and CPU time. Why should we be afraid to use hex editors when needed anyway? ;-=)

Re:Make everything "Just Work"

[Athaulf]

http://ask.slashdot.org/comments.pl?sid=279871&thr eshold=1&commentsort=0&mode=thread&pid=20362805#20 363083

Re:Make everything "Just Work"

[Ex-MislTech]

The caveman first used pictures to indicate what he wanted, after many long years,
they started using words...today we call it language, window users call it CLI, lol.

Re:Make everything "Just Work"

[houghi]

It is easier to install Linux then it is to install Windows. Here a BETA install for openSUSE 10.3. http://houghi.org/moving_pictures/openSUSE10.3_ins tall.php

Re:Make everything "Just Work"

[Anonymous Coward]

Amen, thank you for pretty much posting exactly what I was thinking. I don't exactly consider myself a computer whiz by any means, though I've been building my own PCs for about 10 years now, do a little VB coding just to fiddle, and I can get around and do some things in DOS.

Now, I admit that I did install Ubuntu on my mother in laws old Gateway, however, as long as she can turn it on, log in, surf the net and play solitaire, she's pretty much a happy camper, unfortunately thats about all I can *really* count on Ubuntu to do without at least as much headache as cleaning out a windows machine full of crap every couple of months.

Aside from the mother-in-laws machine, *I* really have tried on several occasions to use Ubuntu as my sole OS. I tried on my laptop for awhile as its too old to be of any use for serious gaming, and I also tried it on my desktop, (which I do use for gaming quite a bit). After spending 3 days tracking down obscure forum references on ways that I *might* be able to get the wireless card working, failing miserably I might add, I finally gave up and went back to XP. After spending almost as much time trying to get 3D support enabled on my 9800pro and Enemy Territory installed and working with a custom .cfg I was like, screw this. Sad thing is, Enemy Territory was a game actually made for *nix. First I installed it into what ever the default location was using sudo, only to learn that I had to remove it and use some other bizarro command line to get it installed cause sudo is, apparently a no no for installing ET, although it was the only way I was able to do it without, yesssss....trolling the Ubuntu forums and noobing it up in the Ubuntu IRC channels for a couple of hours. Until programs MADE for *nix are clickable and installable without any other hassles? It's going to continue to be in the realm of minimal adoption.

For the record, I HATE using MS software, I'm just dying for someone else to come up with something as easy to use.

GUI integration

[Anonymous Coward]

I would focus developers on integrating a GUI into the kernel. It makes the OS more stable, right?

Re:GUI integration

[khellendros1984]

It would make it higher performance, but less stable. Tie something into the kernel, and there's a bigger chance it'll hose the system as it goes down.

Re:GUI integration

[Anonymous Coward]

I would refocus your speech so it no longer originates from your ass.

Re:GUI integration

[piojo]

I would focus developers on integrating a GUI into the kernel. It makes the OS more stable, right? Think of the kernel as some sort of Giant Robotic Overlord that does everything. It knows everything. It is everywhere. And it does all this very, very quickly.

The gui is more like a sleeping cow that gets woken up every five years (in cow years) when a user clicks it or types a key.

If the Robotic Overlord had the cow tied to it all the time, it couldn't efficiently perform its other duties.

Re:GUI integration

[sowth]

It is good from a performance standpoint to put drawing functions into the kernel. That is why they made DRI. But actual GUI functions would just be a mess. Toolkits and such work fine in userland. My wish would be if they integrated graphics into the virtual terminal switching so one could easily have multiple graphical apps beyond running them on a X server.

Figuring out how to make a framebuffer app not screw up when the virtual console was switched really really sucked. Maybe things have improved since I tried last, but I don't think so. I don't think my proprietary driver NVidia card even supports the framebuffer at all anyway. This is why I just use X all the time. Most apps are written for it anyway. Oh well.

That's easy

[Anonymous Coward]

Kill desktop Linux since it's already DOA.

Stop wasting time and develop for Mac OSX instead.

Re:That's easy

[khellendros1984]

Unless Mac OS opened itself up (both source code-wise, and in regards to the hardware it will run on), it just wouldn't be the same.

Re:That's easy

[moderators_are_w*nke]

A major reason why everything just works on Mac OS is because they only have a limited subset of hardware to support.

My reasons for using a Mac are simple, I don't have to set it up, my hardware just works, it has MS Office for work as well as being a unix platform for running my code in the right environment. Also the GUI works well, simple things like expose and spotlight are really useful day to day.

Package management is also nice, in most cases installing an app is a case of dragging it into applications and to uninstall you just drag it into the bin. Where setup programs are required that are pretty straighforward and consistent. Windows could learn a lot from this, most Windows installer are overcomplex and unnecessary. I really don't understand why having an MSI installer package is part of the Vista logo requirements.

Re:That's easy

[Ruie]

A major reason why everything just works on Mac OS is because they only have a limited subset of hardware to support.

My reasons for using a Mac are simple, I don't have to set it up, my hardware just works, it has MS Office for work as well as being a unix platform for running my code in the right environment. Also the GUI works well, simple things like expose and spotlight are really useful day to day.

Which is all fine and dandy, until you want to do something that has not been done before - at least not by 99% of users. For example, suppose your application requires more than 8 cores or more than 64GB of RAM. Or both.

With Linux you just buy a bigger box from IBM or Sun and off you go tuning your application to actually use more hardware.

With Mac OS, you also buy a bigger box from IBM or Sun and port your app to Linux first..

Re:That's easy

[moderators_are_w*nke]

Fair point, but there really aren't that many applications tha fit into that category (mainly scientific I would imagine).

If you develop in Java, as I do, this is not an issue anyway because your application will run on Mac OS or Linux without any changes. Obviously Java is not suitable for all applications, especially in the scientific community who have vast amounts of fortran they would like to reuse but as someone who develops webapps most of the time, its never been a problem for me.

Disk throttling

[Anonymous Coward]

Stop disk hog processes from slowing everything else. If it's disk-intensive it is expected to take a while anyway.

Re:Disk throttling

[ChrisMaple]

Excellent suggestion. I find that cron jobs like updatedb and tripwire slow my system to a crawl, even when I renice all aspects of the cron jobs to +19 and lower the jobs I want running fast.

UI Consistancy

[sqrt(2)]

Al the time on Linux it seems every program looks different and out of place, the only ones that fit are the ones that come with the DE and so are made to look all the same. Take a look at KDE apps on Gnome or vice versa. On windows everything uses standard widgets and themes. And I'm not talking about stuff like Winamp that uses a skin, but take a look at Pidgin on the windows platform, an open source project that looks completely at home on a Gnome desktop. In XP or Vista, the menus and windows aren't drawn the same as other apps. It's more of an annoyance than a mission critical problem, but it really detracts from the professional and complete image that I know Linux is capable of showing.

Re:UI Consistancy

[scapermoya]

except that KDE and gnome apps, in a UI sense, are kinda sorta mutually exclusive, no?

Re:UI Consistancy

[Anonymous Coward]

"On windows everything uses standard widgets and themes"? WTF? Do you even use Windows at all?

Re:UI Consistancy

[timeOday]

Well, the original poster did repeatedly say "kernel," not UI widgets.

(Personally I think common *functionality* (such as all programs using the same sound subsystem, e.g. alsa vs. oss, or printing) is far more important than widgets and other eye candy anyways, but to each his own!)

Re:UI Consistancy

[Orange Crush]

except that KDE and gnome apps, in a UI sense, are kinda sorta mutually exclusive, no?

But there's still the great schism between KDE/Qt apps and GNOME/GTK apps. Regardless of which DE you use, you'll encounter applications that use both toolkits and they're going to look inconsistent and comparatively out of place.

Re:UI Consistancy

[Orange Crush]

Well, the original poster did repeatedly say "kernel," not UI widgets.

Preceded by "applications or", which would include widgets and DEs.

Re:UI Consistancy

[EvanED]

Why's that?

Why couldn't Gnome change their toolkit so that it produces widgets that have the look and feel of KDE's, or KDE (or maybe Trolltech) change their toolkit so it emulates Gnome?

I can thing of three reasons:
1. Developer egos
2. Other features seem more pressing, so get priority
3. It might break compatibility with stuff like themes written for one or the other

Re:UI Consistancy

[EvanED]

I use Windows as my primary OS. While it's not an ideal world, and there are plenty of inconsistencies (everything from quick little Java apps that someone pounded out to big guns like MS Office), on a whole, Windows is far more consistent than my Linux setup. It's like comparing night and day.

Re:UI Consistancy

[EvanED]

I would say that the real culprits are the applications that properly use neither of the big ui toolkits, integrating with nothing, and looking / feeling out of place everywhere!

Agreed. That said, IMO in an ideal world, even the latter distinction would be unimportant from a look-and-feel point of view.

(I'm willing to forgive Firefox and OpenOffice because, at least under my setup, both look very similar to Gnome apps, to the point where I wasn't sure if they were using GTK for their widgets or not.)

For example, thanks to some laziness and lack of vision, tabbed interfaces have caught on. ....

I've liked the Fluxbox idea for a while. It's too bad that more window managers don't support something like that. The Fluxbox idea's also more flexible than typical tab implementations, though Pidgin's are pretty good too.

(That said, in the absence of popular support for WM-managed tabs, I think that application-managed tabs are often better than nothing. There are also sometimes cases where tabs done at a level below the top-level window is the right way to go, though the WM/widget toolkit could still provide the tools to build them and make them consistent.)

(sorry about the long-winded, rambly spaghetti posting by the way)

Heh, no problem. I just posted one of those, though even worse. ;-)

The task bar's preview has no way of showing anything but what the window has currently open via what amounts to its own built in window manager that switches between documents, just like what can already be done with top-level windows!

I haven't really used Vista and don't use IE, so I'm not sure what you're referring to exactly, but I think that some of this opacity is one of the big benefits of tabs.

I wouldn't would want alt-tab for instance to cycle through all of the tabs I have open in Firefox, and I don't want the taskbar to show separate icons for them. This is mostly because of the number -- I have far more tabs open than I do total windows, and I don't want my FF tabs to overwhelm the number of other applications. (I also hate the taskbar grouping feature and turn it off as one of the first actions on a new install or new system. If you like that, the problem is rather lessened I think.)

Having this problem is not innate to Window Manager-managed tabs, but the ideal WM would be cognizant of the issue and would at least allow some configuration of how to handle it.

(In other words, the present situation is, as you say with "the window has currently open via what amounts to its own built in window manager that switches between documents", the WM switches between windows and the application has a different WM that switches between documents. I assert that getting rid of the idea of the distinction between switching between applications and switching between documents is not a good idea, and that we should still be left with the ability to have this sort of two-level affect, just with both levels managed by the WM. If that means that "window" is replaced by "group of windows" and "document" is replaced by "tab within a group", that's fine.)

Re:UI Consistancy

[Anonymous Coward]

And I'm not talking about stuff like Winamp that uses a skin, but take a look at Pidgin on the windows platform, an open source project that looks completely at home on a Gnome desktop. In XP or Vista, the menus and windows aren't drawn the same as other apps.

So, Pidgin on XP or Vista isn't consistent with the Windows UI, ant this somehow becomes a problem with Linux?

IMHO it's a problem on every desktop environment. Run something from outside, and it won't be consistent. Heck, a lot of times you will find programs that were written specifically for that desktop environment, and still doesn't fit (Winamp).

It doesn't matter if you are talking about GTK apps on Windows, KDE apps on Gnome, CDE apps on OSX, or DOS programs on Windows. They are always going to be inconsistent with the desktop they are running on.

Real time desktop

[Anonymous Coward]

I'd like a system where audio and video never glitch, and the GUI is always responsive.

Re:Real time desktop

[Anonymous Coward]

I'd like a system where audio and video never glitch, and the GUI is always responsive.

QNX

Next?

Re:Real time desktop

[faragon]

QNX and LynxOS are great, but closed source, ergo without future.

My bet is that in the future Linux will be both 100% POSIX compliant and able to run as a true Hard Real Time OS.

Re:Real time desktop

[tepples]

I'd like a system where audio and video never glitch, and the GUI is always responsive. QNX

Next? I'd like a system where audio and video never glitch, and the GUI is always responsive, and the applications that I use are available for the system, and hardware designed for use with the system is easily available.

Software development tools

[iplayfast]

I know that gcc is a great compiler, and vi and emacs are wonderful, but I really miss the convenience of select and drop gui development. I also like IDE's, with context sensitive help, and class completion and all the other things they do.

It seems to me that windows development tools are ahead of Linux in these regards, and it would behoove Linux developers to make development as easy as possible.

Re:Software development tools

[jeevesbond]

I also like IDE's, with context sensitive help, and class completion and all the other things they do.

Honestly not trying to flame you here, but have you tried Eclipse? Am pretty sure it does what you need: for Java anyway, it certainly has basic support (syntax highlighting etc.) for other languages like C/C++ and PHP too. If you're into .NET there's Monodevelop, which is an IDE for that.

Re:Software development tools

[Anonymous Coward]

There is also REAL Basic if you want a cross-platform, VB-like, "simple" language.

Re:Software development tools

[sohp]

Well, aside from what the other reply pointed out about the IDEs you clearly have missed in your exploration, I gather that what you really want is VB6 for Linux?

Re:Software development tools

[Anonymous Coward]

eclipse

netbeans (recently revamped, much better than before)

codeblocks

Re:Software development tools

[iplayfast]

I'm a C++ programmer, and I've tried eclipse, for C++ it isn't that good. If I was into Java I'm sure I'd turn straight to eclipse.

Re:Software development tools

[iplayfast]

No, I'd rather have something like CodeGear (Borland's) C++Builder.
I've used VB once. It was an interesting half hour...

I refuse to answer that question.

[drolli]

and so should al large part of the slahsdot readers. I think the Bazar is a good development model. Without doubt "devoting a large number of devolopers to a task" is the opposite of that. Again thing which have a factor of 100 in man-years will show up in the lists of the readers.

Re:I refuse to answer that question.

[iamdrscience]

Are you insane? Developers are developers and how they come together to work on an open source project means nothing, the point is that they're working on the project. If you really believe from reading "The Cathedral and the Bazaar" that ESR intended to kick out developers who don't totally fall in line with FSF ideals then you're crazy. In fact, I think you'd have to be crazier than ESR himself which is a truly amazing feat.

Re:I refuse to answer that question.

[chromatic]

If you really believe from reading "The Cathedral and the Bazaar" that ESR intended to kick out developers who don't totally fall in line with FSF ideals then you're crazy.

Well yeah, but that's because CatB is a critique of the FSF's development process.

Re:I refuse to answer that question.

[drolli]

I think you got my point wrong. My point is: there are a hell lot of things which i whish to see in linux. I also praise the capabilities of the leaders of the development process to steer the larger projects. However i believe that the linux development process right now is going a good way. Let's just say that larger and larger shops open on the bazar and they adress things which they (and me) whished to see addressed. Actually it seems that a real lot of companies found different roles in linux and are doing well in focusing on their specific small part of the problem. Let's take e.g. Desktop Support: Even and old command line freak like me, whos last system operated mainly by GUI was OS/2 found it effective and simple to use the GUI for nearly everything under Ubuntu (please no fundamentalist discussions about the incoherence of Linux GUI programming). So obvioulsy the development refocused itself. Take the kernel-related issues of that (the Scheduler), and they are adressed. My experience is: the more "belief on focused" deveopment was present in the community, becaused in some projects development power was wasted on things which just where no accepted afterwards.

The right question would not have been "how would you refocus", but "what is most annoying to you", because i do neither believe in my own capability to anwer the first question nor in a large parts of the slasdot reader capability to do so.

Fix RPM Dependency Hell

[Chlorus]

Ensure greater collaboration between RPM repositories/distros so we don't have errors such as mess as we do right now. As is: "You installed Program 1. It needs program 2." **installs program 2** "You installed program 1. It still needs program 2. It looks like you installed program 2A. "

Re:Fix RPM Dependency Hell

[Anonymous Coward]

Umm... when was the last time you encountered this?

Re:Fix RPM Dependency Hell

[Anonymous Coward]

Seen it before. Usually, compiling program 1 from source fixes it. Never seen one that ./configure;make;make install as root wasn't good enough.

Re:Fix RPM Dependency Hell

[Schraegstrichpunkt]

Install Debian? Install anything that's been released within the last 2 years?

Re:Fix RPM Dependency Hell

[CronoCloud]

Someone reported something like this on the Yellow Dog Linux mailing list a few days ago. It still happens.

Re:Fix RPM Dependency Hell

[Anonymous Coward]

Sounds like someone is ready to come join the Debian camp...

Re:Fix RPM Dependency Hell

[donaldm]

You can have dependency problems is you just use RPM's however there are tools such as "app-get" and "yum" that reduce this issue to virtually zero although you do have to be careful of the repos you pick since they have to work with each other. Were people run into issues is having to many enabled repos.

To have all repos enabled and accepting all is just plain silly since you are going to have issues. What I do is to activate one repo at a time in a preferred order with the primary repo first. I never do an installation or an update without checking what I am getting and I will only answer "yes" when I am satisfied.

The Hurd

[John Hasler]

if you were able to devote a 'significant' number of resources (read: high-quality developers) to a particular app or area of the kernel, and were able to set the focus for those resources (stability, performance, new features, etc.), what application or kernel area would you attempt to improve, and what would aspect you focus on improving?

I'd budget $1M/year for a minimum of five years for full-time work on the Hurd. No, it isn't Linux but it is an alternative kernel with interesting features that is sadly stagnating.

Re:The Hurd

[Verte]

We don't even have a USB driver yet. The Hurd really could use some more coders. That said, it looks like it's coming along nicely. Markus is porting the Hurd to L4, but the Coyotos team is working pretty slow, they seem to be focusing on the CapIDL more than anything [which makes sense I guess, otherwise we would have a kernel we couldn't write anything for]. I don't think it'd take many more developers to get it up to Linux' standard, to be honest, so we really need to generate more interest in the hacking community. I mean, what is more romantic than microkernel hacking? Especially when it's something as forward thinking as Hurd-on-Coyotos.

Re:The Hurd

[Slashcrap]

I'd budget $1M/year for a minimum of five years for full-time work on the Hurd. No, it isn't Linux but it is an alternative kernel with interesting features that is sadly stagnating.

No, it doesn't have interesting features. If it had interesting features it would not be stagnating and lacking in developers. OpenBSD is an alternative kernel with interesting features and that is why it has enough developers and support to be usable.

There are no end of unfinished OS projects with a couple of developers that move to a completely different kernel every six months and will never, ever be finished. I don't see how the Hurd stands out from them. I foresee your $5M funding nothing except a huge amount of meetings, committee decisions, politics and rewrites. You'd be better off selecting a random microkernel based OS project on Sourceforge and giving it to them. They might actually produce something.

I don't want to discourage anyone from working on the Hurd if they find it interesting, but giving them $5M and expecting something useful to come out is just ridiculous.

Re:The Hurd

[ksd1337]

I think they should merge the Hurd with Linux, but keep the Hurd project alive.

Games

[Anonymous Coward]

Get me my popular Windows games on Linux, and I'll switch in a hearbeat.

Re:Games

[tepples]

Get me my popular Windows games on Linux, and I'll switch in a hearbeat.

Get me my popular Xbox games on PlayStation 3, and I'll switch in a hearbeat.

What's the difference?

Might I Suggest...

[bealzabobs_youruncle]

that anyone who thinks that CLI usage is not a feature of Linux think again? This topic is 12 minutes old and three post have already suggested we bury the command line; part of what makes Linux so fast, flexible and customizable is access to virtually every setting from a text editor. This is not something that needs to be changed, instead change your mindset that this is not Windows.

If you are looking for a completely GUI drive *nix I would say OS X is your best bet (yes, I know you can use the CLI in OS X, but you never have to unless you so desire).

Re:Might I Suggest...

[mikesd81]

I agree. I do alot of work from shells or xterms even when I'm trying to customize something instead of clicking through dozens of windows and screens I can just edit a config file make a setting save it and presto. Actually, the only thing I really require X for is surf the web and grab mail.

Re:Might I Suggest...

[grumbel]

We definitively shouldn't bury the command line, but on the other side we shouldn't just let it stay the way it currently is forever. Currently there is a huge gap between the command line and the GUI, the most you can do in terms of interaction between the two might be copy&paste of a file name, but thats basically it. What I would like to see is some more monad/msh or XMLTerm'ish stuff where you don't just deal with lines of text, but have proper objects that you can move around between command line and other applications freely. It might also nice to have more elaborate output then just text, i.e. thumbnails and stuff. And most importantly the GUI simply should work better together with the command line, for every thing that I can do with the GUI, I want a way to access that with the command line and visa verse.

Re:Might I Suggest...

[avanderveen]

I agree with you in that the command line is amazing and should be kept, but we should not have to always deal with it. A good example would be some installers in Linux that are completely GUI based and meant to be user-friendly, but in order to launch them you have to chmod them or whatever and then run it. Basically, if people expect Linux to be adopted and to still be Linux at heart, the command line should exist and not be crippled, but developers should create more user-friendly apps that focus on the UI and useabiliy. That's why people are switching to Macs, apparently they "just work" (I disagree, but that's another topic) and Linux needs to do this too.

Re:Might I Suggest...

[smallpaul]

that anyone who thinks that CLI usage is not a feature of Linux think again? This topic is 12 minutes old and three post have already suggested we bury the command line; part of what makes Linux so fast, flexible and customizable is access to virtually every setting from a text editor. This is not something that needs to be changed, instead change your mindset that this is not Windows.

You've set up a false dichotomy. The other guys are saying that every setting should be available from a GUI. They did not say that the settings should be unavailable from the command line. Can we put away the "You're with us or you're against us" mentality? A person could be in favour of BOTH more powerful command lines and text files AND more powerful GUIs. The two are very complimentary. Most of the progress Linux has made over the last decade is because people have put aside those false dichotomies. In fact, this is often true for the technology industry in general...

Re:Might I Suggest...

[jeevesbond]

This topic is 12 minutes old and three post have already suggested we bury the command line; part of what makes Linux so fast, flexible and customizable is access to virtually every setting from a text editor.

Whilst I totally agree with what follows after the semi-colon in this sentence am not so sure about the part prior to it. All we're seeing is that people do not want to be forced into changing settings--am assuming, on their desktop machines--using the command line. This does not mean we should 'bury' the command line, or stop using text files to hold settings! In fact you've made my point for me:

If you are looking for a completely GUI drive *nix I would say OS X is your best bet (yes, I know you can use the CLI in OS X, but you never have to unless you so desire).

Aye, there's the rub! The user should be able to choose between a GUI configuration interface or editing a text file: everyone's a winner! Also a GUI should be able to read/write text configuration files whilst handling seperate user changes to those files gracefully.

In fact I'd spend a lot of the money on getting everyone (or as many projects as I could) to agree to a configuration file format that could easily be interpreted by an application. A one-size-fits-all library could be written to get the settings from file into memory and back again, then it would just be a matter of organising that data into a front-end that's meaningful for the user. The real joy is that with a standard file format, and library to support it, a rudimentary GUI for a new application could be created in minutes.

This is not something that needs to be changed, instead change your mindset that this is not Windows.

This is a very conservative viewpoint, why can things not change? Why can't we have the best of both worlds, with both GUI configuration tools and text files?

Re:Might I Suggest...

[ConanG]

(yes, I know you can use the CLI in OS X, but you never have to unless you so desire) That's exactly how it should be. I don't think the CLI needs to be buried (nor do I suspect most people do, either). Right now, it's too in-your-face. That's why a lot of people like OS X. The OS gets out of your way and lets you get things done. The CLI is part of that. If you need to go mucking about in the guts, you can. Most people don't need or want to go mucking about, so it stays out of the way.

I think that's one of the biggest issues with Linux right now. It's designed on the premise that people need or want to go mucking about in the guts of the OS. Windows is designed on the premise that users are to stupid to do so. OS X is designed on the premise that people don't need or want to mess with the OS (but can if they so choose). For me, OS X wins...

Re:Might I Suggest...

[Chandon Seldon]

for every thing that I can do with the GUI, I want a way to access that with the command line and visa verse.

This is false. You don't actually want a way to do everything with the GUI that can be scripted. For exceptionally esoteric features that are only of interest to programmers, sysadmins, and CLI-aware power users it's perfectly reasonable to have them only be accessable though command line options or a non-GUI config file.

That's how most programs work. Even Windows with WSH and the Registry. Hell, even Windows video games have options that can only be accessed by editing some config.ini file somewhere. If you tried to fit all that stuff in the GUI, you'd never be able to find any of it.

Re:Might I Suggest...

[EvanED]

None of those posts suggest burying the CLI.

They are suggesting making the CLI unnecessary to do most tasks, something that would be a virtue, not a vice. Power users and gurus could still use it to get the added power and speed.

The fact that you suggest that if you want a GUI you should use OS X is exactly proof of their point. There's really no reason that they can't coexit, and the only issue with making them do so would be that the effort required to bring Linux up to OS X's standards would distract developers from things that you feel are more important. This is a valid concern, but it's rather a different shade of argument than the one you're apparently making.

Re:Might I Suggest...

[Chandon Seldon]

In fact I'd spend a lot of the money on getting everyone (or as many projects as I could) to agree to a configuration file format that could easily be interpreted by an application. A one-size-fits-all library could be written to get the settings from file into memory and back again, then it would just be a matter of organising that data into a front-end that's meaningful for the user. The real joy is that with a standard file format, and library to support it, a rudimentary GUI for a new application could be created in minutes.

This idea keeps getting suggested, but I see no evidence that it would actually work (or even be a good idea) in practice. Can a single format really work efficiently and elegantly for every text-based config file on a Linux system? Everyone who's invented a new file format for their program thought they had a good reason to do so - do you really know more about their programs requirements than they did? Backwards compatibility is important, given that, would any distros be willing switch to your new format and break every package they upgraded? Would a generic config-editor GUI actually be any better, in practice, than a text editor is?

Does this issue even matter? Are there really any text based config files that desktop users need to edit?

As a data point: The only text-based config files I've had to edit on the Ubuntu machine I'm sitting at right now is my ~/.vimrc - and I think the world is fine without a GUI for that.

Re:Might I Suggest...

[SpacePunk]

Your right. Personally, I think computing when down the shitter when we stopped using punch cards and paper tape.

Re:Might I Suggest...

[Anonymous Coward]

"I think that's one of the biggest issues with Linux right now. It's designed on the premise that people need or want to go mucking about in the guts of the OS."

And that's a problem how, exactly?

If you want Windows, use Windows - don't piss on everyone else's parade and try to make Linux into something it isn't, just because it's cool to hate Bill Gates.

Re:Might I Suggest...

[ConanG]

If you read the entire post you'd see I said I think OS X wins. I'm pretty much done with Windows (just a couple of games now and then in Boot Camp).

The question asked is how I would like to see Linux improved. I'm basically saying that I don't think the user should be expected to need to muck about. If they want to, no problem. Most distros of Linux almost demand it. OS X doesn't. Therefore, I like OS X more.

Re:Might I Suggest...

[oatworm]

Things I've had to hit up the command line for and wouldn't put up with from Windows or OS X on my Ubuntu Feisty Fawn installation:

Installation - I have a Compaq Presario V6000-series laptop with an AMD chipset. Apparently, it has some BIOS issues that require special flags in the initial boot loader screen on the Ubuntu install CD. To find these flags, I had to hit the Ubuntu forums; there was absolutely no way that I could've known which flags to plug in unless I had prior experience with such issues. Ideally, Ubuntu should have figured out that I might have a troubled AMD chipset and provided the correct flags for me. This could probably fall under "hardware support". Alternatively, a quick GUI that listed what advanced options were available to me would've been helpful, too. Heck, even BSD's boot loader will give you a basic list of commands to muck around with, and it's text-based. That would've been more helpful than Ubuntu's CD.
Wireless - I know wireless is a big issue on Linux, especially with Broadcom devices like the one in my laptop. That said, getting wireless working eventually required the use of ndiswrapper, and, if there's a clean GUI for that on Ubuntu (not saying there isn't - didn't look all that hard), I didn't see it recommended. Again, this is probably as much a "hardware support" issue as anything; ideally, the drivers would've already been in Linux and we wouldn't have these problems.

Other wish-list items:
Drivers - Look, I get the Linux driver model and why some people find it great. It's wonderful that, when I plug a USB mouse into my laptop, it immediately works and doesn't play the "10-Mississippi" game that Windows XP plays, even on initial login. I understand that's because all of the drivers are pre-compiled and included into the kernel. Great. That said, if I decide to run the most up-to-date video drivers for a Linux-based computer and I don't hit up their respective repositories (something Ubuntu is getting better about, by the way), why must I recompile my drivers each and every time I install a kernel update? I don't mean a technical "why" - I know WHY, technically, and I even get the philosophical WHY, in that I understand it has to do with keeping open-source halfway open. That said, from a practical standpoint, I think that the first distro that comes up with a halfway decent driver API to link against is going to be the first desktop Linux distro that becomes ready for prime-time, because, from that point on, if someone decides they want bleeding-edge drivers, they won't be punished with repeated driver recompilations after every kernel update.
GUI - Don't get too carried away on having every single function in the GUI. Microsoft already does that and it's a royal pain in the ass. In fact, though I like some of the functions of Powershell, I think it's another fine example of Microsoft trying to do too much with their widgets. Sometimes less is more. If I have to keep hitting the '?' key to figure out which one of the 8,000+ verb-noun commands I want to use, that product is going to be less powerful than the one with only 500 commands to deal with that I actually have a halfway decent chance of remembering. That said, I'd still like to be able to do basic administration on a Linux machine without hitting up the command line. I'm okay if mastering administration requires shell-scripting skills, but let's keep the basic stuff... basic.
Regression Testing - The latest kernel update for Feisty Fawn ruins ACPI, making it where my laptop doesn't know whether it's on battery or not. More than anything, I'd like to throw more resources at preventing these sorts of shenanigans - heck, let's do this for Windows, too.
VIM configuration GUI - But only because I'm feeling ironic. :-)

Re:Might I Suggest...

[Chandon Seldon]

So... you've had to use the command line because you tried to install an OS on unsupported hardware. You get zero pity from me. You chose unsupported hardware when fully supported hardware has been readily available for years. You wouldn't complain about having to use the command line to install Mac OS X on the same system, or to install Windows on a PS3, why is Ubuntu any different?

Re:Might I Suggest...

[oatworm]

Sorry, thought of one more thing...

An enterprise-level accounting package. Think of any of Sage's or Best's products. Something that the bookkeepers could use that would run on Linux, is scalable past 20 employees (i.e. not QuickBooks), has some real inventory management (i.e. not QuickBooks), and is relatively stable. I already know about GnuCash, and it looks promising. Finish that and I can start pushing Linux for my customers instead of grudgingly throwing Windows boxen at them. Without that single bit of software, most of my customers won't touch Linux with a ten foot pole because they can't do business with it.

PS: I'm writing all of this from my Ubuntu-running laptop and I'm quite happy with it... even though it took me two days to get it working right. Heh.

Re:Might I Suggest...

[CopaceticOpus]

I agree that the CLI needs to keep an important role, but I think we should get to the point where the average desktop user can do most everything without ever having to use it.

I also have an idea to make the CLI more friendly. Even for those of us who aren't afraid of it, it can be a pain to have to recall switches or reread man pages to use a command that is only run occasionally. It might be nice to have a wizard for building CLI commands. For example, let's say I need to extract a .tar but I don't remember the specific details. I type "wizard tar" and see a text-based page like this:

=====
tar - Archiving program that stores and extracts files from an archive file known as a tarfile.

Would you like to:

c - Create a new tarfile
x - Extract files from an existing archive
t - List the contents of an archive
(...)

1 - Cancel
2 - See a list of other related programs
=====

Selecting one of those options might take you to a second wizard page with more options based on your first choice. When you're finished, it would dump you back to the CLI with your command entered but not yet executed. You could then run it as is, or edit it if needed. By seeing the command, if you ran wizard frequently you'd eventually learn it and not need the wizard.

(Yes, I know that suggesting a wizard for the CLI lowers my geek cred quite a bit. No, I don't think CLIppy would also be a good idea!)

Re:Might I Suggest...

[ozmanjusri]

I have a Compaq Presario V6000-series laptop with an AMD chipset. Apparently, it has some BIOS issues that require special flags in the initial boot loader screen on the Ubuntu install CD.

I have the same laptop (with the Nvidia 6150), and while I haven't tried Ubuntu on it, Sabayon installed and runs faultlessly.

I did have to change one menu setting for Compiz/Beryl, but apart from that everything worked out of the box, including the Broadcom wifi. The screenmode was correctly set to the LCD's native resolution, and I had the option of enabling 3D desktop acceleration during the install. It's a revelation after XP on the same machine - fast, stable and fun.

I like Ubuntu, and respect their goals, but I don't think it's the best distro for computers which are non-mainstream. Having said that, I suspect it's more likely that the AMD chipset was not available when Feisty was released, and if you try Gutsy, it'll all just work.

Re:Might I Suggest...

[shadanan]

And as a perfect example of this, take a look at MediaCoder http://mediacoder.sourceforge.net/ which is a GUI front end to mencoder. The thing works well, but has a substantial learning curve - might as well learn the mencoder command.

Re:Might I Suggest...

[entgod]

I always thought X was for xterms

Re:Might I Suggest...

[thaneross]

Currently, there are many "rough edges" in configuring a Linux distro from the GUI. There's a reason why you don't see dozens of excellent graphical configuration tools for Linux, and it's not because of some inherent superiority of the command line. It's because writing something like YAST is very difficult; far more so than necessary. The crux of the issue is the lack of consistency / machine readability in configuration files. The xorg.conf is the first example that comes to mind; it's quite frankly a mess even to edit by hand. Try making manual changes and watch the gui mangle the file as it fails to interpret syntax outside it's expected generated style.

Compare crontab, fstab, passwd, httpd.conf... pretty much anything in /etc/ and you'll see the problem. All of these certainly can be programmatically altered, but requires significant effort to create a clean bi-directional implementation that consistently works. This creates our present situation of often "hitting a brick wall" when trying to use the GUI and having no alternative but to do it by hand.

My preference would be files in XML. Now, say what you will about it, but XML is one of the few methods of storing data that is both easily human & machine readable. Without any additional work, you'd have countless tools that could edit and validate your config files. I'd take the trivial bloat over what we have now any day of the week.

Re:Might I Suggest...

[NickFortune]

This is a very conservative viewpoint, why can things not change? Why can't we have the best of both worlds, with both GUI configuration tools and text files?

That's a very good point. Trying to make your users work in a certain way is very much a Microsoft viewpoint. Actually, that's not strictly fair - it's a proprietary viewpoint. Software houses have limited resources, so they need to close choose a single way of doing things and concentrate their resources.

That's not something that's achievable in the FOSS world, nor (imho) is it desirable. People work on what interests them, which makes FOSS devs rather harder to reallocate than their professional counterparts. Tell a developer he can't work on Project X, and there's a fair chance his second option will be Nothing At All. Or maybe Project-X-On-BSD. There's no benefit to be gained. And who is to say that, in the long term, Project X may not prove to be the Next Big Thing.

I think this is the flaw behind all the posts bemoaning "imagine how much further ahead we'd be without all these distros/desktops/what-have you". It assumes that a developer working Gnome is one lost to KDE, and I don't think the maths works out like that, any more than the RIAA's self serving assumption that each downloaded MP3 equals a lost sale.

I think we should stop viewing the diversity of Linux software asa bug, and start looking at it as something to be celebrated. The question is: how to explain that to the someone steeped in the Windows mindset?

Re:Might I Suggest...

[Anonymous Coward]

If you are looking for a completely GUI drive *nix I would say OS X is your best bet (yes, I know you can use the CLI in OS X, but you never have to unless you so desire). This is somewhere where OS X is clearly ahead of Linux. It should be possible to completely drive Linux from the GUI and still use the CLI but not have to rely on it. Unless you so desire.

Re:Might I Suggest...

[Kjella]

Look, flat text files are just that - flat text files. I am CLI-aware but I don't think text files are particularly good - they have no validation of either parameter names, values or even types of value. They have no easily navigateable structure, nor any user friendly UI elements. Typical example:

# Controls max number of connections
connectionsMax = 10
# Enable SSL
ssl = true

I think it should probably have been defined somewhere in XML

<field name = "connectionsMax">
  <use>required</use>
  <type>numeric</type>
  <validation type = "range">
    <min>0</min>
    <max>1000</max>
  </validation>
  <help>
    <short>"Controls max number of connections"</short>
  </help>
</field>
<field name = "ssl">
  <use>optional</use>
  <type>boolean</type>
  <default>false</default>
  <help>
    <short>"Enable SSL"</short>
    <long>"SSL enables secure connections protected by encryption to and from your server. [Blahblahblah]"</long>
  </help>
</field>

Then, to get a plain txt you'd do something like configedit --cli, and it creates a virtual text file that it'll validate on save against the XML. Or an export/import to let you edit in vi/emacs/flamewar of the day. Or configedit --tui to get a text user interface like the debian installer. Or configedit --x11, configedit --gnome or configedit --kde to get a graphical edit mode where connectionsMax is a spinbox and enable SSL is a checkbox.

I think that would be far superior to handling text files directly, and where you can add a *lot* more dependency rules, ui hints etc. which would make the GUI smarter without clogging up the text file for a CLI user. The only thing is that it must be really easy for developers, like maybe even so easy that if you add a new variable it's automatically added as a string variable with no validation. This, along with default UI mode can probably be stored in some sort of preferences. That way the user gets an error: "Field "ssl": "flase" is not a valid value. Valid values are "true", "false" while it doesn't get in the way of the developer.

Re:Might I Suggest...

[Kjella]

Oh yeah, and doh... a BSD library that'll parse and extract these values for you so you don't need to write your own XML parser.

Re:Might I Suggest...

[Anonymous Coward]

Parent is overrated! This sounds like a cop-out to me... if its a feature that the user is expected to touch, it should be accessible in some sane way. (editing registry keys does not count, nor does finding obscure config files - they should not need to understand how a filesystem works to change settings in their app) Just because thats "how most programs work" doesn't mean it isn't a bit sloppy. With some thought you can design a GUI for many things.

Re:Might I Suggest...

[scwizard]

In fact I'd spend a lot of the money on getting everyone (or as many projects as I could) to agree to a configuration file format that could easily be interpreted by an application. A one-size-fits-all library could be written to get the settings from file into memory and back again

Qt has such a feature last time I checked. I don't understand why it's not used for more linux apps...

Re:Might I Suggest...

[CoughDropAddict]

I don't think you realize how ungodly complex such a thing would need to be to actually replace common configuration files. Just look at how the absolutely most trivial example of configuration ballooned into 20 lines of XML. Configuration gets a lot more complex than that. If you don't believe me, just look at the enormous complexity and verbosity of XML Schema, which is trying to solve a problem of similar complexity.

Re:Might I Suggest...

[mpeg4codec]

So does GNOME. It's called GConf and it stores data hierarchically not too much unlike the Windows registry. Much like an LDAP server, the frontend has a unified API, but the backend can be any data store. The default is XML files in /etc/gconf, but nothing prevents you from storing it in a legit RDBMS. There are libraries for accessing it in most languages, as well.

Why didn't it catch on? Most [all?] GNOME projects use it now, and a handful of others do. I'm sure the Qt mirror-mirror version of GConf is much the same for KDE apps.

Why wouldn't I use it? Why bother with forcing the user to have another dependency when a dotfile is more than good enough for a standalone console app. In a more perfect world, where I could count on the user having GConf or Qt's version, maybe it'd be different. But for now, I'm afraid I have to stick to a simpler text file configuration for most things, which unfortunately means one parser per app.

Re:Might I Suggest...

[tepples]

Again, [configuration of 802.11b/g networking] is probably as much a "hardware support" issue as anything; ideally, the drivers would've already been in Linux and we wouldn't have these problems. But would the U.S. Federal Communications Commission (or foreign counterparts) allow binary blobs used by wireless drivers to become free software?

Re:Might I Suggest...

[tepples]

So... you've had to use the command line because you tried to install an OS on unsupported hardware. You get zero pity from me. You chose unsupported hardware when fully supported hardware has been readily available for years. When unsupported hardware plus a Windows license costs less than supported hardware, what would one choose? When donated unsupported hardware (such as a birthday gift or gift to non-profit organization) costs $0, what would one choose? When unsupported hardware that was paid for before one considered switching to another operating system costs $0, what would one choose?

Re:Might I Suggest...

[Anonymous Coward]

I don't want Qt if I'd just need to parse a config file.
That said, I am quite happy with libxml2 and a limited shell parser (VAR="def" style format)

Re:Might I Suggest...

[tepples]

My preference would be files in XML. Now, say what you will about it, but XML is one of the few methods of storing data that is both easily human & machine readable. So is a delimited table such as fstab, and so is INI. What we really need isn't standardization on XML as much as a schema: a solid definition of each file's syntax and some level of its semantics. We'd need a schema with XML anyway; otherwise, you're just editing a DOM tree and likely making it inconsistent.

Re:Might I Suggest...

[Anonymous Coward]

"part of what makes Linux so fast, flexible and customizable is access to virtually every setting from a text editor. This is not something that needs to be changed, instead change your mindset that this is not Windows."

Why don't you change your own mindset and figure that Having a GUI developed ad advanced enough to be used in place of the CLI, on a system that has a CLI just as advanced is a very, very good thing. The idea is that resorting to the CLI becomes a matter of personal preference, not obligation, as it should be.

"(yes, I know you can use the CLI in OS X, but you never have to unless you so desire)."

That's the idea, Apple got it right. OS X's CLI possesses every bit of power that you'd expect from the CLI on a BSD system, except they've put considerable work into the GUI as well, and as such, the use of CLI the majority of the time is entirely a matter of personal preference.

Even Windows has begun going down this road, with the introduction of PowerShell, which is an optional component for XP/2k3/Vista, but will be integrated into server 2008 and Vienna. Until now Windows and Linux were opposites in that regard, the former with a flimsy CLI tacked onto a the GUI (and now with a rather powerful CLI tacked onto it), and the later with a relatively flimsy GUI tacked onto the CLI. What's with the resistance against bringing both GUI and CLI on equal footing?

Re:Might I Suggest...

[dbodner]

that anyone who thinks that CLI usage is not a feature of Linux think again? This topic is 12 minutes old and three post have already suggested we bury the command line; part of what makes Linux so fast, flexible and customizable is access to virtually every setting from a text editor. This is not something that needs to be changed, instead change your mindset that this is not Windows. You should always have the ability to drop to a command line. Editing files by hand should always be an OPTION. However, being able to do all day to day tasks via an interface should be an obtainable goal, as long as it doesn't limit the functionality and ability to drop to a command line. You should always be able to drop to a command line, but it should very rarely be necessary.

Re:Might I Suggest...

[grumbel]

The config file (which really isn't meant to be user editable) gets more complex, the process editing it however become much easier, since a proper editor doesn't just show "name = value", but can also show the documentation, possible values, validate your input and all that stuff. Automating stuff also becomes a "cfgtool --set name value thing and not a horror of grep/sed/ed hacks.

Re:Might I Suggest...

[thaneross]

Yep. I agree. A schema is what I was trying to get at.

Re:Might I Suggest...

[Planesdragon]

This topic is 12 minutes old and three post have already suggested we bury the command line No one suggested that you make the command line inaccessible. What they suggested was that you stop making the Linux GUI as crippled as the Windows 2000 command line.

Sure, the CLI is great. (I love the command-line in MySQL). But that's no excuse for making a half-assed GUI. Either make the GUI at least as expansive and full-featured as Windows, or drop the entire "config GUI" and just load up the darn text file.

Re:Might I Suggest...

[Anonymous Coward]

Even Windows with WSH and the Registry. Hell,

Now when I look for advice on registry hell I will come to this page. Thanks a lot, pal.

Re:Might I Suggest...

[Chandon Seldon]

When unsupported hardware plus a Windows license costs less than supported hardware, what would one choose?

I would chose the system that runs the software that I want to use. Every time. Unsupported hardware is completely useless, so there's no reason for me to even consider buying it.

Re:Might I Suggest...

[Chandon Seldon]

Is there any reason that any of the features that you describe are incompatible with a flat text config file that can still be hand edited? Unix systems are pretty damn good at dealing with text files - throwing away all those capabilities for no actual gains would be a mistake.

Re:Might I Suggest...

[Chandon Seldon]

When donated unsupported hardware (such as a birthday gift or gift to non-profit organization) costs $0, what would one choose? When unsupported hardware that was paid for before one considered switching to another operating system costs $0, what would one choose?

In this case, the cost of seriously switching to another operating system includes the cost of any necessary hardware alterations. When you have an unsupported wireless card, running Ubuntu costs $30. In the worst case, switching to Ubuntu may have a similar monetary cost to switching to a Mac - you may have to buy a whole new computer with Ubuntu pre-installed.

Sure, it would be nice if Ubuntu supported every piece of hardware in the world - but that's simply not possible. Not even Windows can do that. Luckily, we don't actually need to support every piece of hardware - we just need to be able to get hardware that is supported, and that's not a problem.

Re:Might I Suggest...

[anon mouse-cow-aard]

That sucks big time. heres why:

First. It mixes in meta-data with data. Metadata should be somewhere else, because, well obviously, if I have 1500 configuration files for slight variations of the same thing, then Im going to have 1499 pointless copies of metadata. keep the validation information and real data separate, dammit.

Second. XML parsing is 10 to 100 times slower than parsing a text file. You have clearly never written a text file parser and an XML parser. Using a reasonable language like python, text file parsing is trivial.
XML parsing is simply not to be attempted without a library, its that painful. When you are going to read 10,000 config files, it matters. If you adopt this scheme widely, you will end up reading that large number of files. XML structure doesnt add anything useful. It is not easier to read than a standard text file. It is not more concise. It will be expensive.

Third. when upgrading your software, all the metadata can change, If you want to upgrade and see what the differences are... good luck! Comments, upper bounds, etc... will all change. Keeping meta data separate from configuration, allows you to diff the meta data separately from the configuration info.

Fourth. I18n and i10n. If you wanted to do your approach thoroughly, you would have 150 translations of all the textual meta data in the same file. That file is going to be one big sucker. Put it in a separate file? fine. do it consistently instead of treating one language as special.

Fifth. If you have All this XML meta data in a separate file and the text file stays in exactly the same format, you still get all the goodness, and you havent lost anything. The only requirement for this is to have a couple of text file parsers. There are not that many flavours of such text file formats, so some XML tag could identify the grammar approach. The text file, concise, to the point, easily parsed by experienced humans, remains untouched.

Sixth: you completely miss the point of text files. They are a user interface. To me, a text file with a simple, easy to understand grammar, together with an application that identifies format errors is the easiest, most flexible, most effective way to manage application configuration. I regularly strip out comments from configuration files because to me, 99% of the comments are noise. If I wanted documentation on the format, I would look in a man page. Configuration information is a list of settings.
The point of text files is that they are human editable, human auditable, diffable, revision controllable. When you make the file format too noisy for a human to parse, you lose the entire benefit of a text file in the first place. It might as well be binary, because no human can safely edit the file anymore (Its too !@%#$@! complicated to edit the file and get all the tags and metadata correct.) The user of a text configuration file is a human. Making him type, or understand five times or ten times as much text to configure the same information is deeply, irretrievably, wrong.

Re:Might I Suggest...

[Blackknight]

Try SQLedger.

Re:Might I Suggest...

[grumbel]

How to write the config settings themselves down is really a secondary issue, the important part is to have all the documentation, default values, types and stuff in a machine readable form (i.e. not just man-page or a comment in the config file itself) so that you can validate the config files after or even while editing. GConf already does that (/etc/gconf/schemas/) so its an already partly solved problem, except that there still a lot of tools around that have their own at hoc format, which is kind of part of the crux, Gnome and KDE do their GUI stuff, but a lot of their improvements have a hard time to get used by non-GUI apps that could still profit from them.

The main issue with 'flat text config' is simply that it isn't a well defined format, every app has its own little format, some might even look the same, but behave differently. So its impossible to build a editor that handles them all properly.

Re:Might I Suggest...

[dodobh]

Cool, now can you you imagine FINDING the correct line in large files? Do you realise HOW BIG XML files get? The simplicity of text files lies in the fact that they have a minimum of clutter for me to waste my time. Especially when I am working of dialup over a cellphone.

Hell, GNOME uses XML config files. It's next to impossible to figure out the correct magic value to set some field to to get it to do what I want.

You do realise that all that extraneous information should be recorded in the manual page for that file in section 5? Alternatively, applications with lots of options can ship their documentation in alternative formats as well, like HTML or Texinfo, or docbook, or ...

Re:Might I Suggest...

[poot_rootbeer]

If you are looking for a completely GUI drive *nix I would say OS X is your best bet

A good marketing manager should never say "Our product comes pretty close to meeting your needs, but we're not willing to go that last mile, so I think you should go to our competitor instead." You may not work as a Marketing Manager for Linux, but as an informal advocate, your stance should be the same.

What Linux needs GUI usability expertise. And that doesn't mean support for transparent widgets or yet another window manager -- it means somebody who understands the psychological and sociological aspects of human-computer interaction, and can put a consistent and accessible veneer on top of one of the most technically refined operating systems there is.

Apple accomplished this when they developed OS X, because they had HCI experts on staff. What is holding the Linux community back?

Re:Might I Suggest...

[Reziac]

[trots out same suggestion I've made every time this topic comes up; mods self -1, Redundant]

Why not an interface that can both show you the available config options in a GUI, *and* (in an adjacent subwindow) display the actual textfile, so you can SEE what you really did by clicking the GUI's checkbox?

Most HTML editors already do effectively the same thing, by having both WYSIWYG and raw-HTML modes, and you can use whichever interface you prefer. Also lets you learn by seeing what your changes actually DID.

Re:Might I Suggest...

[oatworm]

I will freely point out that I didn't research how well my laptop might be supported by Ubuntu when I got it. In fact, it really wasn't a big deal whether Ubuntu worked right or not; I got the laptop to have something I could play with and use on the road, and thought it'd be fun to throw Ubuntu on there instead of relying on XP. I even tried PC-BSD on it, which is how I discovered the hard way that my laptop isn't just unsupported on BSD - it's impossible (or at least more difficult than I'm willing to overcome) to boot due to Compaq's flaky UltraDMA implementation and the laptop's BIOS's complete inability to disable it. At least Ubuntu boots.

That said, the question asked was what I think Linux should focus on. I think supporting hardware would be a great start. If I wanted to be limited in my hardware selections because my OS won't work on anything else, I'd get a Mac; it seems Linux runs best on Intel hardware anyways, so what's the difference? That said, I understand that, realistically, there's no way for Linux to support every single chipset or vendor in the world, especially since there are many vendors that don't want to be supported on Linux, so I'm willing to settle for a nicer way of dealing with unsupported hardware. I'm still happy with Linux, my laptop is still running Ubuntu, and I'm very happy with that... I just think there's room for improvement, that's all.

One universal install method...

[mikesd81]

That just freaking works. I never understood why every distro can't just use the same install method. Whatever it may be, rpm, apt, yast whatever. And I don't mean the 3 step make install method. Wouldn't it be great to go grab a package from freshmeat or sourceforge and...oh look that's the package type I need....

Re:One universal install method...

[tajmorton]

Wouldn't it be great to go grab a package from freshmeat or sourceforge and...oh look that's the package type I need....

See the Autopackage project, and there's a fairly large amount of software packaged using it. It works, but distros don't like it because they're afraid that the developers packages might mess up your system, so they refuse to support it. *shrug*

Re:One universal install method...

[garcia]

I never understood why every distro can't just use the same install method.

Because none of them are setup the same way and honestly, I don't want them to be? apt-get update ; apt-get install foo. Looks like a two step method where I don't have to think all that much and I still feel like I have control over my system.

Let's not fuck up the way things are just because some believe it would be better. Choice is a big part of the draw to Linux, why make it more like Windows when a lot of people want it to be anything but?

Re:One universal install method...

[Schraegstrichpunkt]

The problem is that "Linux"[1] is not a single operating system, and there are subtle binary incompatibilities between various distros. Use Debian for a while, then switch to something like Mandriva. You'll get the idea pretty quickly.

-

[1] or "GNU/Linux"---it doesn't matter here

Re:One universal install method...

[wizardforce]

I never understood why every distro can't just use the same install method.

what you want is windows not linux. windows does everything in a cookie cutter sort of way and look where it got them... the current state of security/flaws that we see in windows is a consequence of that sterile environment. we need a diverse system that can survive in a way windows can not. things are done differently not "incorrectly." Now if the different flavors of Linux made each installation method simpler to carry out while maintaining a layer of security and diversity then all is well. Personally I would rather have a secure, efficinet OS that has 5-10 seperate methods of installing things than a windows clone that only hs one way and all the inherant flaws associated with it. But really what you have in mind should be the focus of a small part of the community that has its own niche like the others. some of us want total ease of installation while others want customization and power, so we can still retain the choice to move to the new system if we want or continue down another route.

Re:One universal install method...

[Anonymous Coward]

How about a secure, efficient OS that has 1-2 seperate methods of installing things, like OS X?

Re:One universal install method...

[644bd346996]

Are you saying that the architecture of the packaging system has an effect on the overall security of the system? That isn't really possible unless the packaging system in question is so byzantine that it becomes difficult to repackage application security updates in a timely fashion.

From a security standpoint, a universal package format is a wonderful ideal. As soon as a fix is available, the application developers can release a new (signed) binary that can be tested by a very large audience. Once the update has been tested, it can be applied by any user, without the extra delay from the distro maintainers.

You also seem to think that ease of use and customizability are mutually exclusive. This is an extremely pessimistic and wrong viewpoint, and should never be applied to software development.

Re:One universal install method...

[644bd346996]

Is it really that much harder to type yum update; yum install foo?

Let's face it: All software has dependencies. Those dependencies (and optional dependencies) are determined by the application developers, not the distro packagers. Almost all packaging systems out there are simply cases of re-inventing the wheel with slight, trivial differences. There are no technological barriers to implementing a standardized packaging system. It would make things a lot easier for maintainers (both the original authors and the distro maintainers), and would make things a much easier for commercial software developers that have to do all their own packaging. And at what cost to the user? None. You can always install a front end you are comfortable with.

Re:One universal install method...

[wizardforce]

what I am saying is that we need to have diversity in regard to operating systems, it is a bad idea to reduce diversity in a system like this... if we can get away with a certain level of diversity while making things simple so be it. I just can't help but hold biology as an example what not to do- when species become less diverse they are prone to disease just like OSes tend yto be when you take for example Windows- very low diversity and a clear disregard for the customization we are used to in Linux. As for the gap between customization and ease of use, I am not saying it is impossible to have the two together but I am saying that the way software development has evolved tends to reinforce the viewpoint that a customizable system is going to have more complexity at some level GUI or not that a less customizable system will tend to have. Linux does well because you can really screw with things at a basic level and still have a nice GUI too and it should continue to have the best of both worlds.

Re:One universal install method...

[EvanED]

That's fine.

That's the elegance of splitting implementation from interface that's at the core of why programming works. You can have one set of tools -- or even just one set of CLI and GUI interfaces -- that will take care of that stuff behind your back.

Then if you move from Red Hat to Ubuntu you don't need to relearn how to install programs just because Red Hat uses Yum and Ubuntu uses apt-get (or Aptitude or Synaptic or whatever you want to consider it using), because what the user would do to install an application on both distros would be the same, even though what actually goes on behind the scenes would differ.

Re:One universal install method...

[EvanED]

What? You're saying that Windows is insecure because it provides cookie-cutter ways of doing things?

Windows's insecurity comes from some fundamental security policy decisions, such as run as admin, and some outright coding bugs like buffer overflows.

The security mechanisms Windows provides are fine, and in fact rather more expressive than anything you get on a traditional Unix-like system.

Re:One universal install method...

[wizardforce]

what I'm saying is that operating systems should have soem diversity in how they work much like in healthy ecosystems in biology. when you start restricting diversity in a system it becomes a hell of a lot more vulnerable. two main examples of this: the lack of diversity in crops specifically bannanas; there is one major breed which is now being wiped out by a fungus- had there been any diversity in the species it wouldn't be so much a of a problem. this isn't the first time it happened either. the breed we use today is entirely different from the strain that was destroyed in the past for a similar reason. the other example is penicillin; this antibiotic has worked for MILLIONS of years fighting bacteria just fine for the fungus that makes it. that is until we started using it all across the planet; we made it a common antibiotic and now we have created a real mess with antibiotic resistance within a matter of decades. it is unimaginably better for security to have a diverse set of OSes rather than a small number of unique ways of doing things.

Re:One universal install method...

[chdig]

Let's face it: All software has dependencies. When I install software on Windows, it just works (95% of the time). You're right for Linux, all software has dependencies, while with Windows there aren't any. Maybe that's another place where Windows has a system that makes more sense to people.

Linux installs are a nightmare. Compare LAMP setup on Ubuntu to WAMP on windows, and WAMP wins in ease of setup. By a long shot!

Re:One universal install method...

[wellingj]

What you couldn't handle the Ubuntu Server CD? Begone Elmer FUD! LAMP stack was installed on my circa '02 netvista in all of 20 minutes. How was that a nightmare? The real nightmare comes when you have to secure your Windows box. Have fun with that.

Re:One universal install method...

[BKX]

The reason that Windows appears to not have dependencies is because all the dependencies are included in the package. This works on Windows because all software gets it's own directory and can put whatever junk it wants in it. Under Linux, all software (generally, I know there's /opt) resides in the same standard location, so that if two packages have the same dependency they'd end up fucking each other up. We don't want that. That's why distros have their own packaging systems.

Furthermore, in the end its much easier to have your distro handle the package management anyway. Have you ever tried to keep up with software updates on a Windows system? Probably not. You probably just stick with the same version while people exploit bugs that were fixed years ago, and random hackers fight over who owns your box. Under Linux, it's a one or two command process to update EVERY package on your entire system automatically. Windows blows it.

Re:One universal install method...

[jon_anderson_ca]

Each distro likes their own package format, but they're all pretty much the same now. Whether you type "apt-get/aptitude", "emerge" or "yum", pretty much every distro involves typing "something install the-package".

Most distros have some kind of GUI that's accessible with a link like "Add/Remove Software", too.

Re:One universal install method...

[Eli Gottlieb]

I know it's been said elsewhere but...

APPDIRS! APPDIRS! APPDIRS! Also known as "bundles" under OSX, they are the closest I've ever seen anything come to an ideal packaging solution. You have directories for programs sorted by category: /Applications, /Libraries, /Daemons, etc., and each "bundle" is a directory in one of those hierarchies with a name like (gasp at the sheer simplicity) "bash-3.1" (you can add a symlink called bash if you want, too). That bundle/appdir contains every system-wide file owned by the package, just like under Windoze. User-specific configuration files go somewhere in users' home directories, just like under Linux.

To install software, put its bundle in the appropriate hierarchy. To uninstall software, remove a bundle and remove personal settings.

Admittedly, tools like apt or emerge that can read and display package repositories, build custom packages (ports and portage), or resolve dependencies still have great use, but in an appdir system they rest upon a much simpler, more solid foundation.

Re:One universal install method...

[lachlan76]

The complaints seem fairly legitimate to me.

Digital Video Editing

[RockHorn]

Digital Video Editing - hands down.

We have converted to Ubuntu for all computers in our house now, and the only reason I'm looking elsewhere is for a stable and *easy to use* digital video editing solution. Surprising really, given the amount of video people record lately.

The hopeful projects are Kdenlive (still seems immature), and Cinelerra (can't for the life of me get it to install/run). Kino isn't an option, as you can't include still images.

With DV editing, Linux would indeed be desktop ready.

Re:Digital Video Editing

[Salsaman]

http://www.getdeb.net/app.php?name=LiVES

WINE

[jstomel]

Admit it, wine sucks and there are lots of programs that will never be ported. I want wine to be integrated and almost invisible, like the Classic interface in OSX.

Re:WINE

[Paradigm_Complex]

Wine's awesome and I'm often surprised at how well any given application works. Double-clicking on a windows executable runs it like any other program; it's easy for someone not to even realize that it isn't native. True, WINE lags behind the ever changing Windows landscape, but if you don't need the absolute newest programs out there (considering this is compared to "classic" a little age seems acceptable), you'll be fine.

Re:WINE

[rastilin]

It's not as good as you think. Psychonauts only runs with the last two versions, since they finally implemented multi threaded 3d and Far Cry doesn't appear to run without difficulty even now. It's true that you don't need the latest applications, but odds are you don't technically need lots of things on your system possibly including a computer. I still want things to work regardless if other people think they don't matter and even after 10 years of development wine still doesn't cut it. Yes, I know it's Microsoft's fault for designing an impenetrable system but acknowledging that still doesn't get Bioshock working.

Re:WINE

[Paradigm_Complex]

I don't know about Psychonauts, but Far Cry is DX9 (that was its call to fame, the game play was quite disappointing). WINE has DX8 down fairly well, things like Half Life 2 run fine. As I said, WINE lags behind a bit... in the case of Direct X, it seems just over a generation behind. Given time, it'll tackle DX9 as it currently is with DX8.

Re:WINE

[Verte]

No, I'd have to agree with the GP. An application I used to use at my last job would crash when docking toolbars, opening dialog boxes in some circumstances, etc. There are a lot of FIXMEs and missing functionality throughout WINE. Sure, it works, sometimes, and with some fiddling you can get applications to mostly work, but I'd agree that it does need a lot of work itself.

Re:WINE

[Anonymous Coward]

I believe a much more compatible and stable WINE is essential if Linux is ever to have a large user base.
For the office it's that ancient custom business application that will never be ported (if the source even still exists).
For home use it's their favorite old windows game.

It seems like far too much effort is being put into the project to prove itself by making it run the newest software releases when it's really the older applications that matter so much more. At least when it comes to more casual users who consider using Linux as their primary system for anything beyond basic web & email.

I suppose it's just a matter of developer interest, the new stuff is always more exciting to tinker with than the often poorly written older applications.

Re:WINE

[rastilin]

Yes but that's exactly the problem. In a generation there'll be new games that I'll want to play. The problem doesn't end there though, because there are still bugs in games that work almost perfectly.Not that virtualization will be all that better but a great many problems will be avoided.

Re:WINE

[MostAwesomeDude]

Wine is already one of the highest-priority, work-intensive projects in the GNU/Linux world. They release every two weeks, and each release is a significant improvement over the last. While I can't argue that Wine is perfect (it's not;) I think that more work needs to be done in the kernel. There's dozens of regressions on amd64 and ppc that aren't caught because they work on x86. Also, we need an updated DRM interface for ATI Radeons and nVidia Geforces. (DRM in this case is Direct Rendering Manager, the kernel code that makes accelerated everything possible.)

Re:WINE

[buddyglass]

This is actually one of the apps I had in mind when I posted the question. I was trying to think of a short list of apps and/or kernel areas that, if they were improved, would make me much more likely to run Linux as my primary desktop. Wine is one. An "ideal" Wine would let me run necessary apps without having to suffer the overhead of a virtualized OS instance. Each wine release seems to fix a ton of bugs, and yet many of the highest profile apps still don't work without massive tweaking.

The other app I had in mind is Firefox. If I were donating resources to Firefox, I'd task them with making it bulletproof from a performance / stability point of view. No more having to close firefox periodically because memory leaks. Make it more performant than Opera.

Re:WINE

[mpeg4codec]

Apple does have the home turf advantage of having the source to OS 9. Not to mention, Apple isn't afraid of deprecating and even making old software obsolete. Versions of OS X running on Intel Macs do not have support for Classic applications, and most of their customers simply don't care. Microsoft is stuck supporting legacy code written for 16 bit versions of Windows right next to the latest and flashiest modern apps.

It's a different business model with a different customer base. Unfortunately, those who are on the sidelines trying to implement a mostly undocumented API are stuck with what they can reverse engineer. That leads to a project that is admittedly less polished than one where the source of the API being implemented is available.

Re:WINE

[MrCrassic]

I'll second this. If I needed full Windows compatibility (as I seem to need with Outlook 2003), I would run a VM with some low-resource Windows OS on it (I'm using Server 2003 on it) and let that handle it. For anything else, CrossOver Office (a WINE fork) works flawlessly and without much hard work involved. Nothing beats using Word without using Windows!

UI needs a lot more consistency

[Whuffo]

One of the GNU / Linux weak points is the wide variety of user interface implementations. Simple and common functions should be accessed in the same way in every application. Look how many different keystroke combinations / menu selections are used to exit a program - sure, us nerds can keep them all straight without any trouble but what about the unwashed masses?

Why is it that all the developers seem to be able to code to a standard API - but they can't even come close to agreement on the way a program is operated? Maybe it's time to create a UI standard for Linux apps?

This would go a long way towards making Linux the favored choice for desktop machines. Ease of use is a great way to unseat the dominant OS; it's not exactly easy to use and it's very possible to beat them at this game.

Re:UI needs a lot more consistency

[Anonymous Coward]

Actually, Windows is ultra-easy to use. (And I use the "ultra"-prefix only for special occasions.)

Re:UI needs a lot more consistency

[Anonymous Coward]

I believe nerds ARE the unwashed masses. The rest of us take showers regularly.

Re:UI needs a lot more consistency

[DamnStupidElf]

Why is it that all the developers seem to be able to code to a standard API - but they can't even come close to agreement on the way a program is operated? Maybe it's time to create a UI standard for Linux apps?

Probably because there are no universal UI guidelines that fit in every case, and no one knows of the "best" way to design a UI. There are almost certainly better UI models that will be found in the future, and part of open source development is trying to discover those models. Modal dialogs with "Yes, No, Cancel" may not be part of the optimal solution (it's hard to imagine them being the best for any task, really), so exploration into other representations and interaction models is a good thing.

For one thing, modal dialogs are essentially points in the program flow where the user is forced to make a decision. In many cases, the decision can be made automatically or simply removed from the program flow if the correct UI model is used. All the dialogs warning about saving documents before closing programs are probably the worst thing to happen in UIs since their creation. The proper solution is to save and version everything the user does, and never lose work (to the maximum extent possible) when the application is closed or crashes. That requires support from the filesystem and a UI design that emphasizes the versioned nature of files and data. Forcing users to "log on" and "log off" and closing all their applications and opening them up again is another genuine UI mistake that's a holdover from the days when memory was so tiny that overlays were popular and it was simply a necessity of operating a computer. There is no reason that modern operating systems should not present a persistent interface to the user. Hibernation is the closest thing to persistence that's currently available, with suspension a close second, but neither are truly persistent. They still rely on the old model of individually managing each piece of data.

A Common GUI and a Linus-like Dictator for same

[Anonymous Coward]

Many development projects benefit from competing teams and an open marketplace of ideas. UI design is not one of those areas. Its far, far, more important to be conistent than it is to be "optimal" for any given task/operation.

With Linux right now, there's one kernel, because Linus runs that end of the house as a dictator, and it works.

The GUI world is a nightmare of competing standards, mish-mashed interface concepts, and generally poor design. Apple has proved that you *can* put a beautiful, conistent UI on top of a unix kernel. They did it by ruthlessly enforcing standards, and they did it by having one guy ultimately calling the shots.

It's time for the Linux world to do the same thing.

Re:A Common GUI and a Linus-like Dictator for same

[Malekin]

I don't buy your assertion that Apple ruthlessly enforces its UI standards. In fact, Apple Inc. seems quite intent on throwing its own human interface guidelines to the wind. The do provide an API and toolset that makes it easy to create a passable interface, and they do give out industry awards, but I'd be very surprised to hear of an example of them ever telling a developer their interface wasn't up to scratch.

The consistency of user interface across the software running under Mac OS X is the result of a developer culture that cares about these things, and a user base that expects consistently high UI standards. Changing Linux developer and user culture to implement the same sort of thing would be quite the challenge - culture has a massive amount of inertia, and rarely responds to orders.

Re:A Common GUI and a Linus-like Dictator for same

[bky1701]

So what GUI do we pick? This is always the question when it comes to "Can't we just agree on something" arguments. And really, theres only 1 kernel only because only that kernel is Linux. Compare the BSDs, that do very similar things to Linux, and you'll see that the kernel and GUI state of affairs aren't much different.

Re:A Common GUI and a Linus-like Dictator for same

[EvanED]

Just because it's a hard question doesn't mean it doesn't need an answer.

Re:A Common GUI and a Linus-like Dictator for same

[bky1701]

And I am sure as soon as we have a global language, someone will have that answer.

Re:A Common GUI and a Linus-like Dictator for same

[EvanED]

Hey, I didn't say that I actually expected an answer. ;-)

But I think that while there isn't one, the inconsistencies between especially KDE and Gnome apps will hinder Linux's adoption. I don't think it will actually act as a blocker to adoption for almost anyone, but it will slow it, because it adds a reason to not switch.

Re:A Common GUI and a Linus-like Dictator for same

[Repossessed]

So do you want it to look like GNOME or KDE? Either one you pick you'll be outnumbered by the people that want the other, or something else.

Re:A Common GUI and a Linus-like Dictator for same

[dbIII]

CDE showed that nobody really wanted a common desktop environment.

Re:A Common GUI and a Linus-like Dictator for same

[bky1701]

Well, I do have to agree some things maybe nice- a common basic API interface between code and gnome/KDE/xfce- leaving more complex parts to be handled on a per-desktop basis (maybe not at all, depending on the program). I don't really program apps like that, but as far as I know such does not exist.

Different apps isn't really a problem for most people (you have to remember, most people just use what comes with the computer) but apps acting strange is. There are some cases where an app is only for 1 (or worse, just x directly) and odd things happen when used on another. Some distros are better than others in this area- no problems with Ubuntu, but I had missing taskbar icons and some strange font problems on fedora. Gnome->KDE also seems to work better than KDE->Gnome.

While different GUIs are here to stay and for good reason, we could still treat the negative symptoms better than we are now.

Re:A Common GUI and a Linus-like Dictator for same

[WNight]

KDE and Gnome are filling different niches. To you they're only eye-candy. So use a KDE theme in Gnome or vice-versa. But behind the scenes they're coded in a different fashion to meet different goals.

How many systems are you recommending people switch between? My mom's been using Gnome since I installed Ubuntu, KDE might be a bit confusing out of the blue, so I *don't* have her system configured to randomly switch GUIs... Vista or OS X would be a bit of a surprise too.

Besides, if you think the Mac is consistent, look at iTunes. Most of Apple's UI guidelines tossed out the window. Quicktime player for Windows was another laughably bad example.

It's funny how to authoritarians everything seems like a good excuse for a dictator to tell everyone what to do.

Re:A Common GUI and a Linus-like Dictator for same

[WNight]

Which is better, Kittens or Apples?

Which is better, a stripped down Linux that runs on my PDA, or a full-featured kitchen-sink distro that comes with everything pre-selected?

That's not just a hard question, it's a trick question. My PDA won't run with a desktop distro, my desktop would be crippled with a PDA distro. Neither one should be replaced with the other.

That's precisely why Windows is a lame server. Try turning the GUI off for more speed/ram.

Re:A Common GUI and a Linus-like Dictator for same

[EvanED]

Yes, but that's a different question in some sense.

Why do we have Gnome and KDE? (Actually I should say GTK and Qt...) They both do about the same thing, they both run and don't run on about the same systems. My understanding is that we have both originally because of licensing issues, but we have both now because they think that the systems should go in different directions, and also because of developer egos.

But as a user, I don't care about that. What I see is that Gaim doesn't fit in with the rest of my KDE apps, because it uses GTK widgets instead of Qt.

This is the sort of inconsistency that I'm talking about, not that you need systems tuned differently for different applications.

Re:A Common GUI and a Linus-like Dictator for same

[WNight]

It seems like the answer for that is a KDE-like Gnome theme, installed by default when you install KDE. Otherwise the different UI seems to suggest that the app is going to be different.

But people act like the technologies involved (Qt, etc) need to be pruned. IMHO they just need to be tweaked a little bit to look alike, where they'll be used together.

One Word

[infonography]

Ninjas

Face it its an excellent way to deal with all sorts of troublesome issues. If you don't believe me go on Youtube and ask a Ninja.

Re:One Word

[lordtoran]

I have one of them working here as a spam assassin. Couldn't live without.

Unfortunatly...

[Anonymous Coward]

....if I had to pick, I'd have them spend their time reverse engineering drivers for popular hardware from manufacturers who don't publish their API's. Like it or not, significantly more hardware "just works" on Windows that doesn't work on Linux. And if (and maybe it's still an if) Linux wants to be a real competitor on the desktop market, we need to realize unpacking hard-to-find tarballs and tweaking your system configuration isn't viable.

Re:Unfortunatly...

[SEWilco]

Before they start reverse-engineering drivers, first clearly document the manufacturer's failure. For years I've been avoiding giving such manufacturers any of my money. I'm quite willing to help them fail faster in the marketplace.

Re:Unfortunatly...

[Anonymous Coward]

A good idea, and good for you on this. But misses the point.

The linux community is not large enough, and currently never will be large enough, to dictate the hardware marketplace. Vocal minority, sure. But most computer buyers/users get "whatever Dell/HP/Gateway/etc" installs in the machine. They may know the more expensive graphics card will play games better, but darned if most of them have a bloody clue where they'll get drivers.

The point being made is that if Linux wants to compete for Microsoft's territory on the desktop, it needs to do at least as good a job supporting whatever hardware is out there. Not just the ones from enlightened manufacturers.

Re:Unfortunatly...

[EvanED]

Why are you so confidant that they will fail faster because of their choice not to support Linux?

It's quite possible that people have sat around, pounded out some numbers, and said "providing support for Linux will currently cost us more than the added business that it will bring in", and they consider the loss of your sale part of the cost of business.

Just because they lost a sale doesn't mean that they are losing out on the whole.

(Of course, this isn't in any way meant to suggest that you SHOULD buy hardware from companies that say this, that your decision is wrong, etc. Also, it seems unlikely that on popular pieces of hardware (for instance, video cards from a certain manufacturer with a 3-letter name) this tradeoff would go against Linux. I just get tired of hearing the same things over and over again when they really aren't all that correct.)

Re:Unfortunatly...

[SEWilco]

I wonder if they also calculated how much they lose from Linux developers not helping them debug their software; I also am less likely to help them. Outside developers, that is, as a significant fraction of a company's developers are certain to be Linux users (that's what I hear when I talk to them).

Kernel and Xorg Usabilty

[grumbel]

KDE and Gnome do a lot of things for usability, but some usability quirks have their root deep down in the kernel (awkward handling of CD/DVD, lack of stable ABI for kernel modules, userspace-fs could need some additions, kernel features that need a kernel recompile instead of just a module, etc.) or Xorg (hot-pluging of input devices isn't supported, no real graphical configuration tool, way to easy to get a non working configurations, etc.) and can't be fixed elsewhere no matter how much wrapper magic you throw at them. Lack of a real widespread distribution portable package format would be another major issue.

Re:Kernel and Xorg Usabilty

[bersl2]

I think Xorg 7.3 is scheduled for release on the 29th. Yes, as in a few days. I don't think they'll make that date, but hotplugging on input and output seems to be somewhat imminent.

Mod article "-1, Troll"

[Anonymous Coward]

Seriously. what do you expect by this "article?" The author does not even attempt to make this an interesting / productive discussion, by narrowing a focus or even distinguishing between "linux" and "GNU/linux" or F/OSS software.

It's like asking, "What don't you like about your life? How would you improve your life?"

Open Source 3D and better interactivity

[Omnifarious]

The first of those two is self-explanatory. High-quality, high-performance Free 3D drivers for good hardware.

The second...

I want some (not all) kernel developers to stop using throughput based metrics to measure performance, and instead use a metric based on interactive performance. I have a suggestion for such a metric...

The time between user input and the user input having a noticeable affect on an output device like a display. And I don't think this time should be as short as possible, though that's a good goal. The time should be as consistent as possible while remaining short. I propose a metric that measures this latency and plots the standard deviation of the latency and uses that as the main metric with that average latency being a secondary metric.

Re:Open Source 3D and better interactivity

[rawler]

Agreed, but I would say this is not only a matter of performance. On the fuzzy perspective, it's about making the user feeling in control. One of the things that currently make me feel out of control is the numerous times when I find myself out of reach with the OS, due to that SOME application have begun some heavy thrashing, and I can't even regain enough control to kill the application.

What I would like to focus some effort on, kernel-wise, is to
  * Skip IOnice. Process niceness/priority should basically not differentiate between what resource it's trying to access. Either it's a prioritized process, or it's not.
  * Making process priorities inherited, through the resource-blocking-chain. Basically, the effective priority of a process should always be the highest of it's own priority, and all the priorities of the processes that may be waiting on it. In this scenario, put the X-server on top, make sure that background tasks are run through an event post-subscribe-mechanism, and more or less for free, you'd get a system that magically focuses on what the user needs feedback on right now. In a server-setting, elevating the priority of httpd, for instance, will ensure that whatever might be blocking httpd (MySQL, syslog, other background-services on the same host) will always be prioritized over, say, cron-jobs and similar.

In user-space there's tons of stuff to do;
  * Increase parallelism to further utilize resources. (I once did some tests indicating that parallelizing the reading of all the application startup resources in kwrite could cut the worst-case loading time by as much as up to 80%, depending on system, fragmentation and configuration)
  * Decrease memory footprint. (COME ON?!)
  * Take a long serious look at X11, and walk away while we can
    - Revamping most of the graphics subsystem is about time.
    - KGI has some interesting points, regarding parallel sessions. It doesn't matter that your kernel might be running stable, if the graphics card have left you without a working terminal.
    - A completely OpenGL-rendered desktop is about time. Accelerated compositing is nice of course, but hardly a substitute for ubiquitous vector graphics. (Resolution-independent, of course)
  * Hierarchical file systems work only for hierarchical people.
  * Fixing parallel sessions. My computer should be strong enough for several desktop-sessiosn. So why does the current solutions suck in terms of memory-usage, graphics issues, switch-times, desktop-incompatibilities....

On the usability side, there's a ton of stuff to do as well. For instance, you can check out my current pet-project at http://wiki.kde.org/KShortcutAssistant, but there's a lot of other issues to address.

Preferences, doc files and print settings

[JoeCommodore]

I've seen some great preferences and some lousy ones, and then there are a bunch that are still only accessible via text editors. Getting some of the guys that are doing a bang up job on the good ones to help with the bad ones would be a start (i.,e. get the RH guys from system-config-samba to help on the gsambad project.) Also get a nice recoverable video preferences system (to go back to a one-size-fits most mode if you totally mess up the monitor settings).

How about including some of the documentation files WITH the app packages (sometimes they are there as an often overlooked separate install, sometimes there are none included. :-/)

Last but not least would be print settings - some are nearly great (Open Office) where everything just about works (i.e. Tiling should not loose stuff in the print margins). And other are near non-existant (Inkscape).

Those are the things I think are past due.

Most linux problems are "desktop" issues...

[bytesmythe]

1) Wifi networking
    Too many wifi cards (especially the Broadcom chipsets) are painfully difficult to get working correctly. WPA2 encryption support is flaky. Wired and wifi should switch gracefully.

2) Better sound support
    There are too many conflicting ways of producing sound, some of which dislike working together. Midi support should be built-in. Currently, it's a pain to install. Hopefully KDE's Phonon subsystem will help here.

3) Better a/v
    Too many movies have unsynced audio and video. Also, many codecs are unsupported. Yes, I know they're proprietary, but I don't really care. Ubuntu is making codec installation easier, but frequently the codecs only work with some particular backend. (For instance, even with mp3 support installed for gxine, Amarok (a KDE app) still needs to install it's own. The desktop environment should provide a generic way for apps play audio, and if a KDE app is running under a Gnome environment, it should be able to "just work".) Don't forget the wonderful closed-source
graphics card drivers!

4) Easier windowing subsystem
    No one should have to edit xorg.conf to get anything working. Fortunately, the next release of X windows is supposed to finally do away with this by adding dynamic configuration with xrandr. Also, it will be nice when CompizFusion is more robust. Lots of people really like the eye-candy, and I find some of the features useful.

5) Applications
    It should be easier to keep applications up-to-date. I love Ubuntu, but it drives me nuts seeing bug fixes or major enhancements to applications that I can't easily obtain because either the OS updates don't include application upgrades, or the OS repositories are simply not adequately maintained. I don't want to have to
litter my package manager with repositories, or manually install packages just to keep my apps updated.

6) Laptop support
    Suspend and resume don't always work very well. Some laptops don't come back, and frequently networking
is messed up.

Re:Most linux problems are "desktop" issues...

[Chandon Seldon]

Your points #1 and #6 are hardware support issues. You chose your hardware - that makes them your problem, not someone else's problem.

You wouldn't buy a PS3 or a PowerPC Mac and expect it to run Windows. Computers with Windows pre-installed are no different - it's only guaranteed to run Windows. If you want to run some other OS, it's up to you to make sure that you buy hardware that lets you do that. That hardware isn't uncommon, hard to find, or more expensive - it's just the next box over on the shelf. Blaming other people because you wouldn't move your arm two feet and grab a different box is absurd.

Re:Most linux problems are "desktop" issues...

[g_rampage]

That's missing the entire point. They want linux to support more hardware so that becomes less of an issue.

Clearly if someone buys a card with the intention of using it under linux then they are stupid for not making sure it's compatible but it would be still be easier/better if there were more options.

Re:Most linux problems are "desktop" issues...

[Chandon Seldon]

And how do you propose we use this supposed development money to wrest hardware documentation from uncooperative vendors?

Actually, that's pretty easy. You buy Broadcom.

Re:Most linux problems are "desktop" issues...

[InlawBiker]

I agree completely. 98% of the problems I've had you listed above. Most of those can be narrowed down to driver issues, i.e. hardware support. Wifi, audio, laptop sleep/awaken, xorg - most of my problems with X were due to crappy ATI drivers. Have you ever plugged an external monitor into a Mac? It's automatically detected and turned on, like magic. That's the way it should work. A/V issues are a function of too many competing codecs and players. It will sort itself out. We are in the infancy of Linux desktop. Stuff we take for granted on Windows or Mac is still being worked out on Linux, but it will come around. On the server it's worked for a long time, because I compile and tailor everything how I want it. The desktop changes at an exponentially higher rate, it's a lot more complicated.

Re:Most linux problems are "desktop" issues...

[Anonymous Coward]

Laptop power management on Linux has been crappy for ages. Pretty much since APM was abandoned in favor of ACPI, actually. Suspending to disk has been in a work in progress for an eternity, and the rest of the power management landscape isn't so hot, either.

I run Linux on my laptop, and I'm fine with it, but I certainly never expect to be able to hibernate it with any sort of reliability.

Re:Most linux problems are "desktop" issues...

[Anonymous Coward]

4) Easier windowing subsystem
What if, and I know this is a radical idea, but what if X would automatically fall back to a reliable driver (e.g. vesa) when using unsupported hardware? I've heard that Windows users don't have to edit files in a text console to get a display before they have a graphics driver for their new card!

Re:Most linux problems are "desktop" issues...

[bytesmythe]

I specifically purchased a laptop with Intel wireless and graphics because of the drivers. WPA2 encryption works, but only if you put all the magic voodoo programs in place. I am perfectly able to do so, but I should not have to. This is a higher level software issue, not a driver issue. There is no good program that will automagically manage the wireless connections, use encryption, and seamlessly switch between wired and wifi. The closest program is probably wicd.

Anyway, the topic post simply asked about areas I would like to devote a number of developers. What I want is a bunch of broadcom/atheros/ralink/nvidia/ati/amd/intel/hp/ca non/samsung/etc. programmers writing open-source drivers.

Re:Most linux problems are "desktop" issues...

[Anonymous Coward]

Bang, you've pretty much just described the major issues with Linux distros. I agree with all of those.

Re:Most linux problems are "desktop" issues...

[Fophillips]

No one should have to edit xorg.conf to get anything working. Fortunately, the next release of X windows is supposed to finally do away with this by adding dynamic configuration with xrandr. Also, it will be nice when CompizFusion is more robust. Lots of people really like the eye-candy, and I find some of the features useful.

On my current install I haven't touched xorg.con, I just ran

hwd -x

and it sorted it all out for me.

It should be easier to keep applications up-to-date. I love Ubuntu, but it drives me nuts seeing bug fixes or major enhancements to applications that I can't easily obtain because either the OS updates don't include application upgrades, or the OS repositories are simply not adequately maintained. I don't want to have to litter my package manager with repositories, or manually install packages just to keep my apps updated

That depends on the distro you use. All my apps on Arch are up to date, Ubuntu and virtually all Debian based distros have become a dumping ground for old software.

Re:Most linux problems are "desktop" issues...

[Kjella]

No, it's not easy to find for mortal men. Your average store or webshop doesn't list linux compatibility in my experience. You can be 100% sure that the people in the store have no clue. You have to dig up different lists for everything you want. Want to buy a printer? Better check the foomatic support pages. Want to buy a scanner? Better check the SANE support pages. And if you're buying a whole PC... seriously, I don't know all the places I need to look. Often the "it works fine" messages mean "it works fine after I spent hours on google figuring out how to hack it to work right".

Your average store is full of Linux lemons, and they look exactly the same. Your average person expects whatever he buys to work, because that's exactly what he's used to with Windows, the days of "Plug&Pray" are over. You never, ever have to go into a computer store and ask "Is this compatible with Windows?". The question hasn't even crossed their mind. So they do exactly the same, and come home with a paperweight under Linux. Linux very seriously need a compatibility program, and I don't mean an obscure web page somewhere, I mean a proper icon that's tagged on boxes and listed on specs everywhere.

Re:Most linux problems are "desktop" issues...

[LWATCDR]

The lack of a stable binary driver api makes hardware support for Linux difficult for the hardware vendors.
1. They can not include a driver CD with the product that will install and just work.
2. They can not offer a single package on their website that will just install and work.
These are issues even with FOSS drivers! I have had to recompile or use the /force option to get drivers to load. That fails the easy to use test.

Re:Most linux problems are "desktop" issues...

[Anonymous Coward]

Your average store is full of Linux lemons, and they look exactly the same. Your average person expects whatever he buys to work, because that's exactly what he's used to with Windows, the days of "Plug&Pray" are over. You never, ever have to go into a computer store and ask "Is this compatible with Windows?".

It's not necessary to ask, because it says "Designed for Microsoft Windows" on the box. If it doesn't, it probably doesn't work.

I have a machine in my living room that doesn't have a "designed for Microsoft Windows" sticker. Want to give installing Windows a try? You might want to use an original disc, though, it refuses to read most "CDR edition" discs. So, either an original silver DVD, or an original black(!) CD-ROM. That machine is claimed to run Linux just fine (but you'd need to order it directly from Sony Computer Entertainment (wherever)).

Linux supports lots of hardware, including lots of hardware Windows doesn't support. The problem is not lack of hardware support (Windows' lack is even bigger), but that you basically needs to bring Google along when going shopping, because noone cares about Linux support. The manufacturer doesn't care enough to write it on the box (this is changing, the USB hd tray I bought mentioned Linux in big friendly letters), the sales people don't care enough to be able to answer, and so on.

In short: What is lacking is customers. Enough customers to make the manufacturer care, enough customers to make the sales person care.

Fix-up the standard toolchain

[Hoplite3]

There are lots of exciting new things in linux: HAL, DBUS, UDEV, etc. These have changed the way hardware is detected and activated, mostly for the best.

But many of the new tools that deal with this stuff are GUI programs like network manager. Now, network manager is a good program, but wouldn't it be cool for it to coordinate with ifup/ifdown? You know, update the classic commands so they use the new systems. I think there could be either a new generation of CLI tools or a re-vamping of the old.

I'd also like to see more development of client/daemon type programs, where large pieces of the execution are separated from the GUI/CLI interface. This could make a lot of programs more fault-tolerant. Like bit torrent clients that could continue to download when the user logs out, etc.

Re:Fix-up the standard toolchain

[Anonymous Coward]

Run your bittorrent client in the nifty tool called screen. (You won't be able to use a GUI client with this solution though. rtorrent looks like it's going to be able to do away with some of the more useful edges that are so common with GUI bittorrent clients, but it isn't usable just yet (at least the version that is included in Etch (I'll have to switch to testing soon, I guess)). The reasons are: parts of the interface don't even fit into 142 columns, and that's about how many you'd usually be able to fit on a 1024x768 display; the program seems to insist on checking hashes every time (unless the download had finished already, whence you supposedly can pass an option that prevents this).)
With screen, your controlling terminal could disintegrate, and you would still be able to regain access to your program by reattaching it (-d -r $pid)).

Linux? What's Linux?

[NotQuiteReal]

I don't know where to get Linux...

I can get Ubuntu, Gentoo, Debian, RedHat, etc., etc... I can't get "Linux".

There's your problem...

Re:Linux? What's Linux?

[Anonymous Coward]

I second that. The huge number of Linux distros is one of Linux's greatest strengths, and also one its greatest weaknesses. The Linux community is much too fractured. There is a lot of effort being duplicated unnecessarily. Look at GNOME and KDE, for example. Why does Linux need two desktop environments? It doesn't! Microsoft--and Apple too, I suppose--must be having a great laugh at all this!

Firefox is a great example of what can be achieved when the open source community focuses on a single project. Why, after all, did the Mozilla Foundation discontinue the Mozilla application suite? It was a smart move! Of course, now the Mozilla application suite is back in the form of SeaMonkey, which just seems like a major step backwards.

Re:Linux? What's Linux?

[Chandon Seldon]

Linux is available on kernel.org - the current version is 2.6.22.5

Re:Linux? What's Linux?

[rastilin]

Personally I disagree with that. I've always felt that a fractured community comes naturally when people can do their own thing but that this was in no way a problem. Since everyone has a different goal, they can make something that fits their ideals. There's two important principles

1. The people who design distros would not necessarily want to help on a different project if their work was unrequired.
2. The distributions exchange massive amounts of code. There is surprisingly little work being duplicated without necessity.

Linux doesn't need two desktop environments, which is a good example of duplicated work. But both desktop environments are going in different directions, if they were all on the same project, it's entirely possible that nothing would get done and development would slow down.

Re:Linux? What's Linux?

[Geoffreyerffoeg]

"Sadly, a kernel by itself gets you nowhere. To get a working system you need a shell, compilers, a library etc. These are separate parts and may be under a stricter (or even looser) copyright. Most of the tools used with linux are GNU software and are under the GNU copyleft. These tools aren't in the distribution - ask me (or GNU) for more info."

-Linus Torvalds, September 1991

Re:Linux? What's Linux?

[Anonymous Coward]

What's w32krnl.exe?

Re:Linux? What's Linux?

[Anonymous Coward]

This is why I tell people to get Ubuntu. Even if they ask about Linux, just get Ubuntu and worry about learning the stuff you posted about earlier. This whole 'too many distros' argument is getting a bit tired, we have a dominant, stable, beginners distro now. Stop moaning and get recommending. :)

Re:Linux? What's Linux?

[MPAB]

When it comes to praising or publicity, everyone refers to Linux as the distros that can surf the web, read mail, create documents, act as servers, act as firewalls, etc.
Alas, when it comes to defend against those that say linux is command-line intensive, has buggy and unsupported software, lacks drivers for wireless or video, is painful to install software on, etc. the answer is "Linux is only a kernel, go blame the distros".

This is the win-win scheme I've come across lots of times. And I think the first lack of consistency that must be adressed in Linux is this one. Whether it is just a kernel or a collection of distros, it should be the same for good and for bad.

Re:Linux? What's Linux?

[entgod]

I agree that Linux doesn't need two desktops. It needs, and has, more. Most of the people will always just stick to the most popular few but there are quite a many people who don't want all the windowish bulk they carry with them. I for one wouldn't like to use kde nor gnome but if all the Linux users were forced into using something like ion3 there would be a whole lot of more windowers.

Re:Linux? What's Linux?

[buddyglass]

You're missing the point of my question. Regardless of which distro I choose, I'd still be running the linux kernel (i.e. similar hardware support issues) and I'd still be running many of teh same apps (Firefox, etc.) If you feel like the best way to spur linux adoption is to improve a particular distro, then that's a valid answer. Maybe not the one I'd give, but an answer nonetheless.

Re:Linux? What's Linux?

[Anonymous Coward]

Linux is the kernel.

The thing we see is a desktop, from which starting point we run applications. To finetune them, we can use CLI. The host of 'just-under-the-skin' applications that make all this work used to primarily be GNU, but there have been some notable additions over the years.

The 'failure' (if that is what anyone wants to call the fact that over 1 per cent of all personal computing devices already run these 'unorganised' collections of software) lies in the fact that the GNU/Linux world hasn't the cash behind it to seriously 'brand' the software in a hurry, the kind of hurry that would stop these kinds of discussions cold.

The original question was, given the resources, what should be done? Most of the suggestions I have seen here relate to market related things, such as drivers and so on. Well, I don't know what it takes to get hardware manufacturers to issue or support good drivers, but that is what is most needed.

Yes, and maybe easier ways to install apps. It would be good to have a wide range of apps run across the plethora of desktops as advertised (well, they don't advertise much, and thank goodness for that).

Re:Linux? What's Linux?

[tepples]

Alas, when it comes to defend against those that say linux is command-line intensive, has buggy and unsupported software, lacks drivers for wireless or video, is painful to install software on, etc. the answer is "Linux is only a kernel, go blame the distros". Then complain that GNU/Linux "is command-line intensive, has buggy and unsupported software, [...] is painful to install software on, etc." Still, drivers for 802.11b/g networking are a kernel problem, but probably not any fault of the kernel developers. A lot of it is the fault of the FCC (and foreign counterparts) with its "robustness rules" against shipping wireless transmitters that can be modified by the end user to increase transmission power beyond the legal limit in a given jurisdiction.

Better developer tools for creating gui apps

[ErikThompson]

Creating gui apps should be as easy on linux as it is on windows. gtk and glade are pretty good but aren't as easy to control the layout and size of the widgets as they are in windows. If there was a Python IDE that was as feature rich and similar to Visual Studio it would make the transition for Windows developers writing programs for Linux a whole lot easier and more attractive.

Re:Better developer tools for creating gui apps

[Anonymous Coward]

Certainly in the native code arena, qt is better than anything available on windows -- that's why a lot of people use qt on windows even if they don't need to ship on multiple platforms.

Re:Better developer tools for creating gui apps

[Repossessed]

Ow ow ow.... please don't write GUIs in Python. Or C#, or god forbid Java. I moved to Linux to get *away* from the ridiculous hardware inflation Windows has. (A note to any Firefox developers out there...)

Re:Better developer tools for creating gui apps

[sohp]

I take it you're a VB6 developer?

Re:Better developer tools for creating gui apps

[Chandon Seldon]

Is it really that you need better tools, or is it that you're used to some existing tools and don't want to take the time to find and learn anything new and potentially different?

Open Source programmers definitely develop what they personally need first - if there were any blatant shortcomings in the development tools, that problem would have been fixed long before many other apps got written. I suggest you figure out how the many existing programmers are managing.

Re:Better developer tools for creating gui apps

[ErikThompson]

I have taken the time to learn something different and am fairly used to using wxPython and and am also figuring out gtk. I personally don't write as much gui code as I used to so it isn't a huge issue for me and I can take my time learning. I was just answering the question on how I would allocate resources into improving Linux. It seems to me that making it as easy as possible for developers to migrate over or start fresh is the way to go. It would mean more people writing code on and for Linux distributions.

Its been in development for almost 20 years

[Anonymous Coward]

There may be more problems than just time and money can solve with the Hurd.

Thourough context sensitive help...

[AmazingRuss]

Seems like everytime I click the help menu, I get some skeleton outline, if anything at all. I don't mind googling around for the information, but if usage is going to grow outside of the techie segment, the help systems are going to have to catch up with Windows 95 era chm files, at least. I'm not talking about technically, but rather actually having some useful content in the systems. I understand that writing documentation is no fun, so I don't hold out much hope for this.

Sure would be nice though.

Educational software

[psherma1]

For students and teachers. Make it so useful for the schools they'd be crazy not to use it. All else would follow...

Obligatory

[qweqwe321]

While you're at it, you can fund the development of Duke Nukem Forever, too!

The defacto desktop distro for the general public

[the_rajah]

seems to be tending toward Ubuntu. Simple to install, no command line needed for the casual non-geek user, intuitive GUI for the convert from Windows. I'm using it right now on my Dell desktop that came pre-installed with it. I already had it running on another machine, but I wanted to vote with my wallet for linux to be available Joe Public as already installed on a commodity computer.

Additional software is easy to install. I'm using Opera to post this and, even though it's not available through the Ubuntu add/remove, getting it downloaded and installed required no CLI use.

I've got another machine with Fedora 7 and have played with a lot of other distros including SuSE, Knoppix, Mepis, and some of the Ubuntu forks like Kubuntu, Xubuntu and Mint.

We have a guest room in our house that is complete with a guest computer.. It runs Ubuntu Feisty and our guests have not had a problem using it. I just tell them that it's not Windows, it's Linux. Here's the browser and feel free to use any of the networked printers if you need to print something.

refocus Linux?

[guardia]

or the free software movement? We're talking about the "bazaar" here... It's fine the way it is now. We've got great distros like Fedora, Ubuntu, etc. Everything's working pretty well, really. The problems are mostly political (read: multimedia codecs, patents and such) and somewhat technical (read: Microsoft Office proprietary files, etc.)... that won't go away easily even if you put thousands of hours of the best programmer or lawyers on them. I hope the rumors about Red Hat's effort on some licensing deal for codecs is true, that should make for some interesting evening news. :)

Build more doghouses

[Anonymous Coward]

One quote I've heard is that programmers want to build palaces, not doghouses. What this usually leads to is "roll your own" approaches to items that should be done in a standard way (case in point: print handling). And, like it or not, Open Source is very prone to building palaces.

In many ways, this is because of the "do it yourself" spirit that made open source great. People are passionate about their software, and they want it to be as good and as "can you top this?" full featured as it could be. But, regrettably, this means different projects will handle the same thing differently, or have different priorities. The wheel gets reinvented more in open source projects than it does in general. I've seen a lot of really good programmers spend hours writing a framework that's exactly what they want rather than take something off the shelf that's perfectly servicable but has features they don't want, or has a few quirks needing coding around.

How would I use 100 good programmers to fix this? By stepping back from the project-by-project trees and seeing the forest. What's usually used together on the same system? How do two otherwise similar programs implement similar concepts differently? Then, take the best (or an amalgam of the best of multiple approaches), and contribute it to the projects that are off rewriting what's been written before.

"Every project for itself" is in many ways an approach that's outlived it's usefulness.

Do everything at the same time...

[martijnd]

I think Linux has always had this "everything at the same time" feeling to it; so things move ever so slowly. Some languish, some die, sometimes people get angry about it and things get fixed. So many people pulling in different directions ; many projects died but their best ideas live on.

It used to be a nightmare to configure hardware -- its now easier than installing XP on a Vista machine. X had (and has) so many problems it wasn't funny; but these days you can click around for days and it mostly just works. Wine was a joke for years -- but I can run my favorite online (DirectX) games at decent frame rates and progress is fast. For years it felt like all Linux coders lived in the USA; now proper Unicode support & multi language support make for example Chinese/Japanese input much easier.

Linux is a giant wave always moving slightly behind the edge, companies can make money by living on the bleeding edge. But slowly all of them get washed away.

Junk the cruft.

[gwern]

We could ask for a lot of stuff, but let's face it, most of what is going to be suggested wouldn't really help and would have short-term benefits at best: Linux already has plenty of device drivers, those developers couldn't do much more than is already being done, and certainly not the several percent needed for a real shift; the various desktop environments are just not going to merge, let's face it, and so we're stuck with the horrors of diversity in our visual interfaces; new software certainly isn't desperately needed cf. all the people living perfectly happy with existing software; and so on. What would be useful would be if those developers could go around and either: figure out some way to make X not suck. There are better ways out there, guys; if X really is the best way to provide a GUI, then we might as well cut our wrists open now. Heck, X wasn't even the best solution back when they were first writing it in the '80s. A lot of our GUI and usability complaints can be either directly or indirectly traced back to the fact that X is so dominant. Let's have them write a better system (a Y, perhaps), add in a compatibility layer, and reap the benefits. More radically, we could have them go around and try to persuade people to adopt a reasonable proposal: don't explicitly support any hardware created or software written before 1990. How much cruft has built up on the altar of backwards compatibility? How much disk space is taken up for APIs, programs, options etc that were deprecated back in the '90s? We should simplify and cleanup; should your pre-90s whatever still work, that's swell. But we need to move on! This is a radical proposal, I know, but the unseen weight of all that old shit is a hidden tax on every innovation and every process that occurs in the Linux world. A good long-term investment would be to try to pay off some of that debt, if you will. Clear out the rubble and build our modern stuff on a clean foundation.

osx like desktop

[ZipprHead]

On the desktop, hands down. If there was an awesome, advanced, easy to maintain and use ui ( cough osx like cough). It would take hold.

Yes I Am writing this from my iphone.

Re:osx like desktop

[calebt3]

That is up to Gnome and KDE and the likes. As well as those that make GTK themes. And it can already be done. Just look at this site: http://www.taimila.com/?q=node/11

Also, Ubuntu 7.10 will include compiz-fusion, and it's parent project, Beryl, is pretty neat.

Re:osx like desktop

[Eli Gottlieb]

Or better yet, why not take the open-source base of OSX (say... OpenDarwin and the BSD userland) and write a full suite of F/OSS workalike software to replace Apple's non-free bits?

If you wanted it to run on Linux, you could port the thing to a GNU userland and... say... MkLinux or L4Linux.

Better removable media support, and...

[Anonymous Coward]

I haven't used Linux since Fedora Core 4, so maybe some of my suggestions aren't applicable now, but here they are:

* I know there's auto-mount, but inserting and removing CD-ROMs always seemed to cause problems, particularly if the drive was a USB drive that was spontaneously added.

* Try to improve upon the block device numbering scheme; "/dev/hdc#" is mostly fine, but
the numbering is arcane.

* Clean out some of the ancient junk in all of the configuration files. It's time to get rid of terminal definitions (e.g., VT100, etc) and all of the crazy modem definitions (e.g., some ancient 300 baud thing, etc).

Graphics Card Support

[Ledsock]

One of the most lacking things in Linux is out-of-the-box GPU support. Yes, they will usually work, but rarely to their full potential. I feel that this is one of the biggest things holding Linux back from mainstream adoption. Like it or not, eye candy sells people. Look at OS X or Vista. Both are loaded with eye grabbing features (not that it's helped MS's sales of Vista as much as they've wanted).

The other reason decent graphical support is key is due to games. Windows supporters often say that they need it to play games and there aren't any games for Linux or OS X. OS X is slowly gaining support since its switch to Intel, and Mac sales have improved likewise. Linux may arguably support more graphics cards than OS X, but the big difference is with OS X, you can tell by looking at the box whether it'll work or not. Linux often involves more detailed searching with Google to find compatibility. Not an issue for most slashdotters, but for our friends and relatives who don't like to tinker with forums, the dreaded command line, and technical settings, it can be a major headache.

Aside from persuading people to try Linux, graphics can be a bottle neck on systems and lead to performance and stability issues. Having a better graphical support system may also help developers in creating a more unified UI layer.

Make syscalls more orthogonal

[Anonymous Coward]

I'd try to add various non-posix syscalls that make the kernel more orthogonal.

Examples:

Versions of fork() and exec() that return a file descriptor that you can poll/select/epoll on to wait for a return code, instead of having to wait/waitpid for SIGCHILD.

Similarly, a tid could be changed into a file descriptor as well so that threads could watch each other's status in a simpler way.

Change brk() so that you can give it an offset like sbrk(), and make it return a pointer to the start of the added memory. This would allow use in multithreaded programs without locks.

Also, I'd love to have scatter-gather read/write calls. readv/writev are nice, but what if you want to read from multiple offsets in a file into multiple non-adjacent buffers? Or write multiple data chunks into "random" places within a data file. Currently you need to do a data copy first.

An aio barrier of lighter weight than fsync() would be nice to.

Finally, an interface for something like the solaris "iotop" would be really helpful. I'd like to know what program is accessing disk, or the network at any given time, or even which user.

Re:Make syscalls more orthogonal

[Dadoo]

Versions of fork() and exec()

I always thought the whole fork/exec mechanism was a little goofy, but personally, I'd be happy if they made a version of exec() that didn't have any limitations on command-line length. The fact that I may or may not require xargs on a command like "ls a*" in this day and age is ridiculous.

Re:Make syscalls more orthogonal

[EvanED]

You know, I've always thought that fork/exec is a bit goofy too (as compared to a system call that just launches a new process, like NtCreateProcess), but there are very good reasons for it.

Essentially, it allows you to change things about the environment before running the new program. Want to redirect standard input or output? You can do that between fork and exec. Want to change the environment variables the new process runs under? You can do that between fork and exec. Want to do that in Windows, with CreateProcess? You have to pass them as parameters, which means that everyone has to pay at least a little attention to these things that you may want to do but usually won't.

fork(2) is a zero-argument function, and execve(2) has three. CreateProcess has nine in parameters, one of which is a pointer to a struct with 14 meaningful fields. Not all of the added complexity is due to the decision to make a CreateProcess call instead of a fork/exec-type call, but a lot of it is. (E.g. three fields of the struct are handles to stdin, out, and err; one argument is the environment to run in; another is the working directory the program should run in. All of these could be removed with a fork/exec interface. Many other arguments and fields may or may not be able to be removed nicely.)

Fire the freetards.

[sakusha]

The biggest encumbrance I see in linux is the radical freetard mission, which is expressed in the licensing. Licenses like GNU have eliminated the possibility of real commercial development where it most counts. Nobody can create proprietary improvements and capitalize on their investment.

Let me give an example. Apple is the largest unix vendor today because they picked the BSD distro, which is under the MIT license. This allows them to create a proprietary GUI layer (Aqua) while still contributing freely to the underlying open source OS code. MacOS X works because a commercial enterprise has devoted significant development efforts into the weak spots of linux, specifically, consistency of operation and function.

Change the license. Fire the freetards like RMS. Get real commercial development going, and get everyone paid for what they're coding. For linux to succeed, it has to be an ecosystem where everyone makes money.

Re:Fire the freetards.

[Anonymous Coward]

>>For linux to succeed, it has to be an ecosystem where everyone makes money.

umm except that it's been succeeding, even more so recently, without this.

Re:Fire the freetards.

[sakusha]

Gee, let me see how well linux has succeeded without anyone making money.

Linux desktop marketshare: ~0.01%
MacOS X desktop marketshare: 5.9%, up 1% from last year

It's the license. Apple's use of BSD and the MIT license allows companies like Microsoft to write apps that can never appear on Linux. Like it or not, that is the barrier to linux on the desktop. If major developers won't touch the platform due to licensing issues, you can improve the platform all you want and there will still only be amateur apps (like it is now).

Re:Fire the freetards.

[Verte]

Linux desktop market share is usually placed between 0.7-3.0%.

Microsoft could easily write apps for Linux, the license doesn't forbid that at all. It just forbids people taking the OS, sticking a GUI/DRM/etc on it, and selling it back to you.

Look, sakusha, it looks like Linux isn't for you. If you want OSX, use OSX. If you want the most popular operating system [because that seems like the metric you favor], use Windows. If you want a reliable operating system that works well and won't be made propriety next week, if you want an operating system that you can modify and contribute freely to, GNU/Linux is the obvious choice. This is its most important feature, wherever it lacks technically it can be fixed by any interested party, but one thing is assured to you: it is free. Everything else is just implementation detail.

Re:Fire the freetards.

[Anonymous Coward]

Linux desktop marketshare: ~0.01% MacOS X desktop marketshare: 5.9%, up 1% from last year

Oh really, and where did you get those figures from: your arse? My arse? Steve Jobs' arse? Get the Facts...errr... I mean Compare's arse? Or was it from some crappy Web visitor statistics: how many jobless, emo, fuckwits visited MyArse.com in a given time period using MacOS X due to a post they'd seen on Slashdot with the title: 'ATTN:SWITCHEURS'.

So they thought to themselves, 'Wow, I'm not a switcheur! I'm 1337 enough to own a Mac!' So they went out and bought Macs, because they're fucking gullible, like you. Then they visited MyArse.com where your statistics are pulled from.

Here are my statistics:

Windows 74.1 %
Linux 20.6 %
Macintosh 2.6 %

All my extremely popular, yet undisclosed, Websites show roughly the same figures. People love the license, really it all depends on which orifice you pull your stats from and what the target audience of that orifice is.

Re:Fire the freetards.

[skeeto]

It is obvious you have no idea what you are talking about. Free software != non-commercial software. You can sell free software all day long and no one can do anything about it. In fact, the FSF encourages you to charge money when distributing free software. "Free" is not about price, but freedom.

I was going to write more but it is actually pretty obvious that you are just trolling. I am done.

Re:Fire the freetards.

[jb.hl.com]

Nothing is stopping anyone producing a proprietary graphics layer for Linux like Aqua. All that is required is the will and the cash, and at the moment neither appear to be in ready supply.

Which is a shame, because I'm sure there are many talented coders out there that could make something a hundred times better than crappy old X11.

Re:Fire the freetards.

[jb.hl.com]

It just forbids people taking the OS, sticking a GUI/DRM/etc on it, and selling it back to you.

Excepting DRM, isn't that exactly what Red Hat does?

great commercial software on the desktop

[Anonymous Coward]

That's the problem in a nutshell, isn't it?

There's no ecosystem in place for it. Vendors like Adobe probably think, if we did port our flagship apps to Linux, people would just stick it on a DVD along with 200 FOSS packages and pass it around to their friends as free software. And they'd get stuck with a massive tech support burden where often as not, the problem would be with hardware compatibility and driver problems, rather than with their applications. Sure, they have to deal with these type issues on Windows, but Microsoft has close to 95 pct share so it's worth their while.

question sabotages the answer

[epine]

The people responding here do realize, almost by definition, that "refocussing" Linux involves a hypothetical mode of intervention that stands at complete contrast to everything Linux has so far represented. There is no vagrant pool of talent at this level, nor is there a mechanism to confine this pool of talent to pie-in-the-sky wishlish thinking.

As a point of reference, this text has been in the OpenBSD dhcpd man page for as long as I can remember:

We realize that it would be nice if one could send a SIGHUP to the server
and have it reload the database. This is not technically impossible, but
it would require a great deal of work, our resources are extremely limited,
and they can be better spent elsewhere. So please don't complain about
this on the mailing list unless you're prepared to fund a project
to implement this feature, or prepared to do it yourself.

So far it has been extremely effective in scaring off any useful contribution. There are plenty of dirty jobs that need doing. Many small dirty jobs is worth more than a heaping serving of pie.

Re:question sabotages the answer

[value_added]

As a point of reference, this text has been in the OpenBSD dhcpd man page for as long as I can remember ...

An even better point of reference are the installers.

With respect to the BSDs, for example, new users typically bitch or are so put off by it that they refuse to consider it. Most everyone else doesn't care both for multiple reasons, among which is the reason cited in your DHCPD example, that none of use can't fathom the concept of repeatedly installing/reinstalling the OS to fix a problem, and that a full-blown GUI installer with limited features for customisation or automation is as counter productive as it is a waste of time.

At any rate, the bitch fest is entertaining to read. ;-)

From a personal computing standpoint.

[sdfad1]

Your questions concern things we can add to today's software to improve them. Let me flip that around. Why not REMOVE some software to improve the computing experience for humanity as a whole?

Why, in the 21st century, are we still stuck on 1970's paradigms? Why Linux? Why Unix? Why do we still work on layered and opaque systems? Is the dominance of the C-language, Unix-operating-system duo today a historical accident? Or was it purely evolution: survival of the fittest (fittest may even be far from the best technological solution).

Let me play the devil's advocate here by suggesting that we need to now rethink the old technology that is a systems environment that is not well integrated and extendable. Operating systems today are 'dead' and static (once compiled) monolithic programs. To bring dynamism to the platform, we then layer on abstractions like shells, (bash, csh, zsh, etc), scripting languages (sed, grep, awk, perl,+ newer friends), windowing systems, etc. Our compiler toolchain consist of separate tools that don't talk to each other except through what they see in front of them, on the 'bit conveyor belt'.

Eg the autotools -- M4, autoconf, automake, configure, make (roughly in that sequence). Here's a quick test, is there a way we can define a variable that is checked in each of those stages, to control certain (compilation, configuration, other settings) options? No. Instead, we pass on values through multiple languages/interfaces or filesystems in the form of strings, macros, environment variables, files, etc. Who's gonna maintain all these software, to fix and sometimes perpetuate(!) bugs?

May I suggest that this is not a problem with the userspace apps, but rather a systemic defect in the system we use today. I wish the operating system is open all the way (and I don't think this conflicts with the need for security in the system). I wish for a system that is extensible at runtime, one with a single language all the way, simple models of abstraction and no unnecessary barriers between programmer and user. I wish the computer would take care of the details, and help me do my work. I wish life was good.

Disclaimer: I do not have all the solutions to my complains, so I'm just a sufferer ranting. Ok, this is also not quite what you asked for, but it's probably not that offtopic, I'm just off on a slight tangent.

Re:From a personal computing standpoint.

[Eli Gottlieb]

I mostly agree with you, but you should face that we will never, ever run systems powered entirely by one language. Different languages solve different problems, and binary (+ interpreters that were themselves compiled to binary) provides enough of a common platform for most applications.

On the other hand, the use of a single high-level language (like bash, AREXX, or AppleScript) to tie multiple applications and their functionality together would be very nice.

Re:From a personal computing standpoint.

[sdfad1]

Sup Eli. I remember you from comp.lang.lisp!

Yes, I am (half) aware of the practicalities of todays world. However, as a Lisper, I find your lack of faith disturbin'. :)

Re:From a personal computing standpoint.

[Eli Gottlieb]

Have you taken a look at Scsh? I, personally, would like to see AREXX/AppleScript-like application scripting abilities added to something like that so it could drive GUI apps as well as command-line ones. Then I'd pretty much consider it perfect.

Re:From a personal computing standpoint.

[sdfad1]

Yes, but I had a lot of trouble trying to understand what it was all about (confusing syntax/semantics, at least for me). Nowadays, I have to work on windows too, so it's a mongrel mix of Common Lisp (as much as possible), Emacs, cygwin bash (or bash through ext:run-command in clisp). The problem is that most major applications don't really want to talk to other programs, so this is still quite limited. So it's still not quite the same as my wish, which is a full, live system all the way down to the metal, whatever it may be -- Lisp, Smalltalk, Forth, etc. It's great being able to physically see objects and manipulate them, but at the lowest level, we are still stuck with an opaque OS.

3D

[G3ckoG33k]

While it has been said before, I believe 3D is the path. Please, please, provide support on par with Windows for any 3D graphics hardware, whether inside a computer or a console.

That is the path to success on the desktop. Today, I cannot even run OpenGL apps or any 3D apps on the lastest and greatest 3D graphics hardware from AMD (formerly ATI), the Radeon 2900XT. Why? There are no drivers. They have focused entirely on Windows, and consider Linux a niche market not worth the effort. Because of that, my family do not have a Linux only machine, which is also why I dual boot. The Radeon 2900XT support may well come to Linux, "when it is ready".

Please, take 3D support in Linux more seriously, whether you are hardware manufacturer or a software developer.

Re:3D

[Chandon Seldon]

The software 3D support in X is actually damn good. On hardware that the manufacturer is willing to provide development information for (*cough* Intel *cough*), there is perfect 3D support. The lack of support on Nvidia / AMD cards is simply the fault of those companies.

Re:3D

[Anonymous Coward]

Perfect? The Intel X3100 graphics drivers still have some way to go. Try Google Earth on one, for example... unusable. Many other 3D applications have lots of issues with the newest intel drivers too. On the other hand, Intel drivers are improved quicker than any other graphics driver I have experienced

The Elektra Project

[Thaidog]

http://elektra.g4ii.com/Main_Page

I think it's at least worth trying such an implementation. Ok... now bring on the "It Windows again" haters...

It's so easy!

[copponex]

If you want to win the desktop war, you can, in a few years, by asking yourself a single question:

Could my grandmother (who is already "sort of" computer savvy) use this without calling me every five minutes?

It's been a minute since I've used Linux as my desktop, but if users are still being forced to edit text files to change common program preferences, you'd better get used to your third seat behind Windows and OS X. I'm not telling you to have some crazy xml schema with a billion pieces fronted by a hefty GUI - I'm just asking you for the option of using a lightweight GUI to parse and store my preferences to the same text file.

Keep your CLI, and -color-code-for-Klingon-language-support options, but don't even try to force that on every day users. Leave stuff exposed so you can work your admin magic, but build some sane GUIs for everyone else unless you enjoy end-user support.

Re:It's so easy!

[Chandon Seldon]

Could my grandmother (who is already "sort of" computer savvy) use this without calling me every five minutes?

Dunno. My Dad doesn't seem to have any problem with the Ubuntu machine I gave him, and he hadn't used a computer since Windows 95. I'm pretty sure he hasn't magically learned how to edit obscure config files or anything. It's not like he's not doing anything at all with it either - he had no problem setting up his printer, for example.

Re:It's so easy!

[Verte]

I don't get it. How is changing:

option 1: true

to

option 1: false

any more difficult than finding the menu that lets you make the change and clicking the button? I think people will get used to text config files, and soon it will be GUI configuration tools that feel dated.

Re:It's so easy!

[Psychor]

The difference is that options within an application GUI are offered to you, are obviously relevant to that application, and are usually faster and simpler to reconfigure. Editing a text configuration file requires that you a) find the file (locations vary between distributions, and sometimes multiple versions are stored, making finding the active one tough) b) understand the format of that configuration file, which is not standardised (may require checking documentation) c) ensure the file is saved in the right format (e.g. some editors screw up line breaks which a novice user might not expect). While it's still not hard to change a true option to false, if you want to add an option, it's often not clear from config files what options are available, and reading documentation is unavoidable.

And that's assuming everything goes according to plan. Error checking within a GUI normally makes it impossible to invoke invalid options. However a mistake in a text file could just cause an application to silently fail, which then usually requires finding the relevant log file and deciphering error messages, costing more time.

This is speaking very generally in terms of standard GUI applications. Obviously for console applications (and certain other specific/unusually complex apps) a text configuration file makes a lot of sense.

Re:It's so easy!

[Verte]

I seem to somehow missed that you'd said common program preferences, so I apologize. Those really should be accessible via the GUI. Weather text file or GUI though, options should always be easy to understand. I've got a pretty serious gripe with Firefox's "about:config". Very few of those options are self documenting.

Re:It's so easy!

[Valpis]

Doesn't matter what kind of ssytem we build and provide, how easy things are, even if there is just one button on the screen which says "CLICK ON ME" and makes all the things they want totally automatic. They will call us and ask if they should click on that button

Re:It's so easy!

[jrieth50]

Agree fully - a friend of mine got a new Dell the other day that arrived DOA (frozen on the BIOS info screen) - despite the obvious "linguistic" issues that arise when using Dell tech support - before he got off the phone he asked how long the typical support call lasted. His response? TWO HOURS.

Ask yourself, if it takes on average 2 hours to walk your typical support center caller through what are probably most often simple mundane tasks, how do you think those calls would go if they were supporting a typical linux distro?

"ok, now grep glxinfo for the following..." - yea right.

Re:It's so easy!

[FooBarWidget]

I know my grandmother can't use Windows (or OS X) even if she did call me every 5 minutes.

Re:It's so easy!

[naapo]

I'm just asking you for the option of using a lightweight GUI to parse and store my preferences to the same text file.

I felt the same way, but then I found two excellent general-purpose GUI configuration editors for Linux. They're called emacs and vi, and with bot of them you can edit the config files on graphical screen. No more flipping the switches on the front panel! The only problem is I haven't yet been able to decide which one is better.

Re:It's so easy!

[$pace6host]

If you want to win the desktop war, you can, in a few years, by asking yourself a single question: Could my grandmother (who is already "sort of" computer savvy) use this without calling me every five minutes?

Well, I can only offer you my anecdotal evidence. Last year, my Mom (age 72 at the time) had a Windows XP box that my brother gave her. I think she's a fairly typical senior user -- she wants to browse around on the internet (she shops for air travel, gets directions, reads email), use a word processor (she writes business correspondence, articles for the newsletter, etc.), and play games (Free cell, Mah Jong). My brother and I regularly provide phone support for her, and/or remote in to fix things. One time mid-last year, after a "my computer doesn't work" call, I stopped in to visit Mom, and found her PC in an endless boot cycle. Somehow, something had corrupted NTOSKRNL.EXE, which is a fairly rare problem that I don't expect any non-techie to fix. I searched around, but couldn't find the repair / install disks. I don't have a spare copy of NTOSKRNL.EXE, so I used my laptop to burn an Ubuntu disk, and went through the repartition / install procedure. When it was done, I arranged the desktop icons so that Firefox was where it used to be, OO.o Writer was where Word used to be, Freecell and Mah Jong were where they used to be, set the theme to look vaguely Windows-ish, imported her Firefox bookmarks, and set it up so she could get to all her old documents. She had no problem using it for the next few months. Oddly, I actually took fewer support calls from her during that time. She's back on Windows now -- my brother just ended up replacing the computer, it was a little dated anyway. At one point, Mom mentioned that she liked what I had done to the computer (Ubuntu). Apparently, it didn't get stuck at shutdown "Installing Update 1 of 3" every night, it didn't pop up Zone Alarm notices, didn't want her to reboot to complete updates, and it didn't didn't take 5 minutes to log on while starting virus scanners. I think that there are actually a lot of people out there who would be perfectly happy running Ubuntu instead of Windows, they really don't care what OS they're on. They are just unlikely to ever have a computer with Ubuntu on it as an option, since very few come from the store that way.

It's been a minute since I've used Linux as my desktop, but if users are still being forced to edit text files to change common program preferences, you'd better get used to your third seat behind Windows and OS X. I'm not telling you to have some crazy xml schema with a billion pieces fronted by a hefty GUI - I'm just asking you for the option of using a lightweight GUI to parse and store my preferences to the same text file.

It must have been a while -- for her setup, I didn't edit any config files by hand. Now, you probably can't set up a web server w/o editing config files by hand, but you should be able to set up a modern distro like Ubuntu for a typical user w/o it.

What I'd like to see is something that the Linux developer community can't do: Increased support by 3rd parties, like digital camera or MP3 manufacturers. What the Linux community has done is amazing, and though it takes a few months, most of these things are supported eventually, but I'm sure Mom would find it confusing to buy an iPod or SanDisk player and plug it in if she was still running Ubuntu.

Re:It's so easy!

[Anonymous Coward]

My grandmother has to ask very often with Windows. It's not like Windows is soo much better than Linux in terms of usability. The only reason that people don't seem to have a problem with windows is that they 'grow up' with it. Linux is different, and it should stay such. It should not really that hard to use Linux. People just need to realize that Linux is different. Mac OS X is completely different from Windows, and you have to ask, "How is this done?" And that is how people view OS X. It is NOT a windows replacement, it is a completely different OS.

How long have computers been around? More than 20 years. Why do we expect people to 'not have to learn' when it comes to using computers? We have cars... are they so user-friendly that we don't have to take driving courses? I think it's the same thing for computers. If they can't use Linux because they are so old they think they can't learn, then by all means, let them not use Linux. If the whole world used Linux, would Linux be better from it? In some ways, yeah, but in many other ways it would not. I think we should just leave Linux the way it is, based on CLI. Improve stuff, but don't make it so 'user-friendly' that it drives power users up the wall. CLI is great. Sure you gotta learn it, but once you know it, its very fast and efficient to use. If all you want to do is the basics of computing, why care what OS you use. Windows is good enough for something as stupid as checking your mail. Why change to Linux just for the sake of 'using Linux'. Each OS is suited for different kinds of people. I highly highly doubt that Linux will ever become more used in the Desktop world than Windows --- this has nothing to do with which is better --- people just grow up using Windows, and they have no reason to change. At least most of them.

If I am in windows, and I want to know what DNS servers I am using, I have to go through all these other interfaces -- and Windows is slow on that kind of stuff, even if you have a fast computer. Linux on the other hand all you have to do is open a terminal and type "cat /etc/resolv.conf" -- two things you have to do. Not 10. But do we all need to know what DNS servers we are using? Not most of us. So let it the way it is. If they need to know on Linux, then they can learn it the way it is. Linux wasn't made for computer-illiterate people. It was made for developers. Let's not waste the developers' time by asking them to make useless GUIs.

Re:It's so easy!

[CuteAlien]

I've heard the mantra of "lets make Linux easier to use for " quite often. And I don't like it. My mantra is "lets make Linux easier to use for it's _current_ users". You heard it right - make it easier for the technique-fans, for the developers and for the sysops. Because these people are the current Linux market. Don't piss it off trying to reach people who are very far from getting interested. Currently nearly everyone has it's regular fights with the Linux system. We still like it, because of it's power, but the rough edges are there.

I think it would be a very bad idea to suddenly go for the grandmother market, when all the strong points of linux are of interest to the technical people. Extend it by adding feature by feature for the "a little less" technical guy.

It's Beautiful Just the Way It Is

[scruffy]

Don't let the perfect be the enemy of the good.

Unified packaging system...

[644bd346996]

...with good support for building from source. We obviously need a standard package format with robust support for complex dependencies. The build-from-source part is also really important: right now, no distro (not even Gentoo!) makes it easy to, for example, compile your own Mesa libraries and have them used by the pre-compiled X server. Right now, you can pull a project from cvs/svn and do a make install. But it will overwrite the version from the package and break dependencies. This greatly raises the barrier of entry for testing new code, making the "open source" aspect of Linux software far more accessible.

Once we have a unified packaging system, the meaning of a "linux distro" will change. There will be a lot more sharing of work for the base system, and separate distros will really become sets of config files with just a few changes from the upstream code. Kubuntu is a great example of this: it is a low-maintenance specialization of Ubuntu.

Re:Unified packaging system...

[Anonymous Coward]

I think you're wrong about that, you're just not very familiar with how it's done. On my Debian systems, for example, almost all of my APT repositories also have a related deb-src line pointing to the source repository. With one command, I can automatically pull down the source package and install any necessary dependencies. With another command, I can produce the dpkg (and optionally install it). Before I actually build the dpkg, I have the option to fiddle with the sources as I please. The most useful reason I find for doing this is backporting a package that's only being compiled for experimental releases (which are pretty good about staying on the bleeding edge), and thus won't work with my current libraries.

Some distributions make it easy enough to package-ize source installs, but you need to recognize that turning a source tarball into a package isn't fundamentally a simple problem (which is why I generally leave it up to the dedicated Debian package maintainers, and then swipe their work :) ). I don't think this is a problem that needs to be solved. Perhaps made easier, but it's hardly what I would think of as critical.

I think the requests for better platform-level multimedia support (for heaven sakes, does every single application need to build its own custom interface to ffmpeg), and more consistent package management (of which your suggestion could be a part) would be the most useful.

That said, this is all an exercise in mental wanking, because it's a pure hypothetical and I have no idea why this is even on /. If someone were actually looking for a project to do, or had the ear of the kernel developers, then it'd be worth discussing, but as it is, people are going to just do things the open source way, and work on what appeals to them most. And that's fine, but let's not delude ourselves that we're going to change anything by just talking about it.

Re:Unified packaging system...

[EEBaum]

On my Debian systems, for example
...
Some distributions make it easy enough to package-ize source installs

This was the parent's point exactly, I think. The problem isn't so much that there's not a WAY to do it, it's that there's not A way to do it. There are too many ways to do it, each with varying levels of effectiveness. Different distros have different levels of support for different methods of package management. If there was a single easy way to handle packages on ANY linux distro, THAT would be nice.

Reduce the amount of kernel-space/root code

[Schraegstrichpunkt]

There's still way too much code that runs with elevated privileges. I want to be able to run downright *malicious* code on my machine, under X11, and not be worried about it compromising my SSH keys, for example.

If this doesn't happen, we're eventually going to see the malicious screensavers in Linux that we've in Windows throughout the last decade.

Re:Reduce the amount of kernel-space/root code

[Gothmolly]

If you're running your code as with your user privs, you've already lost your keys. If you run your code as root, you've already lost your keys. You're talking about a sandbox. This is a non-problem, as there are already half a dozen virtualization options.

Re:Reduce the amount of kernel-space/root code

[Schraegstrichpunkt]

I'm talking about the fact that the probability that there are privilege escalation holes on any given system approaches 1, because the interface between unprivileged and privileged code isn't generally auditable.

Kernel is great, rest is so so

[timmarhy]

The kernel isn't where we need improvment - it's the developer tool kits that are needed. All the open source developer tools are crap, i'm sorry to say it but it's just the truth people. compare gtk to QT (yes i know QT is free, but not if you want closed source which is what many many people want), or compare anything .net to... well i can't even think of anything that's even close. look at the direct x tool kit compared to opengl, sure opengl can do some ok things but developing in it compared to DX is a nightmare. another prime example is reporting tool kits. open source has NOTHING even remotely as good as crystal or ms reporting services.

I program in python and am an avid postgresql fan, just to show i'm not some ms troll. i just want better developer tool kits. THESE are the things holding linux back, not it's kernel or it's admin tools.

Re:Kernel is great, rest is so so

[sohp]

So, you want it to cater to VB programmers?

Re:Kernel is great, rest is so so

[timmarhy]

I want people to actually have a reason to program under linux. Just because not all of us feel the need to fight clunky buggy tool kits all day just to achieve basic tasks, it doesn't makes us "VB programmers". i'm guessing you know exactly what i'm talking about, so you have chosen to go the path of language snobbery because you realise the validity of my comment.

Re:Kernel is great, rest is so so

[Chandon Seldon]

Have you learned to use Emacs? No? Then you don't get to say it's crap.

Seriously though, the tools are there - even full featured and fully-integrated IDEs - you just haven't looked, or if you have looked you've dismissed things like Eclipse out of hand for some arbitrary reason without trying them.

Re:Kernel is great, rest is so so

[timmarhy]

Since when does a text editor like emacs equate to a development tool kit? It's got one purpose - to edit text, which it does well. but it's nothing more.

I've used plenty of open source tools, so i suggest you drop this attitude that I haven't looked because it makes you look childish. The vast majority of OSS tools lack the polish that is required for professional's to choose them over commercial products.

I will even give you a very specific example. I recently went looking for visual studio plugin for SVN. I tried anhk, an OSS version and visualsvn, which is $50 and pretty much just a wrapper for tortise svn. visualsvn pissed on anhk from great heights. well worth the $50. the reason? lack of polish, shoddy features and poorly thought out interface.

these failings are a common theme through out OSS. I guess it's indicative of the development model many of these project follow, nerds get interested in something and work on it till it fits their needs then stop work, where commercial products have customers to please. They work on a project till it fits the customers needs, not their own.

you'll notice OSS projects that do stand out are all commercially funded.

Don't get me wrong, if i'm not willing to pay i won't complain about shoddy software, but this article asked the question so i'm answering it.

Re:Kernel is great, rest is so so

[amorsen]

Since when does a text editor like emacs equate to a development tool kit? It's got one purpose - to edit text, which it does well. but it's nothing more.

You have no idea what emacs actually is. I have mostly given up on it for text editing, but I use for lots of other things.

Re:Kernel is great, rest is so so

[Chandon Seldon]

I will even give you a very specific example. I recently went looking for visual studio plugin for SVN. I tried anhk, an OSS version and visualsvn, which is $50 and pretty much just a wrapper for tortise svn. visualsvn pissed on anhk from great heights. well worth the $50. the reason? lack of polish, shoddy features and poorly thought out interface.

So... you're saying that for a plugin for a proprietary Windows program, another proprietary program happened to be higher quality than one specific open source program? I'm entirely willing to believe that, and I don't see how it reflects in any way on the quality of F/OSS software on a F/OSS platform.

Focus on where linux is strong

[ryanw]

Linux is an amazing server. Since Linux and all the components uses the GPL I don't imagine anyone ever being able to invest any major amounts of efforts to refine the gui for consumers. Linux is very strong as a server. I really think people should just abandon efforts on trying to force Linux to be a widely used Workstation. It's a great workstation for developers and programmers. Not even close to ready for consumers.

Focus on the server side, consolidate efforts of all the developers to make it perform even better and be more secure and have more features and drivers as a server, not a workstation.

Re:Focus on where linux is strong

[Chandon Seldon]

Have you even used a modern Ubuntu desktop? On recent and supported hardware?

There's a reason that Dell is selling Ubuntu desktops, and it's not just because there were 25 very vocal geeks on the internet or whatever.

Re:Focus on where linux is strong

[timmarhy]

That seems to be your answer to everything, doesn't it Seldon. If anyone has an opinion that conflicts with your own, they must be idiot's who haven't used it, because honestly how could they be right and you wrong!

Re:Focus on where linux is strong

[jrieth50]

Speaking as a computer enthusiast, but the kind that knows absolutely no code whatsoever and relies entirely on the collective wisdom of sites like ubuntuforums and google groups, I've had a more than satisfactory experience with the Ubuntu desktop, and before that SuSE, and prior to that Red Hat. My current desktop running on my Thinkpad T42 rivals even MacOSX in terms of both functionality and looks, and really wasn't THAT difficult to set up.

The only thing holding linux back from a desktop platform that could easily be chosen over competing products are 1) hardware support approaching universality, 2) driver support from day 1 from hw manufacturers (see #1), and 3) simplified/universal/or cross-distro package installation.

I've been using linux for quite a while, and I like to play around alot with it, but it seems now like its just one short breath away from being an obvious alternative for anyone that does not REQUIRE that they run some very Windows-centric software like anything made by Adobe or any latest and greatest PC games. (Which COULD run just as effectively had they been developed with linux users in mind from the start.) Or linux could be one short breath from being doomed to servers and mainframes and hobbyists for another 10-15 years.

If linux is going to finally rise to the mainstream, it better happen in the next 2 years while MS is reeling from the fail that is Vista. I think the software side from linux kernel and desktop developers will be ready, the question will be whether or not the AMD/ATI's, Broadcoms, and Adobe Apollo's of the world (even Google to some degree) will prevent it from ever becoming a reality because they refused to take the platform serious enough to stop giving driver/software compatibility development anything more than scraps from their dev schedules.

Re:Focus on where linux is strong

[ryanw]

I've been using linux for quite a while, and I like to play around alot with it, but it seems now like its just one short breath away from being an obvious alternative

Yah, I've been on the "linux is one breath behind" team for over 10 years. Trust me. Someday you too will see that the GPL is what holds linux back. Developers that actually want to graduate from school and not be working at a college the rest of their lives need income. Working on Linux is not profitable. Everything you do for linux has to be done for free. I don't understand how Linux will ever rise above software that actually can sustain income to pay their developers a hefty salary.

If there were a Linux fork with the BSD license, you would see progress that might even attract some serious commercial support from Adobe and such.

Re:Focus on where linux is strong

[Anonymous Coward]

Working on Linux is not profitable. Everything you do for linux has to be done for free.
Perhaps you should tell IBM etc. about this. They seem to think that contributing to Linux on a significant scale is and will be profitable for them. They also pay their staff considerable amounts to do this work (IBM staff do not generally work 'for free').

Most significant work on most FOSS projects is in fact paid for by some company or other who believe they will benefit in some way.

Re:Focus on where linux is strong

[Chandon Seldon]

Working on Linux is not profitable. Everything you do for linux has to be done for free. I don't understand how Linux will ever rise above software that actually can sustain income to pay their developers a hefty salary.

If there were a Linux fork with the BSD license, you would see progress that might even attract some serious commercial support from Adobe and such.

That's false. You are either confused or intentionally spreading lies.

The GPL license does not imply that software developers work for free.

Unless you think that the programmers at Red Hat, IBM, MySQL AB, Zimbra, Sun Microsystems, all work for free. If anything, the GPL means more work for software developers - because they can charge money to add features to existing programs.

Selling propretary software for Linux works fine.

Sure, trying to sell an office suite that's worse than Open Office or a web browser that's worse than Firefox isn't going to work. There's no reason why it should. It's not like with Mac OS X either where people manage to sell low quality propretary text editors somehow. But numerous companies manage to make money selling high quality propretary software for Linux. Like the 3D Modeler Maya. Or Quake 4. Or Adobe Acrobat Distiller. Or anything on this list.

Re:Focus on where linux is strong

[Chandon Seldon]

If anyone has an opinion that conflicts with your own, they must be idiot's who haven't used it, because honestly how could they be right and you wrong!

Statistically, with "Linux isn't ready for the desktop" posts on Slashdot (and Digg), that seems to be a reasonably safe assumption.

People are very, very willing to jump from "I couldn't get Slackware to work after spending two hours trying in 1999" to "Ubuntu 7.04 is utter crap". In this case I was wrong. It turns out that ryanw is actually just an anti-GPL troll instead.

Where Linux Applications Fall Short

[Crispin Cowan]

Linux has achieved near parity with Windows in a lot of places. I've been a Linux desktop user since 1995. Between 1995 and 2005, I always used some kind of Windows emulation to run Microsoft Office and FrameMaker, because there is no Linux equivalent. Since 2005, OpenOffice has become sufficiently powerful, compatible, and stable, that I have not felt the need, at all, for Microsoft Office, and so I have completely given up using Windows emulators.

However, Linux is still sorely lacking in 2 key application areas:

• Really good DTP. FrameMaker rocks. Adobe had a Frame version for Linux in beta for a year, and then withdrew it :-( Yes, I know about LaTeX, I wrote 2 theses and 10 papers in LaTeX, and it is not an adequate replacement. Linux needs either a port of FrameMaker, or a clone. OO Writer is an easy-to-use word processor, not a DTP, it is like the difference between a hand saw and a lumber mill.
• Good webcast software. On Windows, WebEx is the clear winner, and they have a solid product. Delivering a demo to remote viewers of a desktop application is no problem. Doing the same thing in Linux is problematic, at best. Instead of one really good product, we have several bad ones. Mostly they just lack maturity, but that still means that I have to spend a lot on plane fare to give effective Linux desktop demos.

Re:Where Linux Applications Fall Short

[chris_sawtell]

Really good DTP.

Tried Scribus?

It's very good indeed.

Re:Where Linux Applications Fall Short

[value_added]

I know about LaTeX, I wrote 2 theses and 10 papers in LaTeX, and it is not an adequate replacement.

Not to sound criticial, but writing 2 theses and 10 papers sounds like what most any student does in their first term in a typical year in high school. Hardly evidence of extended or advanced usage.

I've used Framemaker longer than I care to remember, but I'd choose LaTeX any day of the week for any kind of project. Put another way, I've invested years learning and mastering my text editor and related tools (shell, sed/awk, Perl, cvs, diff, etc.) and would be a fool to throw that away for an approach that's narrow and implies vendor lock-in.

I guess that makes the score 1-1.

Re:Where Linux Applications Fall Short

[I'm just joshin]

2 theses and 10 papers in their first term of high school? I think that "theses" does not mean what you think it means. Thesis: a dissertation embodying results of original research and especially substantiating a specific view; especially : one written by a candidate for an academic degree (Merriam Webster) My brother-in-law's thesis was >800 pages.

Re:Where Linux Applications Fall Short

[Crispin Cowan]

Not to sound criticial, but writing 2 theses and 10 papers sounds like what most any student does in their first term in a typical year in high school. Hardly evidence of extended or advanced usage.

Perhaps "paper" and "thesis" do not mean what you think they mean. The papers are those published on major computer science conferences, and the theses were my masters and PhD theses (about 200 pages each).

... and would be a fool to throw that away for an approach that's narrow and implies vendor lock-in.

Of course. But the question was what I would put emphasis on for Linux, and my point is that the FOSS community has failed to produce an obvious winner in the DTP space. I don't want FrameMaker per se, I want an open source DTP that is just as good or better. Several other responders pointed me at Scribus I'll have to go check it out.

better drivers

[EagleEye101]

... or a better (maybe binary interface) way of creating drivers

CUPS?

[ynososiduts]

I'm surprised no one has said anything about printing. CUPS is a bitch to configure if you're knew to linux. Forget the fact that most printers aren't even *nix compatible, but they still take awhile to configure if they are. I would like to see better print support.

Re:CUPS?

[/dev/trash]

KNEW? KNEW? The word is NEW N-E-W.

Re:CUPS?

[Verte]

I keep hearing this. I remember reading ESR's essay on how horrible it was, long before I installed my first free operating system. But you know what? It was very easy to set up. I'd say it was less than thirty seconds to set up all of my printers [neither really popular, neither really recent, one is USB though] and print the test pages. I find it hard to believe that foomatic gui could be hard to use.

Re:CUPS?

[metamatic]

I'd like to see CUPS actually work printing to my IPP network printer.

Re:CUPS?

[/dev/trash]

nwe is a mistake. KNEW, is just a liberal education gone terribly wrong.

Re:CUPS?

[jimicus]

ESRs essay demonstrates a case of "Microsoft-itis" on the part of the developers of the print config tool in FC1.

IPP (the main protocol CUPS uses) makes it dead easy to share printers over a network. It periodically broadcasts across a network an announcement of what print queues are available on the host. You then have a daemon which listens to these broadcasts and a user interface which presents the user a list of printers it knows about - which is exactly what CUPS does and exactly what most modern networked printers do. Mac OS X uses CUPS under the hood, and it actually works pretty damn well when it's properly configured.

However, "Microsoft-itis" makes this feature a lot less useful. "Microsoft-itis" is where the OS, despite being developed from very early stages to run as part of a network of many, insists on presenting itself to the user as if there is only one computer in the world - the one that they're sat at - and no such thing as networking.

AFAICT, this was compounded in ESRs case by the other computer being configured not to share printers. Laudable from a security perspective but unless you think carefully about what you're going to present to the user, lousy from a UI perspective.

A really good video editor

[kie]

I am pretty happy with my GNU/linux boxes.
They are stable, have good hardware support, provide a good user interface and have
a great set of tools.
From this baseline further progress is being made in these areas (big thanks to all you developers).

The one application missing from my box is having a great video editor
(such as Apple's Final Cut, Premiere or even the new iMovie) that supports HD video,
(preferably licensed under the GPL).
The nearest that I have seen was MainActor which was proprietary, but is not longer sold.
Even that does not support HD.

While I applaud efforts such as Kino, cinelerra and gopchop they lag significantly behind
I am able to live with MainActor for the moment but I dream of a better video editor for GNU/linux.

Re:A really good video editor

[kie]

As it's a Christmas list anyway, I would also add written Chinese character recognition

I posted about it before
http://ask.slashdot.org/comments.pl?sid=197503&cid =16186751

There is still no written Chinese character recognition for GNU/linux on the desktop,
specifically so that a pen and tablet or mouse can be used to write characters

Motorola have already done it on their Ming A1200 phone which runs linux,
I wish i had it for my desktop!

Re:A really good video editor

[Salsaman]

Well, LiVES is trying to be that video editor you dream of...

Re:The defacto desktop distro for the general publ

[Simon80]

I've gotten Opera out of the add/remove utility before, but I couldn't tell you exactly how I did it.

No real innovation

[Anonymous Coward]

I use it everyday and like it, but as things sit right now, it does not bring anything new to the table. My hopes for Linux years ago, were that it would redefine the computer and the way we use it. It does not. It seems that most of the development for linux is centered around duplicating what already exists elsewhere. In the long run, that won't bring more than the geek crowd.

Publicity and Minor Improvements

[thePsychologist]

1) PDF support. Almost all PDF readers on Linux except for Adobe's product have difficulties with large PDF documents. What's with the "LOADING" message that takes forever? Adobe Reader looks horrible (inconsistent with the native GUI). There isn't a single PDF reader besides Adobe Reader that supports subpixel rendering which makes the font rendering hurt my eyes.

2) MIDI support

3) A "configuration manager" that knows most of the contents of the /etc directory and has three windows: a list of text config files, a window that displays the file, and a window with a paragraph or two of explanations and examples on how to change the file.

4) More active development of Fluxbox. It could use more features like shading on mouse wheel scroll and multiple backgrounds for each workspace.

5) A publicity website for Linux! This is probably the most important thing the Linux community could do. Features are nice, but who cares if no one uses them? The website would contain among other things:

-Step by step guide and interactive application to help people select a distribution
-Explanation of all major window manager/desktop environments, again to help people select.
-List of most mature Linux apps with description, screenshots, reviews, and commentary by users
-Discussion forums
-Latest on Linux section: demos of CompizFusion, new apps, tips and tricks, etc.
-Section specifically for articles on switching from Windows difficulties
-User friendly, designed primarily for noobs
-Linux store with quality Linux clothing
-Professional design

Re:Publicity and Minor Improvements

[calebt3]

I definitely agree with number 5. When I first wanted to go to Linux, I checked out Linux.org and it dod not say much that was useful to me. Such as how to download it. I wanted to download Linux, and did not know to look for Ubuntu or SUSE or such. And I did not get any directions there. I would have never switched if I did not know someone in person who had SUSE discs.

Re:Publicity and Minor Improvements

[abinnie]

1) PDF support. Almost all PDF readers on Linux except for Adobe's product have difficulties with large PDF documents. What's with the "LOADING" message that takes forever? Adobe Reader looks horrible (inconsistent with the native GUI). There isn't a single PDF reader besides Adobe Reader that supports subpixel rendering which makes the font rendering hurt my eyes.

KPDF uses subpixel rendering, and loads large documents inside a few seconds on my machine.
And if you're using KDE, it blends right in.

Re:Publicity and Minor Improvements

[Anonymous Coward]

2) Midi Support

This definately is available on Linux - in fact just about every device I have tried plugging in lately has *Just Worked*. What you really need to do is find a Midi connection manager - If you already use jack (Jack is linux's pro audio interface that lets you route audio between apps in the same way you can route midi- if you are into pro audio I can't imagine a reason not to) you can use qjackctl (The jack control panel) or patchage (Patchage is what I use) if you don't use jack you could try kaconnect.

For midi based editing and playback rosegarden is probably the first choice, Hydrogen is a very decent drum machine I leave you to find the other options for yourself.

A midi connection manager put into system settings would probably be the best thing.

Re:Publicity and Minor Improvements

[Verte]

I've never seen a PDF viewer anywhere that is quite as bad as Adobe's. Even on Windows it is flaky. When in Windows, I sincerely miss the Document Viewer that comes with Gnome. How large are we talking about? 2000 pages plus? I don't have many documents with more than 500 pages, so I'm not really sure what kind of size you mean.

I love the rest of your ideas, by the way. While MIDI support on Linux is fairly good now, I think we will take a few more years to surpass Windows as the musician's choice. None of the MIDI editors, for example, tickle me right now; but that is no one's fault but my own. Time to get hacking :)

Re:Publicity and Minor Improvements

[Anonymous Coward]

If you think Adobe Reader is bad, you haven't used OS X's Preview.

Re:Publicity and Minor Improvements

[Anonymous Coward]

How come the only people who know anything about multimedia on Linux are always anonymous cowards?

Re:Publicity and Minor Improvements

[SD-Arcadia]

For #1 : Did you try Foxit PDF? I've replaced Adobe's with this one on Windows because it's very fast and 2mb in size. There is a linux version, closed source, sdk available. http://www.foxitsoftware.com/pdf/desklinux/

Re:Publicity and Minor Improvements

[Anonymous Coward]

5) A publicity website for Linux! This is probably the most important thing the Linux community could do. Features are nice, but who cares if no one uses them? The website would contain among other things:

-Step by step guide and interactive application to help people select a distribution
-Explanation of all major window manager/desktop environments, again to help people select.
-List of most mature Linux apps with description, screenshots, reviews, and commentary by users
-Discussion forums
-Latest on Linux section: demos of CompizFusion, new apps, tips and tricks, etc.
-Section specifically for articles on switching from Windows difficulties
-User friendly, designed primarily for noobs
-Linux store with quality Linux clothing
-Professional design

I've got one more:
-Comprehensive database of in-kernel hardware support, sorted by degree of support/functionality and kernel version (maybe starting with whatever's the current kernel when this mythical tome is first created). If it's in the kernel it should also be in a consolodated, 'official' database/document. And don't forget the motherboard chipsets - they should have sub-entries for all the onboard features/controllers/whatnot, levels of support, etc. This database might benefit from a section that details hardware that isn't supported in the kernel but is supported by proprietary(or just not in-kernel) drivers.

Re:Publicity and Minor Improvements

[EEBaum]

Because if they posted their real identities, the sheer volume of messages to their accounts would simultaneously DDOS the internets.

Re:Publicity and Minor Improvements

[EEBaum]

I think we will take a few more years to surpass Windows as the musician's choice.

When you're done with that, you can work on surpassing Apple.

/upset that so much of the decent music software is not available on Windows.

Re:Publicity and Minor Improvements

[Verte]

What are you thinking of?

Re:Publicity and Minor Improvements

[Reziac]

3) A "configuration manager" that knows most of the contents of the /etc directory and has three windows: a list of text config files, a window that displays the file, and a window with a paragraph or two of explanations and examples on how to change the file.

This goes with what I keep suggesting: a dual-pane config editor, with *both* checkbox-with-clue-balloons and raw text modes available. Use one, the other, or both as suits you, but the idea is to both show the available options (and basic info on their effects) all in one unified place, and let the user SEE what actually happened in the config file, so they can learn from it, if they wish (but aren't compelled to do so).

What I would do

[Orion Blastar]

is improve the kernel so that it can use Windows drivers as well as Linux drivers to solve the hardware incompatibility that Linux suffers from.

If I were to improve an application it would be WINE, adding in better DirectX support and more support for Windows games so we could convert the Gameheads from Windows to Linux with WINE, and it would affect the sales of Windows.

Linux lacks a DRM media player, right now Windows has the advantage with Internet movie rentals and other media files that can play on Windows, but not Linux. You would need a freeware media player that can handle the DRM of iTunes/Quicktime, Windows Media Player, Real Player. Of course it cannot be under the GPL, and it cannot be bundled with Linux, but it can be an optional download for Linux users who want to use it.

If I had millions to spend, I'd have Windows Game Developers develop popular games for Linux to prove that there can be a commercial games market. Get Blizzard, Electronic Arts, 2K, Activision, etc seed money to rewrite or port their popular games to Linux.

If I had millions more to spend, I'd have application companies develop software like Wordperfect Office, Lotus Smartsuite, Photoshop, Quark Express, Paintshop, and other popular applications to Linux. I'd also give Mozilla seed money to write open source versions of Word processors, Spreadsheets, Presentation Software, and other Office software for Linux and multiple platforms that can use the MS-Office file formats, as well as open source file formats. The Mozilla code has an HTML editor that can be the basis for a good Word processor. It can also be used to tweak it into a spreadsheet and presentation software.

If I had millions more to spend, I'd give money to the OSFree project to get OS/2 applications to run under Linux, the Haiku OS project to get BeOS applications to run under Linux, the Amiga Research OS project to get AmigaDOS/AmigaOS applications to run under Linux, and I would fund money into a project to make an application to translate OSX API calls to Linux ones, so Linux can run OSX applications. Then Linux would be able to run almost any software from almost any OS platform, and people won't be able to complain of a lack of applications for Linux.

Re:What I would do

[calebt3]

Think we can get Bill to donate towards the fund that is gonna pay for all this?

Re:What I would do

[tepples]

Linux lacks a DRM media player Why would it need one? Would it be less valid to say "Major labels lack a non-DRM download store"?

I'd also give Mozilla seed money to write open source versions of Word processors, Spreadsheets, Presentation Software, and other Office software for Linux and multiple platforms that can use the MS-Office file formats Why Mozilla? Sun bankrolls OpenOffice.org. As for the image editors that you mention, what does GIMPshop lack that Paint Shop Pro and Photoshop Elements have?

and I would fund money into a project to make an application to translate OSX API calls to Linux ones, so Linux can run OSX applications. You can start by donating to GNUstep to bring it closer to source code compatibility with Mac OS X Cocoa.

Re:What I would do

[Orion Blastar]

You can't fight DRM because the RIAA and MPAA are too strong and the DMCA law enforces DRM on us.

I am tired of using Linux and not being able to listen to audio or watch video clips because the web site uses DRM methods that the Linux media players do not support. It would be easier to write a DRM compliant media player for Linux than to change the law that enforces DRM on everyone.

Why not Mozilla? OpenOffice.Org is almost mostly Java based. I'd like to see a 100% native code alternative to MS-Office available for Linux and other platforms that can use ODF, OOXML, MS-Office, WordPerfect, RTF, etc file formats. I think that Thunderbird already has a good editing system and we can take the core of Thunderbird and turn it into ThunderWord or something and then allow ThunderWord extensions to be written for it Mozilla style. Then ThunderCalc for Spreadsheet, ThunderPoint for Presentation, ThunderBase for Database, bundle them with ThunderBird and call it ThunderOffice or something.

I looked at GNUStep, I agree it needs help badly. Not only that but the cash to hire on expert programmers to improve it more.

The bad thing is that I've been out of work since 2002, and I am not a millionaire, or else I would come up with the investment money to improve these things with Linux.

Re:What I would do

[tepples]

You can't fight DRM because the RIAA and MPAA are too strong Why are the RIAA and MPAA too strong? How did they get that way, and why aren't they losing their grip?

It would be easier to write a DRM compliant media player for Linux than to change the law that enforces DRM on everyone. The law enforces DRM on everyone only when the owners of copyright choose to enforce that law. Not all copyright owners do. Notably, EMI does not.

Why not Mozilla? OpenOffice.Org is almost mostly Java based. And Mozilla is almost mostly JavaScript based, or at least JavaScript based to the extent that OOo is Java based.

I think that Thunderbird already has a good editing system and we can take the core of Thunderbird and turn it into ThunderWord or something and then allow ThunderWord extensions to be written for it Mozilla style. Might one start with Nvu?

First Fix

[Anonymous Coward]

First I would fix Internet Explorer. Oh wait...

stable modular hardware driver support

[akp]

A user should be able to install a set of drivers for a computer and have them work independent of distribution and kernel upgrades. So, for instance, if I go out and buy a new laptop, then once somebody has come up with stable video, sound, acpi, etc. drivers for it, then I should be able to install those drivers on my system and have them work. I should not have to compile a custom kernel, or wait until the new driver is packaged up in the official distribution repositories, or anything like that. And if my distribution comes out with a new kernel update, then I should be able to install it and not have to worry about it messing up my hardware support.

If this results in more proprietary binary drivers, then so be it. We just have to suck it up and deal.

I don't mind having to tell friends that, if they want to run Linux, they have to buy hardware that is supported. I mind having to tell friends that their hardware is supported, but in order to get it to run they have to hand-compile the drivers, or install an alpha distribution, or something else ridiculous like that.

Why all the overlap?

[Anonymous Coward]

For an 'outsider' (using a Mac at home and Windows at work) there seems to be a stack of overlap occuring - why have KDE and Gnome when the combined resources could probably make a single product? Further all the different bits are made by separate groups. It is confusing for someone (a mainstream user like me) to the point where it is 'easier' to just go buy a Mac if you want something different to Windows.

The one thing that's missing is virtualized 3d

[rastilin]

I've been using linux (not GNU/Linux) for over four years now and I'm fairly satisfied with my system. There's plenty of things that could be better but the one thing that drives me up the wall is that I had to give up gaming. Real gaming I mean, like Fear, Stalker, Titan Quest, Farcry, Bioshock and the upcoming Crysis; neither tuxracer or "Savage: Battle for Newerth" quite cuts it in comparison. The one thing that would make a massive and lasting improvement is a properly written DirectX passthrough driver for the open source virtualizer VirtualBox. It's very complicated but technically possible and considering the excellent developers that exist in the open source movement, will probably prove surprisingly quick to develop if people work at it. The main problem with the current implementations is that none of them support Pixel Shaders V2 and Parallels version 3.0 has not been released for linux. It has the major advantage over wine that all the wierdness will work. You won't have to prepare for all the strange undocumented API calls or deal with files cropping up in your home folder to mention just a few things.

Re:start over from the gound up

[Ultra64]

Yes! Starting over from scratch should get things done much more quickly then simply polishing the existing software.

Oh wait, that doesn't make any sense.

802.11 networking

[jnelson4765]

Wireless networking is still a pain in the arse to get working - the drivers behave differently, each card has its own behaviour, etc.

Stabilizing and standardizing the wireless stack, kernel- and user-space, would make the life of those who use laptops suck a lot less.

engage microsoft

[Anonymous Coward]

I'd engage Microsoft and Apple in talks about building their own compatible X-Windows gui and distro of linux. I know its heresy...but thats what it will take for Linux to keep growing.

Refine creative apps

[alfredo]

Bring GIMP up to the point where it would compete with PS. Do the same for audio and video apps.

Work on polish.

Re: Os X vs. Linux

[TaoPhoenix]

I see the suggestion for folks wanting pure GUI to go to OSX. However, I have this Animal Farm premonition of Apple's lock-in being just as bad as Microsoft's at some point. "Then one day, we could not tell the difference".

I think there are ways to let users know that the best raw power lives in the command line, but they need not ever use it if they don't wish to.

I know this sounds crazy, but...

[Arceliar]

Advertising.

I'm dead serious. OS X is a great example of what a little creative advertising can do. I know I'm over simplifying this, but OS X is basically FreeBSD (if I recall) with a proprietary UI installed on extremely well understood hardware to ensure things "just work"

And look at their market shares. It works. And, when coupled with their similarly simplistic accessories (iPod, iPhone, etc) and further tight integration into the OS, well, they've gotten a bit of traction to say the least.

It's not by any means what I'd consider a significant threat to Microsoft, but it's enough that a quickly growing number of major software applications are available on the platform.

I don't see why the same couldn't be done for linux. As I see it, at this point anyway, your average (competent) distro will do just about anything you want it to, out of the box (or after installing a few restricted packages) and with virtually no configuration--provided you have well supported hardware. If linux were advertised, at least enough outside the server market for more everyday people to become aware of it, then there might be more pressure on hardware manufacturers to release driver source code and/or enough hardware specs to actually write a good driver.

The only thing, and I mean the ONLY thing I see holding linux back right now is a vicious cycle. Hardware manufacturers don't make drivers because too few people run linux. Too little software gets ported to linux because too few people run linux. Too few people run linux because of the lack of familiar software and hardware support (that's not to say that there's any lacking in alternatives for most things).

Basically, what I'm trying to say is, despite my loathe of Dell I think the fact that they sell hardware with Ubuntu pre-installed is potentially the greatest thing which could happen to linux, at least likely from the viewpoint of a typical user. I just wish I saw an Ubuntu notebook or desktop in a Dell advertisement every now and again.

</rant>

All that now said, actual code base work wise...two things: unquestionably legal support for more media formats (too bad there'll probably never be a completely legal way to get dvd css support in the US), and virtualization. The latter I still consider to be in an almost infantile stage, but still very effective when properly configured. We still seem to be lacking in a completely GPL virtualization solution in the mainline kernel that's as intuitive to use as VirtualBox. A sleek KVM frontend would be nice.

actually you can

[Anonymous Coward]

Not to detract from your overall argument, because you're 100% right.
But Kino can include still images and generate DV files from them. :-)

Kino is the most amazingly easy-to-use video editor I've ever seen. Love the UI generally, and the features are decent for the stuff I need. (I only edit in DV format anyway.)
My only lament is the lead developer leaving for some KDE project (*sigh* for pete's sake can't I get away from QT once and for all?) ---- let's hope kino doesn't whither and die in the process.

Graphics and Package Management

[WhiteHorse-The Origi]

I want to play a fun game that works. I'd even pay for it. I just want to go look at games, pick one, buy it, and click "install". Right now I play crap-shoots with wine and barely anything works at all. If it does work it takes forever to load and can't read the hard drive or some bs bug.

I think the root of the problem is that there is no standard way to make a game that rocks for Linux and install it. With window$ you just double-click on an exe. With Linux you download a tarball, go to the command line, do a tar zxvf thefile, cd to the directory, type ./configure and then wait. Then it spits out an incomprehensible error message 800 lines before the end of running so you have to scroll back to find out that libgcpoop.so is missing. So you search for that and never find it unless you're lucky and figure out that installing mp3lib gives you libgcpoo.so and then you can proceed to the next error.

So maybe we just need a standard library? Possibly a meta-library that programs can reference at install and automatically install the missing libraries? Heck you could even make a games library that has all functionality of previously built libraries, you just need to know which library to instantiate... This gives rise to the need for a wiki for the standard libraries so that someone can start from scratch and learn all the libraries and use them.

Another thing that would make life great for us all would be a standardized CVS or SVN type package manager that allows us to just double-click a file and it installs from SVN the latest stable or dev package and keeps it up to date. Then developers don't even have to worry about packaging for distros, repositories become extinct, and the developer-to-user lifecycle is improved. You know you could even just write a Firefox extension which will take any SVN link, install and keep current that software.

Maybe if you threw a bunch of programmers at this you'd solve umpteen of the other posts you see here. Many people are reporting distro probs, repository probs, updating probs, lack of standardization, etc.

Oh and while I'm dreaming here, let's just scrap the keyboard/mouse and go with full-blown voice recognition so we don't even have to type. We just say whatever and point, maybe use a gyro-mouse or glove but no keyboard. I think this has already been solved with hidden markov models but you need a bunch of programmers to implement it and make all apps work with it. How cool would it be to just ask your computer a question and it tells you a pretty good answer? You know like the star trek computer. I mean come on people, this IS the 21st century and we DO have 64 bit processors with 4GB RAM...

As for all the "let's do this to make Linux more acceptable to mainstream audiences" comments, I say we should do absolutely nothing to make Linux like that. Forget OSX(Oh s3x) and Micro$oft... They are failing because they suck. Linux is succeeding because we do what we want, not what some corporate mainstream media tells us we want. Let's make Linux what we want, not what we think other people want...

Userland is hurting

[haemish]

1. I'd totally flush X11. It's a 20 year old design that has become a boat anchor for building excellent desktop apps.
2. Build some excellect desktop apps. eg. Gimp & Blender need total overhauls to make them professionally credible.

I'd leave server-side stuff alone: Linux is in excellent shape there.

Re:Userland is hurting

[Eli Gottlieb]

I'm going to quote my design text file for the whole "replace X" thing:

An improved windowing system will not function as the foundation of graphics fun
ctionality. It will run on top of that foundation. For example, it might run o
n top of an SDL implementation that runs without a windowing system under it.

I can temporarily call it the Fork Window System.

Fork runs as a daemon that accepts connections from threads. Applications conne
ct by sending ForkWS a display program they wish interpreted/run (written in a l
anguage like say... PostScript or something). Fork then forks off a thread that
  interprets the display program, drawing the desired window. Connections go onl
y two ways, between one display thread and one application thread. Display prog
rams can perform two functions: actually draw in their window and function as wi
ndow mnagers for sub-windows created using space in their window. Display progr
ams for sub-windows run in their own threads.

When a display program wants to send data to or receive data from the driving ap
plication thread, then they can invoke an IPC interface of some sort, probably s
ynchronous message passing. However, ForkWS itself will asynchronously queue in
put (and possibly output) events to be fed to specific display threads when aske
d.

The design pretty much forces programs into a parallelism model of one UI thread
  communicating with at least one driver application thread, which itself may communicate with several worker threads. Look up "NeWS" on Wikipedia to see my inspiration. A better windowing system was invented, but it lost out because it wasn't an open standard (and was therefore both more expensive to use and more expensive to develop for).

An Idea Worthy for the Linux Kernel Gods

[LifesABeach]

Every interface has been done, and re-done; Over, and over again. All but one. "Voice Input, and Output". The stuff of science fiction has never been fully realized. I plead my case before the Gods of Linux, "Make My Keyboard, and Mouse Obsolete!"

From a longterm everyday user....

[ph1nn]

The 2 things I've always talked about...

#1 - Standardize directories
I know it's great how flexible the operating system is but I would want a UNIX standard for everything when it comes to directory structures (where everything is stored) as there are some variations between distros and I've always found that annoying.

#2 - Standardize a packaging system
Also I would want a standard packaging system. There are Debs, Yums, Rpms, etc. While this is all fine and good think of how great it would be if all coders had to do is make a single package on all their sites for their apps.

The new Pidgin is a great example: A new version came out and my Windows friend (we've all got em) said "oh yea I just grabbed the new installer". And I gotta admit, I don't usually feel like downloading source code and compiling my own anymore. Since there is no package on the main site I goto apt-get it and as per usual it's not updated. I gotta get it weeks later from backports (I use Ubuntu). This is my biggest problem. The NVIDIA drivers are the only example I can think of right now where the installer works fairly well on a variety of distros. But even that has been a pain sometimes and I had to use Envy (which is awesome by the way).

Those are my main 2 things coming from someone who has used Linux (alongside Windows unfortunately) for about a decade. But I would also say improve ext3, the no noatime, nodiratime crap I always gotta throw on boxes is annoying. But that goes with the usual "speed, stability, and security" stuff.

Linux strengths = Linux weaknesses.

[Topherbyte]

It's only natural that the community wants to have its' cake and eat it too. Linux is a damn good server OS, probably the only one I'd ever use for back-end processing.

However, Linux will allow you to cut your wrists if you want to, and quite easily I might add. Whereas a successful desktop OS like OS X will make you jump through enough hoops to question your suicide attempt, yet remain flexible enough to allow you to do most anything. Why can't Linux be like this? Until the developers realize that not every Linux user is a Computer Science grad, Linux will languish in the realm of CLI junkies.

A balance between flexibility and ease-of-use must be found.

Just keep on truckin' ladies and gents. It's a relatively static target you're aiming for.

applications

[Anonymous Coward]

Since the topic is about Linux development and not about kernel dev: 1) Openoffice needs polish - the icons and usability lag behind Frame Maker and many other major word processors. 2) Better integration of Linux with Microsoft Exchange servers and Windows domains. I mean some serious integration and must work out of the box 3) VPN servers and clients - I have been using Linux for a while. I don't know where to start with setting up VPNs for a business that supports all types of clients 4) Linux desktop vendors - missing piece 5) Cut down on resource use by applications and windowing systems. Make them work on cheaper desktops better. Firefox/Thunderbird are two of the worst offenders.

"Foundation software"

[Micah]

Linux is already great technically. The biggest reason I know why it isn't being used in more enterprises, especially nonprofits, is the lack of one type of software: foundation software.

This is the software that controls the financials and business processes of these organizations.

A good offering in this area is a *must* prerequisite to World Domination. It is by far the biggest lacking area I can think of.

There have been some attempts, but they're not good enough yet.

Problem is, we don't just need geeks; we need people in these organizations who actually know in detail what the software should do, and how people expect to use it.

This is software that should tie together a lot of the free software stack -- mail servers, databases, openoffice, and who-knows-what-else.

The kernel or the operating system?

[OverflowingBitBucket]

Kernel?

Bring down the barriers relating to kernel development. We're talking documentation, convenience interfaces between the kernel-level stuff and userland, so forth. Spend a bit of time making kernel modules VERY feature-laden. Make them very easy to play with and ensure there are plenty of user-space tools to help you out. I've mucked around with a lot of stuff and have been developing software for a couple of decades now, but the Linux kernel STILL scares me.

Layer a set of version-consistent APIs above some of the low-level kernel stuff so driver developers don't have to target as many different setups or rely on a compiler being on the system to do their magic. I know this is a very unpopular idea in the kernel circles, but I think it would be very beneficial.

Of course I'm going to get ripped apart on the prior two paragraphs from people who know much more than me in those areas, so let's just say that these are just my thoughts from an external perspective.

Now moving on...

Operating system?

Somebody PLEASE develop a consistent library and API with minimal requirements that can interface with a whole bunch of windowing environments- including GNOME and KDE at a minimum. It should load the specific windowing interfaces dynamically so that using this common library adds no further dependencies to an application that uses it. From this interface I'd want to see fully-customisable keybindings, macros, and GUI controls of various sorts, an ability to hook to interesting events (eg. about to suspend, woken up, user logged out of GUI), info about screen layouts, access to user preferences regarding these applications (window positioning for particular apps, etc), and some assistance in loading and restoring state.

The library could then be taken and developed so that it is so appealing to developers they really have no reason not to use it, and the interfaces appealing to enough of the windowing environment developers so that they want to integrate it as well. It'd have a very liberal license (say BSD) applied to it to keep people using it.

Along with the library you'd have a set of tools that build on a whole bunch of environments (say: GNOME, KDE, and something that uses straight X). They would be used to set up all of the customisation that users could possibly want. The interfaces would have a simple mode for users that like very basic interfaces (actually to keep the people who claim that people want this happy), and a simple checkbox to enable "expert" mode that displays everything in obscene detail. The tools would have a sharing license (say GPL) to keep people pitching in their changes.

And then you'd need a whole bunch of people to promote it to make sure people know about it.

Imagine being able to fully configure all of your graphical apps to act how YOU want, drop in extra controls, keyboard shortcuts, trivially add in macros for remote app control, so forth- all without the developers of those applications needing to worry about it themselves. Why? The library handles it for them. The GNOME people and the KDE people can keep going about things their own way as well, and they'll keep making their own advances too. But they'll both be saved reinventing the same wheel in this common ground, whilst still having full control to take their projects to where they want to go.

I have been awfully tempted to attempt this myself but I know it would be far too large a project for a single person. I'd never finish it on my own, and I'm not interested in the politics it would take to get traction on such a thing.

But I'd love to see it.

Re:The kernel or the operating system?

[Verte]

[Just quickly, in response to the first two paragraphs, I think that's pretty much what /proc is for. It makes a lot of interfacing with the kernel simple.]

As for your library for working with a bunch of environments, I had the contrapositive idea a while back: hear me out. A tool to theme Qt, GTK, and your Window Manager as similar as possible. That way, it wouldn't matter what programmers developed for, because it would all look the same on your system. That's a lot easier than trying to unify different widget sets and the like with an extra interface.

+kudos on the configurable controls, key bindings, macros and scripting for the GUI interface- I think we will head in that direction soon enough. Maybe you should write some small example patches for Qt or GTK?

Re:The kernel or the operating system?

[Anonymous Coward]

www.syllable.org might be interesting for you... ok it has no intention of ever supporting gnome or kde but it has a lot going for it that linux distros tend to ignore.

Re:The kernel or the operating system?

[GreyWolf3000]

I've mucked around with a lot of stuff and have been developing software for a couple of decades now, but the Linux kernel STILL scares me.

You are clearly a much more experienced software developer than I am (I'm only 22 after all), but honestly I found getting into the kernel and hacking it to sit on a very reasonable learning curve. It started for me with a basic character driver I had to develop for work, and grew from there. Maybe what would help you is a reasonably simple task to start with.

Re:The kernel or the operating system?

[OverflowingBitBucket]

[Just quickly, in response to the first two paragraphs, I think that's pretty much what /proc is for. It makes a lot of interfacing with the kernel simple.]

Very true, and no doubt quite useful.

It's going to be hard to elaborate on the sort of things I mean without sounding like a fool since my kernel-level knowledge is exceptionally weak and I am do doubt surrounded by people who forget more kernel-level details in a day than I'll know in my lifetime. Let's just say it would be nice to have an easy "in" for this sort of thing. Basically it would be nice if I could easily defer certain kernel features (say file security) to a separate user-space daemon that I could play around with fairly safely. There are also some things that are more suited to userspace- say anything involving complex parsing or structure building. It'd be nice to have plenty of hooks into the kernel to use such things.

As for your library for working with a bunch of environments, I had the contrapositive idea a while back: hear me out. A tool to theme Qt, GTK, and your Window Manager as similar as possible. That way, it wouldn't matter what programmers developed for, because it would all look the same on your system. That's a lot easier than trying to unify different widget sets and the like with an extra interface.

You'll probably find a fair few {KDE,GNOME} {users,developers} who are quite hostile at making their environment similar to {GNOME,KDE}. But I think if you could somehow develop the sort of unification tookit that would preserve their distinctivenes whilst allowing you to blend the best of both featuresets you'd leave everyone happy.

+kudos on the configurable controls, key bindings, macros and scripting for the GUI interface- I think we will head in that direction soon enough. Maybe you should write some small example patches for Qt or GTK?

The thing I'm talking about is more a complete system than small patches for widget sets. Until the groundwork of what I've described was built up I imagine the maintainers for each of them would look at the patches and wonder what it gains them... then realise it is very little, and drop them. :( However, if it gained them hooks into a whole lot of the features I described previously, I can see developers jumping all over them. I guess it's all about making something that makes window system developers happy, application developers happy, and provide features that the users of the apps want. Unfortuantely that means a lot more work upfront, and then getting developers excited enough to get behind it as well. To be honest it's something I'd love to play around with if I had the time- which of course I don't.

Re:The kernel or the operating system?

[Verte]

A theming tool wouldn't require intervention of the KDE or Gnome developers. You're never going to get them to look exactly the same [I think] or behave exactly the same just through theming, but if it was easy to make a theme that was mostly consistent across the different widget sets you use, it'd be more attractive to mix and match.

As for a scriptable interface, I was thinking it would be simple enough to capture and manipulate calls to the GUI libraries, but that's not enough. For example, clicking on a button may bring up another menu or dialog box, and you won't be able to get that information just catching those calls. So it does need mostly a new paradigm. Or, as many applications have done [Emacs, Autocad, Maple, MagicVLSI?] make all features command line based and make the GUI and keybindings call the command line functions. That way you could separate the interface much like you can define your emacs macros.

This is, of course, the beautiful thing about CLIs in general- they are easy to extend. You can do great things with a GUI if you have a good CLI base. But I guess you didn't need me to tell you that :P

Re:The kernel or the operating system?

[Eli Gottlieb]

Scripting for GUI applications shouldn't actually be too hard. Amiga's AREXX did it twenty years ago, and AppleScript does it today.

Re:The kernel or the operating system?

[dodobh]

Somebody PLEASE develop a consistent library and API with minimal requirements that can interface with a whole bunch of windowing environments- including GNOME and KDE at a minimum. It should load the specific windowing interfaces dynamically so that using this common library adds no further dependencies to an application that uses it. From this interface I'd want to see fully-customisable keybindings, macros, and GUI controls of various sorts, an ability to hook to interesting events (eg. about to suspend, woken up, user logged out of GUI), info about screen layouts, access to user preferences regarding these applications (window positioning for particular apps, etc), and some assistance in loading and restoring state.

xlib?

I have been awfully tempted to attempt this myself but I know it would be far too large a project for a single person. I'd never finish it on my own, and I'm not interested in the politics it would take to get traction on such a thing.

Start the project, show the benefits from what you have, and you might find interested people joining you.

Re:The kernel or the operating system?

[OverflowingBitBucket]

- Somebody PLEASE develop a consistent library and API with minimal requirements that can interface with a whole bunch of windowing environments- including GNOME and KDE at a minimum. It should load the specific windowing interfaces dynamically so that using this common library adds no further dependencies to an application that uses it. From this interface I'd want to see fully-customisable keybindings, macros, and GUI controls of various sorts, an ability to hook to interesting events (eg. about to suspend, woken up, user logged out of GUI), info about screen layouts, access to user preferences regarding these applications (window positioning for particular apps, etc), and some assistance in loading and restoring state.

xlib?

Haha. ;)

But yes, something like that- at least in the sense it is a common base that the various windowing environments use. With emphasis, of course, being on that it is something that they all want to use.

- I have been awfully tempted to attempt this myself but I know it would be far too large a project for a single person. I'd never finish it on my own, and I'm not interested in the politics it would take to get traction on such a thing.

Start the project, show the benefits from what you have, and you might find interested people joining you.

Reasons why I won't be doing this covered in my original post. It's still nice to dream though.

Herding Cats.

[Anonymous Coward]

"How Would You Refocus Linux Development?"

Missing poll item: none of the above.

Re:Herding Cats.

[EvanED]

Herding cats becomes a lot easier if you have catnip. ;-)

re

[Anonymous Coward]

If i had to improve something it would be SDL. SDL lacked funding and so they changed their lookout for the future into a less motivated development because of lack of funding. I think software is the #1 issue in gaining linux adoption (as opposed to hardware or piracy, or proprietary formats) . And i think in the realm of software Games are the biggest incentive. It's either games or office. So i would either fund SDL, and try to make it , combined with openGL a drop in replacement for any of the Direct X's

Linux Improvements?

[Anonymous Coward]

I've been a Linux user since kernel 0.99.14j. I'd like to see:

- More manufacturers willing to provide development/driver/interface info.
- Slightly less painful "gadget" support for USB and ParPort home-brewers.

I'd like to see the following community issues get fixed:
- The KDE and GNOME camps need to stop bickering. KDE has problems and GNOME has
problems. I use 'em both. They're both useless if you don't set them up right.
- vi and emacs both have awful interfaces. I prefer Kate or Gedit.

Linux on the Desktop is already here. I've been using it as my primary OS for the last decade, and normally boot up my secondary OS only to use hardware that NetGear refuses to release driver info for.

Architecture and improvements, not religion

[CaptCanuk]

I'd suggest a couple of different areas for improvement.

First, I'd give up the holy war some kernel developers are having with proprietary drivers. Drivers are proprietary because some companies want to release proprietary drivers. Live with it. Stop fighting it and try to make everyone's experience better. 95%+ of Linux users today (a much MUCH different demographic than from 10 years ago) do not care about open drivers as much as working hardware. If you aren't going to help make people's experience with the OS better, then in the words of Ludacris: "move, get out the way, get out the way".

Second, build infrastructure that's obviously missing. There needs to be a central "registry". Let me rephrase that; a central key-value pair repository. Make it and let everybody know. Make sure drivers (X Server / Kernel) can talk to it and applications can talk to it and read/write.

Third, it's time to optimize for the desktop and the server independently. Make some of the differing parts that really offer those performance gains modular (factory pattern) so that they can be plugged in and compiled easily. Want desktop speed? Get the kernel, drop in Con Kolivas' scheduler and off you go. Want a server oriented kernel, then stick with the other scheduler.

Fourth, stop building reversed apps against a benchmarking utility that exists. You may need to reinvent the measuring stick before you re-invent the wheel. Otherwise you get a wickedly optimized component on the benchmarking utility but ends up lying to you in comparison to real world scenarios.

Re:Architecture and improvements, not religion

[EvanED]

Second, build infrastructure that's obviously missing. There needs to be a central "registry". Let me rephrase that; a central key-value pair repository. Make it and let everybody know. Make sure drivers (X Server / Kernel) can talk to it and applications can talk to it and read/write

I've thought about this for some time, and I'm still on the fence as to whether it's a good idea. I definitely think that the Windows way is wrong. I'm not convinced that a registry-like thing is a bad idea, but you have to do it carefully.

Besides the obvious issues like stability, reliability, security, scalability, and performance, I have what I consider a critical feature: is it easy to figure out what settings a particular program uses and delete, backup, or move them to a new computer?

I can't do that with the Windows registry, not by a longshot. I can do that with Unix dot-files.

A "centralized" registry doesn't preclude being able to do this, but if you can't, then I think it shouldn't be done.

Re:Architecture and improvements, not religion

[CaptCanuk]

That is singularly an implementation issue. You could have an overlayed registry system that splits the locations and priority of the keys in different places on the filesystem i.e. main "registry" is in /etc/conf/conf.conf and group specific registry can be in /etc/conf/groups/.conf and user specific registry would be in ~/.conf/conf.conf. Each one could specific the same key and based on how the key is created originally on the system, it can either be overridden for user specific tasks or group overrides or only system level values. Migration because easy this way and if all applications started using it, I can actually find out what the users last launched browser is if I need to invoke that instead of finding all browsers and opening one.

Re:Architecture and improvements, not religion

[EvanED]

I'm not sure that's enough. There also needs to be a convention in place that means, for instance, that a particular application only reads and writes keys within a certain, easily determinable key range (otherwise what keys to I move over when I want to migrate?), or possibly a couple ranges (like application settings and WM settings or something).

I think I also want to see separate files for each application that correspond to those ranges, instead of a single, monolithic one. This isn't essential (you can back up certain keys in the Windows registry), but I still think I would like it. (Alternately: make a view of the registry as a file system, a kind of ConfFS, that is mounted at ~/.conf, /etc/conf, etc. Actually, now that I've thought of this, I really like it.)

After all, your description above is, at some level, the same as how programs should treat the registry: HKLM is your /etc/conf/conf.conf, and HKCU is ~/.conf/conf.conf. (There isn't really an equivalent to the groups configuration, but that seems more complicated: you need to define an ordering on the groups.) The only difference is that the inheritance should be automatic (and is in your description), while in Windows it has to be written in. (I.e. I think a program would need to explicitly check in one place, then the other to get the settings that way.) Even this isn't true in the case of HKCR, which is built almost exactly as you describe. (It's a union of the appropriate sections from HKCU and HKLM, giving preference to HKCU.)

Re:Architecture and improvements, not religion

[Peaker]

Second, build infrastructure that's obviously missing. There needs to be a central "registry". Let me rephrase that; a central key-value pair repository. Make it and let everybody know. Make sure drivers (X Server / Kernel) can talk to it and applications can talk to it and read/write.

Hey, it already exists! Its called the /etc/yourprogram and ~/.yourprogram directories, and it is far easier to access than the Windows implementation, with simple open/read/write calls!

We don't need more API fragmentation, we need less!
Every new API that does the same thing just makes my tools less useful because there is another aspect of the system that they cannot access.

Not technical, but...

[JosefAssad]

I think (hindsight is always 20 20) that the biggest mistake has been allowing code copyright to disperse and remain untracked. Some contributors are DEAD now. The problem with this is that it makes project relicensing just about impossible, amongst other things. This is one of the reasons why other projects require joint copyright assignment if not straight assignment, such as apache.

Legal environments change; without wanting to descend into a GPLv3 flamewar, I think it's critical that free software projects ensure they have the flexibility to relicense appropriately in the future to adjust.

Linux does not enjoy this distinct advantage.

Where to begin?

[Anonymous Coward]

The kernel needs a disk IO scheduler that doesn't suck.
The kernel needs a consistent driver interface that doesn't require constant driver maintenance between changes. You know, like the memory things that break ati-drivers every kernel.... Or deprecation of old drivers (remember when the Advansys SCSI driver was on its deathbed until someone resurrected it?).

The kernel needs a way to lock hard drives to device nodes in software RAID. (In theory udev should be able to do this, but no one has a way to keep device nodes from sliding around and requiring RAID reconstruction every time a drive is added in a way that disrupts the position of existing nodes. And every now and then your "friends" at gentoo break "baselayout" so RAID initializes before udev runs, etc.)

udev itself is inordinately complicated to script for. In fact, no other OS is this complicated. Plug in a USB stick and it should just be there. Can't there just be one plain text file that maps unique serial #s to device nodes?

Why is glibc so huge?
Why is QT so huge?
Why is X so huge that it had to be broken up into scores of other packages? (And can't we dump support for 1-bit and 4-bit desktops already? And actually get VSYNC? And stop treating 3D like some special case with a funky glx/egl interface? Ever thought that if X11 needs so many new "extensions" to support modern hardware then maybe the protocol isn't "flexible".....it is just broken by design?)

Why are init scripts so dang complicated and different between distros?

Can we dump ALSA? Seriously. That project is so broken the architecture cannot support some SGI hardware, and they certainly haven't fixed ice1724 bugs filed over 3 years ago! I'm told even simple things (like fixing channels that are accidentally swapped in the driver) are difficult to fix.

Why can't network scripts fit into one plain text (not XML!) file? Why do I have to remember where hostname, routing, DNS, and DHCP settings go?

Now how about a consistent API for video, audio, input devices, graphics. Think you SDL? No, how about DirectX. (Except make it not suck either!)

Teach devdelopers their manners. etc.

[chris_sawtell]

Non-technical, but vital none-the-less.

1. Teach developers their manners. Being on many email lists & IRC channels is like being in an amphitheatre full of gladiators. Totally non-productive.

2. Persuade the Kernel-Cabal to be more accepting of some of the near-the-edge patch sets. I'm thinking particularly of the -ck patches, Reiser4, and TuxOnIce. These projects are all pretty good improvements, but do need wider exposure.

3. Purge the fora of all the semi-literate, mis-informed, and out-of-date postings.

4. Set-up some sort of centralised bug clearing house so _all_ the distributors and package authors can access it to fix bugs in a more timely manner. iirc, Mark Shuttleworth suggested something along this line of thought not so long ago. It's important! Let's get it going.

A fourth item

[MMaestro]

A community that doesn't tell the user to RTFM when Microsoft is shoving (seemingly) free, familiar software of most recent version of Windows down everyone's throats.

Re:A fourth item

[Robocoastie]

Actually MMaestro many Linux users are NOT using RTFM as a slang but real advice: http://www.tuxs.org/rtfm.htm there really is a "fine manual". This simple 2 page .pdf (3 counting the title) answers the majority of Linux questions we see over and over again which is why you get answered "RTFM" so much. So seriously, go and Read Tux's Fine Manual.

Re:A fourth item

[networkzombie]

Good morning. If you fine young gentlemen would like supportive, kind and courteous help with your Linux setup and install, we would gladly like to help. But, please, before we supply you with our kind, friendly and courteous support, please head on over to our comprehensive help center which has a friendly, intuitive, and easy to use text version of all the questions we are sick and tired of answering from all you stupid dumbass mutherfucking newbies who need the snot wiped off their faces, god-damn it. Why don't you just read the fucking manual?

Re:A fourth item

[Daengbo]

Beautiful That's about right. After answering the same question for the 100th time, I have less patience. I guess I'm not cut out to be a help desk worker.

I have noticed, though, that both the Ubuntu Forums and the K12OS mailing lists are chocked full of people who seemingly never tire of the same questions. Ubuntu even gives guidance to forum participants that "RTFM" is not an acceptable answer.

Have my list all ready :-)

[Burz]

1. An easy, approachable Hardware Compatibility lookup website. It would consolidate all the compatibility info from kernel & X11 devs, major distros, OEMs and also allow end-users to add their input. FWIW, the HCL at linuxquestions.org is an interesting start but nowhere near exhaustive or current enough to empower Linux users (not hackers) to confidently purchase new equipment.

1a. A certification program for drivers that allows products which meet criteria to bear a special Linux compatibility trademark emblem.

2. Fix the sound architecture. Blocking of sound output still occurs after many years of ALSA. There is no GOOD reason why Harriet shouldn't hear her softphone ring or calendar alarms just because a minimized web page contains a Flash object. Telling her to muck about in the CLI, to buy a pricier multi-channel soundcard, or to learn about sound servers and juggle them is beyond the pale.

3. Create an excellent default IDE for the LSB Desktop environment. The IDE will be geared to target the LSB Desktop spec by default, with desktop applications as the focus. Something you would write a video editor or DVD burner with, not so much a video card or disk driver. GORM on steroids: If it doesn't inspire budding application developers like XCode and Visual Studio then Linux will not inspire application developers to write. Linux will not benefit from many more systems developers at this point because its the apps that matter: The apps sell the platform.

3a. Well-rounded API documentation for the LSB Desktop target, ala MSDN or Apple Developer Connection, eventually integrated with IDE.

4. Enable app developers to become as independent as possible, such that distro managers do not insert themselves between the developers and their users. Distros ought to distribute OS software, and for the most part stay the F*ck away from controlling installation of particular applications. High-level package managers like APT, YUM, etc. should stick to managing (or mangling) the OS dependency tree and leave apps the hell alone! Provide dependency targets in the OS repo like "LSB Desktop", and only one or two others like "Java 6". Then, accept that all the extra stuff you supply on top of LSB is ONLY extra, and will get used when and if the user decides in specific cases.

4a. Ensure those budding app developers can easily share their work with friends and customers. Make appdirs like on OS X and Gobo Linux a standard. Dear God, please.

5. Hacker culture works extremely poorly for application software today. Fund efforts to spread the discipline of user-centered product development. Teach FOSS developers the concepts and ropes of SDLC and Rational Unified Process, with emphasis on adding actor definitions and use-cases to docs and project wikis so that these elements are continually refined and re-thought eventually becoming the centerpiece of requirements. Create use-case instances (scenarios) in close association with unit/app testing scripts. Anything to keep developer minds on the kinds of users and situations the software is meant to satisfy. Encourage budding Business Analysts to do 6-12 month stints with FOSS projects.

6. Create settings persistence (configuration) APIs for crucial system services like X11, Samba, Apache, sound, etc. Get these projects to set and manage their own config files, as no one else seems capable for doing this consistently or well. Maybe when they have to write AND parse their own config data, they will stop creating needlessly bizarre & open-ended formats that umpteen distro tools only understand halfway.

7. Next-generation, object oriented shell based on something like Ruby, Python or even Groovy.

Lastly, all of the above must be in the spirit of fulfilling primary personal computing scenarios like app and driver installation, and configuration of essential services (change screen res, use a network share, etc) in a predictable manner. Unlike MS and Apple, Linux does not yet grok PC land because

Re:Have my list all ready :-)

[philipmather]

Seconded on almost every point but can you imagine what a new generation of OO shell environments would be called? Ruby Shell : rush? Python Shell: pysh? Flex Shell : flesh? ...dear lord no! We have enough naming problems as it is with KDE flinging K's everywhere and the GIMP going strong. 1, 1a, 3a and someone else's point about the /etc directory could almost be bundled by a single application and website. Not sure about 7, shells like bash, csh etc... are either for doing your sys admin work, getting your hands dirty with settings and config or for writting "glue" scripts. In all three of those cases I believe the primary driver should be keeping it simple. Systems Admin is done on a command line because if all their major changes in syntax or features in one version to the next you can still type man "whatever" and spend two minutes figuring it out, change the syntax in a GUI and you have to spend 10 minutes hunting for the bloody button. Changing settings on the command line is again something we shouldn't have to inflict upon people, a "pseudo command line" where a single app presents the /etc/ directory in tree structure on one pane and the textual contents of the selected file in another pane but with each config option replaced with a suitable drop down, radio button or whatever (think about a DTD, XSL, XSLT, conig parsing monster) would probably lower the entry barrier enough to start windows "power users" on their way. Final year degree project anyone, Google code project? My last point, "glue scripts" is where perhaps you were heading. The trouble is you want to be careful that you don't blow the other two functions in the process and then which programming language? It's got to be a universal choice and thats C with csh shirely? Regards, Phil

Re:Have my list all ready :-)

[Arethan]

Well then, with the exception of #7, this hits the nail right on the head. (Even #7 I can see as quite useful.) The problem with Linux has never been lack of applications or lack of effort. It's always been lack of professional polish and lack of consistent direction. I'll agree with most of the other posters that trying to emulate the Windows environment is a pointless endeavor, but Linux will never achieve high desktop penetration without emulating the one thing Windows has always been good at: consistency.

Microsoft ships a new OS about every 5-7 years. That's plenty of time for ISVs to recompile their applications, and fix the few APIs that have changed. Linux, on the other hand, comes out with a fad new distro seemingly every year, all of which want to completely shake everything up with a fresh set of new-hip-fad system libraries that are completely incompatible with any existing systems, and they all release a new version (usually completely overhauled yet again) every 6 months.

If you want Linux to seriously take off for desktop usage, then the parent's list is a great start, but in order to even achieve that with any measure of success, a real standards body would have to be formed, and it would have to (through some fairy tale pixie dust magic) garner the respect and compliance of every major distro on the market. Basically you'd be taking the 10 major distros and compressing them down into a single base image, and the only differences between them would be purely cosmetic, all underlying APIs being completely compatible on a binary level.

If somehow (as a community consisting of everyone always wanting to be the hero, developing the hottest new incompatible doo-dad that does the nifty cool thing everyone will hopefully want so we can all raise ourselves to independent stardom) we can agree on one fucking way to do things for an entire 2 years, the leaps and bounds of progress would be mind blowing. In the end we'd end up tackling pretty much everything on the parent poster's list, plus more. But, unfortunately, it won't happen. The community is far too proud and fickle to ever allow itself to be told how things are supposed to work, even if the standards were created and governed by members of the community itself.

The whole Linux community is built upon the concept of anarchy. With no direction, boundaries, or limitations, you will undoubtedly achieve great progress in coming up with new ideas and ways to do things, and you'll also get 10,000 copy-cats, slight alterations, feature additions that don't belong, and a market so flooded with options that no one can ever decide on which one to use. And yes, you can in fact have too many choices, and Microsoft, Sun, IBM, et al, already know that. (These guys are scared of Linux in the server room, not Linux on the desktop.)

And who's fault is all this chaos? Not Linus' for making Linux, and not even Stallman's for the GNU system (which is where the root of this complete chaos actually exists).
It's your's.
You all took a good idea, and blew it so far out of proportion that you can't even properly work together as a cohesive unit. Everyone wants to be the exotic researcher, and no one wants to be the grunt in the cubicle trenches that actually makes the shit the researchers come up into a usable product. So while I would love to see Linux rise above all of this selfish behavior and actually start to penetrate the desktop domain, I just don't see it happening. Call me a realist, or call this whole post flamebait, offtopic, whatever, but I'm going to keep using Linux for servers, and Windows, OSX, and Solaris for desktops. From the way things are now, how they've always been, and the lack of any real change on the horizon, doing anything else is just stupid.

Re:Have my list all ready :-)

[Anonymous Coward]

4a. Ensure those budding app developers can easily share their work with friends and customers. Make appdirs like on OS X and Gobo Linux a standard. Dear God, please. Application virtualization. It's a killer app for whatever platform provides it to the masses. Apple doesn't count; Apple isn't about the masses, by choice.

If 'Linux' provided uniform, bulletproof app virtualization across all mainstream distros it would fundamentally improve application software. I speculate that ultimately the market will force this on application software. Whoever makes it easy first wins.

I dunno where this story came from. Perhaps there's a wealthy entrepreneur out there with the means to fund something and is seeking ideas. In case you're listening...

A moderate, well bounded challenge is X over RDP. It's been done some; incomplete in extant open source implementations and some (costly) commercial equivalents. What I want is a simple, free, ubiquitous X server that reflects the display+audio to RDP clients. Native remote X is inefficient, intolerant of latency, still lacks the integral support for the simple concept of disconnect/reconnect, and usually requires some 'tunnel' through which all it's messy ports and connections can be multiplexed. VNC (optimized screen scraping) is too slow (by design) for general use. X over RDP is the answer. If it existed it would be understood that nearly all hosted Linux instances would run this software to provide remote GUI service. Very wide use. The kind of thing that a year later you wonder how the hell anyone did without.

Excellent RDP clients already exist for all relevent platforms. Linux lacks a server.

Re:Have my list all ready :-)

[dbcad7]

4. Enable app developers to become as independent as possible, such that distro managers do not insert themselves between the developers and their users. Distros ought to distribute OS software, and for the most part stay the F*ck away from controlling installation of particular applications. High-level package managers like APT, YUM, etc. should stick to managing (or mangling) the OS dependency tree and leave apps the hell alone! Provide dependency targets in the OS repo like "LSB Desktop", and only one or two others like "Java 6". Then, accept that all the extra stuff you supply on top of LSB is ONLY extra, and will get used when and if the user decides in specific cases.

Your as free as you want to be.. fact is so free you can make your own distro that installs applications the way you want to install them. I think maybe you want something with no package management.. good luck with that.. I think APT is just fine thank you. I like that there are distro managers inserting themselves between developers and users ... deal with it.. If you want to be "outside the box" you still can.. but it shouldn't be standard... sheesh.

Re:Have my list all ready :-)

[Burz]

Thanks for your response about OO shells. I think that the MS 'monad' idea will become extremely attractive, and whereas most of my post concerns end-users and budding developers, there will have to be some work even on the sysadmin front.

As it is now, the *nix (incl Linux) domain is turning off a lot of young people because of silly things like commands and switches that have no rhyme/reason, highly-unique and nasty configs like httpd.conf and xorg.conf, and having to learn vi even when they are told that webmin and the desktop environment will (mostly) take care of conf files.

An OO shell would permit options completion and promote greater consistency with option switches, making usage a great deal simpler. In addition, add an editor like 'nano' to the standard so that admins can choose between vi and something more straightforward.

Re:Have my list all ready :-)

[Burz]

Your as free as you want to be.. fact is so free you can make your own distro that installs applications the way you want to install them.

This is a very old and tiresome mantra.

Time and effort are not free.

What you call "no package management", I call package management without hairy and bug-ridden central databases. No user should have to wrestle with a dependency database to install an application of their choice, much less a repository that distributes apps that are poorly understood and rarely used by the ones compiling configuring and packaging those apps. It is one of the main reasons why Linux apps are feature-poor, as devs have learned to dread the hassles of functionality that breaks across distros.

Not only should there BE a box (a Linux standard), but it is time to build a better one that means something to average users.

Re:Have my list all ready :-)

[dbcad7]

No user should have to wrestle with a dependency database to install an application of their choice

That's exactly why there are package managers.

much less a repository that distributes apps that are poorly understood and rarely used by the ones compiling configuring and packaging those apps

So you would have them distribute these bad boys another way ?, with out any control or testing to see if they are broken ?

It is one of the main reasons why Linux apps are feature-poor, as devs have learned to dread the hassles of functionality that breaks across distros.

So it's your feeling that apps are feature poor, and that the package management system is responsible ? ... again, the package management system is supposed to be there to MAINTAIN dependencies. It (APT at least) works pretty darn well.

Re:Have my list all ready :-)

[krondell]

I would like to add to your 3rd point. I'm a windows application developer by day, so of course I use visual studio and MFC. I've also done quite a bit of linux server side app developement, and I would really like to do the equivalent app development on linux, but I can't find a tool that does what I want. I know some people are going to say I should look at GTK+ and Glade. I've spent hours trying to get hello world generated by glade to compile. I want to write applications not look for the latest TIFF or PNG library source on the internet. Somebody needs to build an IDE and windowing framework that rival Visual Studio and MFC in quality, that you can just install, and it just works. If you really wanted to get me excited, write something that allows my MFC code to compile as native X applications. Think about how the number of applications and application developers would explode if you did that. Damn it, just rip it off! It would make the transition from Windows to linux more inviting for businesses, users, and software developers.

Re:Have my list all ready :-)

[anon mouse-cow-aard]

On point 4 you are without a scintilla of clue.

The problem is as follows:

You package has dependencies. It requires the version of glibc and a dozen other libraries you built it with. Do you talk to the developers of glibc and the dozen other libraries? if they want to upgrade their libraries what do you do?

If you want to be independent then you take a copy of glibc that you used to build your app. you install it it in its own appdir so your app can use it. You do that for each of your dozen dependencies. Then you are independent.

Benefits:
-- your app is independent.

costs:
-- no shared library usage because every application is going to want its own copy of twelve libraries.
-- massive disk space cost because every application is going to ...
-- large number of copies of the same f2#@%#! library on the system is usually pointless.
-- this is how windows works, and their term for this situation is DLL hell because you have no clue
which version of a particular DLL is being used by any application.

Now lets say there is a security problem in glibc. You can either:
a) tell your users to screw off, they paid for software and you gave it to them, two year old glibc and all. Security is somebody elses problem.
b) be out of business, in which case refer ex-customers to a)
c) have a server available for your customers to download patches from. Those patches will include patches to all the libraries which you depend on. So you have to validate and produce packages for all
of the twelve libraries which you depend on. You maintain complete packaging, including all dependencies, for each and every distribution you support. You are able to, on request, verify which packages need updating, and perform coherent updates (perhaps multiple dependencies have to be updated together with your application in order for it all to work together.)

Congratulations! you are now running a repository with dependency management. One step further and you can have your own distro.

Of course, you could avoid that by simply listing in the requirements for the software all the library versions they must have (which is fine, unless they feel the need to upgrade their OS at some point, and your app breaks.) but then that forces the user to use repository tools to satisfy dependencies... hmm... why not automate that by having the application know what its dependencies are! ... Again, you have just invented apt-get/aptitude/yum.

The alternative is to cultivate package maintainers for various distros, let them maintain the packages for the distributions, let them understand the weirdnesses of each one and help you with packaging issues.
Use them as filters for bug reports, so that if the problem is distro specific, they look at it first.

All applications use the same glibc (and any other common dependency) saving huge amounts of disk and memory space. Now when a new glibc patch comes out, the build cluster will automatically re-compile your package if needed and any other dependencies against it, and your clients will automatically received (a single copy of) the patched glibc, and the patched application, with zero effort on your part. If you are getting weird problem reports from a particular distro, talk to the packager and try to work it out together. build a relationship, dont swear about his existence. You are wrong, life is not simple, and swearing at people that reveal that it is more complicated than you would like to believe is stupid.

The second reason you are lacking in vision is that you pretend that there is some bright line between the OS and applications. Get a f$#% clue! to the guy writing BIND or djbdns, DNS is an application. in the 1980s, berkeley networking was an optional application stack on

Re:Have my list all ready :-)

[Burz]

But how Linux distros use package managers is a big part of problem, as I pointed out. Part of the reason is that every complex Linux program has a complex set of dependencies, nitpicking about this-and-that component within what should be the OS.

OS X has a package manager, and while it is seldom used, the dependency checks are mostly limited to "OS X v 10.3.9 and above", perhaps the Java and Quicktime versions installed. You don't get an app installation failure that tells the user their OS doesn't have necessary library versions in a piecemeal fashion. "Your repo doesn't have a recent-enough sound server, or xml parsing library or messaging bus" is gibberish to most people: You can't possibly expect it to be meaningful to them. But Linux coders have been accultured to foisting this experience on their users, whereas an OS X coder will try to hit two or three well-defined environments, i.e. the versions of OS X, and include what those environments don't offer into the Appdir.

So you would have them distribute these bad boys another way ?, with out any control or testing to see if they are broken ?

Are you kidding? Testing is the author's responsibility, and for personal computing, distribution should be also. Distro managers can't be expected to know the carefully considered use cases and scenarios that thousands of apps were written to fulfill, so their testing and distributing role should be minimal-to-none.

Fork it.

[chris_sawtell]

And turn it into a micro-kernel. :-)

Refocus Linux:

[Anonymous Coward]

1. Close down linux.
2. Switch to Vista and have fun playing crysis, cod4, hl2:eps2, teamfortress 2, etc.

Bye idiots.

Focus on education and after this, documentation

[cdn-programmer]

IMHO the toughest job facing the OSS community is education: teaching, learning, and how to document.

This is compounded by the issue that most developers do not find documentation fun. If the common perception that geeks tend to be nerdy and poor at communication is true, then we have a triple whammy. This is one reason I say documentation and communication and education is our collective biggest failing.

The learning curves for _any_ of our packages are steep. SysAdmins rejoice in the job security they perceive they gain as their expertize for apache, mysql, postfix, postgreSQL and so forth increases. The thing is each package has so many options that it takes forever to learn how to set them up. At last count Debian boasted over 30,000 packages available. How is one suppose to even know what a small percentage of these packages do? That is much less than to learn how to install, support and maintain them?

But this is just the systems administration arena. The API's and programming is an order of magnitude more difficult to keep up with.

Then the documentation. To use WxWidgets for instance I am faced with over 3,000 pages of main manuals, I need to decide if I use DialogBlocks or CodeBlocks or neither. I need to figure out what each does and what each doesn't, and after I buy Julian Smart's book - its another over 500 pages to read. In spite of the fact he's written DialogBlocks there is no useful information on same in his book. Thanx.

This is only one (1) package. I have not addressed version differences and library dependencies and so forth. I have not considered the issues of limitations and bugs.

To keep up is typically information overload to the gawd-zillion'th degree.

---

M$ recognized this and attempted a solution. From what I can tell in around w95 they pulled all the error messages out of the system. I experienced the great joy of accidently turning off the external SCSI hard drive on a W95 computer while the system was accessing the disk... reading it actually. No error message was reported. We got what looked like "END OF FILE". This was M$ code reading the disk.

Then on another occasion I noted a networking message from NT4.0 had the exact same text as from OS/2. The error number in NT4.0 was missing. Everything else was the same. On a hunch I looked up the message in OS/2 and lo and behold the error number lead to the issue at hand. Of course NT4.0 was no help at all because this information had been removed.

Either it was removed or never put in. I dunno. What I do know is that the systems ability to correctly diagnose was hamstrung.

So what do we have in the OSS world?

1) volumes of crappy documentation layered on more volumes of poorly organized documentation.

2) When problems are found and corrected - no good method exists to upgrade the docs.

Here is an example. Many years ago I ran into a sound configuration issue in Debian Woody. This had to do with esoteric issues of generic SCSI drivers and bad permissions and so forth. I ended up posting in SourceForge a complete description of the problem and how to walk through it and fix it.

Two (2) years later none of this information had been disseminated through the documentation of the package at hand where I had discovered it. Debian was still misconfigured. People were still coming into IRC pleading for assistance on how to get the software running (It was GRIP as I recall).

---

This is just terrible performance and we are not getting much better at it.

There are several websites of documentation. SourceForge does this. IMHO they do it poorly. There are many wiki's dedicated to various packages. Nothing is coordinated. The man and info pages I have in my latest system are still the first place I would like to look for information and they are basically just as bad now as they were in 1997. Probably these documentation sources have not been updated much since 1997. Why not? If there is new

Re:start over from the gound up

[Jeremiah Cornelius]

Here's a man who understands modular design.

Outlook/Exchange Server

[blackpaw]

Clone both at the API and Functionality levels. A Linux client that could talk to Exchange server and a Linux server that could talk to outlook clients would be an office killer.

Re:start over from the gound up

[Solra Bizna]

As far as I can tell, your argument is:

The Linux kernel does not support many new wireless cards.

Therefore, the Linux kernel is a heap of feces and needs to be completely rewritten.

Sorry, but that seems a little... extreme.

-:sigma.SB

Microsoft Exchange

[dwater]

> 2) Better integration of Linux with Microsoft Exchange servers and Windows domains.

Finally, someone saying what I'm thinking...

Of course, wrt Exchange, I'd prefer something completely independent of Microsoft, but functionally equivalent.

Re:Microsoft Exchange

[RidiculousPie]

Openchange is the project you are looking for.

Re:Microsoft Exchange

[dwater]

Well, that does indeed sound promising.

Less kernel development, more UI

[Psychor]

It seems to me that kernel development is already well served by many capable and active developers (I guess this is the 'glamorous' area of Linux dev). If anything, I think there should be less emphasis on adding major new features to the kernel itself and more on ironing out the few existing bugs and having a slower release schedule with better testing.

Most of the major issues are with the desktop and with individual distributions, some areas that I think could use development (many of which seem to suffer a lack of interested developers currently): -

• A lot more focus is needed on ease of installing. Most of the latest generation of graphical distribution installers are great when they work, but in a large percentage of cases fail on modern hardware unless particular command line options are invoked (noacpi, nodmraid etc.) Another example is that the majority of installers won't run correctly in graphical mode on the Geforce 8800 series despite them having been common for almost a year. At the very least fallback to command line troubleshooting options, text mode etc. should be automatic and specialist user knowledge should not be assumed. Also the majority of installers do not offer the option to resize an existing Windows partition to create install space, a common need among novice Linux desktop users.
• Better integrated control tools, fewer obscure configuration files. The Linux command line is one of the OS's great strengths and I don't advocate dropping/obscuring it like some posters. However, all configuration affecting the desktop system/UI should be capable of being accomplished within the GUI. Few if any distributions provide UI tools capable of, for example, installing/managing ATI/Nvidia display drivers, or connecting to a WPA encrypted wireless network. Also many distributions provide multiple GUI tools for managing the same things, e.g. sound settings. This is confusing and inconsistant as well as a frequent cause of conflicts.
• Better selection of software to include in distributions. For example many common desktop distributions include between two and half a dozen different media player applications in their standard package. It would be better to provide one player which is well integrated into the desktop environment and comes sanely configured.
• By the same token decide where features of software are surplus to requirements and streamline it. For example Konqueror is a fully featured web browser with script support etc. but many distributions including it instead install Firefox as a primary web browser. It would be preferable to standardise by major distributions either using Konqueror as a primary web browser, or encouraging the developers to drop this functionality completely and focus on making a more streamlined file browser.
• Some desktop software lacks features which users 'expect' or they are not easily accessible. For example I couldn't make either KDE or Gnome's file browser display a multicolumn list of files like the 'list' view option in Windows, which in my opinion is the fastest way to browse a directory with a very large number of files in. Another example is that the Linux version of VLC doesn't appear to have an option to open files in an existing window if a video is already playing, and always spawns a new one. Note that these are just examples and can quite possibly be accomplished somehow - my point is that these options are not easy accessible from the GUI.
• Poorly chosen or confusing defaults and UI choices. For example some distributions default different audio file types to different media players for no compelling reason (support by both players). Another popular example is hotkeys, even for something as simple as copy/paste, some GUI apps follow the ctrl-c ctrl-v Windows style convention, some use the select/middle-mouse console style convention, and some support both (usually clumsily). At least within a single distribution, these conventions should be standardised.
• UI speed/consistancy issues. The mess

Code clean-up.

[seebs]

Really, that's it.

Let me give an example. I recently stumbled across a chunk of code which "used to work" in 2.6.14, but which had compile errors in 2.6.21, claiming that it was an error to declare an array of negative size. WTF?

Tracing through a series of macros (some with all-lowercase names, in violation of the standard C convention of the last thirty years), I eventually found a complicated expression in a macro expansion, to the effect of BUILD_BUG_ON_ZERO(various && conditions || otherconditions).

That, in turn, is defined to, as you might easily expect, declare an array whose size is negative if and only if the argument evaluates to non-zero.

This has a number of problems.

1. BUILD_BUG_ON_ZERO might be taken to indicate that it is a bug for the condition to be zero. No. It's the same as BUILD_BUG_ON(condition), except that it yields the value (size_t) 0 if the condition is zero, instead of not yielding a value. The name is comprehensible if you're familiar with it, but frankly ill-chosen.
2. Using this macro inside a macro expanded from another macro results in a seriously opaque chunk of code. The error message one gets (about the array declaration) is on a line of code that doesn't declare any arrays, and has no semantic reason to be declaring one. There's nothing to tell you what happened without you carefully following a chain of expansions. The error message is useless.

Really, the kernel is mostly doing pretty well these days. I don't expect to have to reboot a Linux machine because "maybe something got corrupted" or whatever.

But, as someone who has to READ this code, I wish it were written more clearly and better documented. Obviously, this is something any or all of us can work on.

Re:Code clean-up.

[EvanED]

Tracing through a series of macros (some with all-lowercase names, in violation of the standard C convention of the last thirty years)...

I once spent 20 minutes swearing at GCC for giving me a nonsensical error message that I ultimately found out was caused by me using "current" as a local variable name, which conflicted with the "current" CPP macro.

Every now and again there's a macro that (arguably) shouldn't be all caps, but "current"? Give me a break. I was tempted to submit a patch renaming it to CURRENT_I_HATE_YOU_ALL everywhere it's used, but of course never did it.

9 million more potentially users

[Unnamed Chickenheart]

...if someone somehow could make Blizzard agree to port WoW to Linux.

Re:9 million more potentially users

[EEBaum]

If that someone somehow could make Linux users agree to pay for software, maybe it could happen. Until then, don't hold your breath.

Free Software To-Do list:

[Verte]

If I had millions in the bank and a large team of clever, experienced programmers, I'd point them at topics like:

1. X is fat. X.org is working on Modularisation now, but there's a while to go. It could sure use some more programmers. And then, I bet it could be a lot more useful too. Quartz showed us that X could have more features, I think we could push that idea to a new level.
2. Firefox. I heard someone speaking about having different processes for different second level domains, and different threads for each tab within, as well as one for the main gui of course, and one for the download box. I think that's a worthy cause, among other things. There are a lot of other usability things that need to be looked at, and a better caching algorithm.
3. The Hurd and the microkernels that it runs on. The feature list and the ideas coming out of those development circles blow my mind. Personally, I think this is where I'm going to focus my energy soon.
4. Various usability tweaks, such as unification of GUI, CLI, and keybinding.
5. Next-gen file systems!
6. ...
7. Profit!

I'm sure there is a lot more..

Refocus?

[Anonymous Coward]

Refocus? Was it ever focused?

One word..

[beodd]

One word, virtualization. Having good virtualization built in to the kernel. With that Linux could do anything.

Beodd

5.1 optical audio

[l0b0]

This has been my biggest beef with /any/ Linux distribution the last five years or so - S/PDIF, IEC958, optical, or whatever you want to call it is terribly supported. Ubuntu's audio device menu is a huge improvement - At least I can get stereo digital without Googling. No luck with surround there, though.

A bit of biz, a bit of fun

[dorath]

First, I'd like to see something that really takes on Outlook/Exchange.

Second, I'd like to see something along the lines of OpenX: a competitor for DirectX. Sure openGL competes with Direct3d, and there's openAL for DirectSound... But why not bundle up a bunch of open goodness the way DirectX packages together graphics, sound, networking, and input? Wrap it all up in a nice tight IDE and what've you got? Something that'll write games for Linux, Mac, Windows, Wii, and PS3. Compare that to DirectX, which does Windows and xBox, and frankly the only hurting party is the xBox. Sure, sure, it's more complicated than that, but damnit, that's what I want to see.

Re:A bit of biz, a bit of fun

[Eli Gottlieb]

Linux multimedia dev, meet Simple Direct-media Layer. It's been done, it's used, it works, and it's open. The trouble lies in the politics of getting developers to port their games and engines over to SDL from DirectX and other proprietary libraries.

Development will grow with the userbase

[distantbody]

premises:Linux development will speed up as the userbase grows. I think the greatest return on effort would be to focus on ease-of-use, lowering the jargon barrier, explain the FLOSS philosophy, including its advantages AND disadvantages WITHOUT sounding like a Nigerian 419 scam.

Re: Two Options better than 1 or 8?

[TaoPhoenix]

There may be a hint of an idea here.

We all rail at Microsoft and possibly Apple for their Monolith approach to certain things. One famous argument with at least a few grains of truth is that "eight-ish distros confuse potential new users". They certainly confuse me, and I'm even expecting a certain amount of roller-coaster riding.

The previous poster may be right that if forced to choose ONE desktop, we would have ... none at all because of the disagreements. But maybe two of everything is just the right balance for a lot of areas. Two desktops - one ultra-protective of new users, one *promoted* as more advanced that the new user can glance at and think "maybe I can use that next year".

We might be there with the Office apps - some combo of Open Office and your choice of Google Apps or whichever other one eventually becomes emergent. In Browser-land, it might also be there - Firefox/IceWeasel and Opera.

We absolutely have to have a distro that nearly everyone wants to stand behind to present to work environments. I'm absolutely for the experimentation bit at home, but for work, business schools have been teaching standardization for some 60 years now. Apparently Dell thought about this and chose uBuntu. Take your pick of your favorite Newbie-Friendly second distro. Then mark all the others as "For specific purposes" and I think the public will get that. I have a copy of u. Drake that looks benign enough.

Proper core dump

[br00tus]

I have Solaris boxes where, if they crash, a core dump is generated on the file system, and I can examine it and send it to Sun as well.

Linux does not have anything like that. I need to have my machine firmware as well as Red Hat version constantly patched, at a level unneeded with Solaris and Sun. Then I need to go through a rigamarole to set up netdump or the like. Even after that Red Hat or Dell/HP/whoever are often clueless, especially the hardware people. With Sun I can quickly examine the core dump, replace a component, and feel safe I've fixed a problem, when Red Hat is a different story, usually I tag the host as unreliable and that's that, we don't really have the time or luxury to figure out why a server running Red Hat crashed. Anyhow, whatever the problems of working with different components is, or the desire on the part of others perhaps that the corporation I work for devote more resources to this over what Solaris needs, with Sun you get an easy, enterprise solution for deducing why a system crashed, and I feel somewhat confident people at Sun are examining these for consistent problems, with Linux it is all over the place and I suspect most people just usually throw up their hands like I do when a Linux system crashes.

Make Automatix obsolete.

[hot soldering iron]

Many current distros seem to be getting hosed by users applying Automatix to install various programs and A/V codex. It's just a series of scripts, written to help automate the process, but it does have weaknesses and breaks sometimes.

Yes, I know that the distros (Ubuntu among them), can't ship with these because they are proprietary. Someone here already stated,"I don't care." Users DON'T CARE why you can't do something, they just know you failed them (their pov). I think that this also ties into the previously mentioned Uniform Installation Process. If we had one, an Automatix-type script would be so simple that it shouldn't be able to break.

User transparent multi-threading/clustering might be an interesting pond to throw a rock into. I don't care who you are, eventually you'll find yourself wishing your hardware was faster, and wondering what you'll do with your current machine. I think that a lot of people would love to be able to throw a liveCD or a minimal install on a series of networked machines, and have their current "Master" Desktop unit become noticeably faster, and the storage space grow via networked RAID, like Gmail! It should improve the performance of Virtualized Systems and concurrent processes, like running a search for a datafile, improved rendering and animation of your 3D desktop spaces, seeding a torrent, serving a Doom LAN-party, and playing Jethro Tull perfectly in the background.
Okay, the datafile search can stay slow.
Building a Beowulf Cluster back-end to a desktop shouldn't require a BS in CS. With the new multi-core CPUs head for the street soon, I'd imagine that a lot of the same technology will be used.

These are just off-the-cuff opinions of a hobbyist, and I'm sure that none of this is easy.

OpenOffice.Samba .Wireless

[baggins2001]

The one app I would focus my developers on is OpenOffice. This is so close to be something that could actually replace MS Office it's a crying shame. The one piece of functionality that they are missing is file sharing. Over and over again I see where the chance of an office implementing OpenOffice just falls, because you can't share documents(multiple users have the same document open and can edit).

The other is wireless, this is actually where linux could still step ahead while still being so far behind. The implementation for secure wireless on windows systems are weak. Windows can be integrated with other systems to give a wireless system, but so can Linux and if more people would write good documents on how to do it with Linux then it would be seen as a solution. Try to read how to integrate Radius with some of the systems out there. The multiple possibilies make your head spin. If someone would show just one right way

Second Application or service would be Samba. They are doing an excellent job, but if they could use the help I would give it to them to speed up the integration of ldap with Samba domain servers. Samba4 is supposed to have it but will it be fine and polished when they get there, will it have the network management capabilities we are needing for our growing environments and remote management.

UTF-8 and Clipboard

[Anonymous Coward]

1. Make UTF-8 work on EVERYTHING as the default

2. Integrate a GOOD clipboard in every application where it makes sense. Hint: Try to copy tables from a web page into a database application.

This has gone over everyone's heads

[dbIII]

Linux is the kernel. Window subsystems, video codecs, even command line shells are applications.

Re:This has gone over everyone's heads

[bytesmythe]

The original topic post says kernel AND applications. Linux (the kernel) will never be adopted wide-scale unless "linux" (i.e., distros based on the linux kernel) manage to get rid of the annoying problems that hinder adoption.

There are a lot of people who try linux and dump it because some little feature (say, sound) either doesn't work, or even better, breaks during an upgrade for no reason. Until "linux" in the broader sense gets all this stuff fixed, regular users won't stick with it.

Re:This has gone over everyone's heads

[Anonymous Coward]

assclown read the summary kthx.

Three Things

[RAMMS EIN]

Three things, in the order I thought of them:

1. Security. Security is probably the largest actualy _problem_ facing the IT industry. Where in the past security issues could be ignored as chances were security breaches wouldn't happen to _you_, nowadays the situation is different. We've gotten to the point where everything that _can_ be exploited _will_ be exploited. If they're not after your sensitive information, they'll break into your computer to use it for spamming, hosting malware, staging attacks on other systems, or any number of other uses you don't want your computer to be put to.

Linux (and open source in general) has a reputation for being more secure than certain alternatives, but I feel this is largely undeserved. Security is not getting the attention it demands, development continues to be done largely in unsafe languages, and plenty of vulnerabilities are published on a regular basis. If you watch your log files, you will see that attacks (and I'm not just talking mindless attacks aimed at MS software) are also performed on a regular basis, too. It's only a matter of time before the detractors of Linux and open source will be all over the media proclaiming the failures of the open source model, with high-profile compromises of Linux systems for ammo.

2. Modularity. If you've ever tried your hand at developing kernel modules, or even if you have enough experience compiling kernels, you will know that Linux is a mess. Lots of modules link against symbols in other modules, APIs keep changing, and, generally, what works today might work tomorrow, but the only thing that is certain is that it will break at some point. And documentation? Well, it may be there, but if it is, it's often out of date and incorrect.

Some efforts are being made to modularize Linux and provide stable APIs (see, for example, FUSE). This is good. We need more of that.

Ideally, I'd like there to be a single, small and stable interface that modules can use to get all their work done. By using this interface (instead of directly linking to kernel symbols, as is currently common), modules can truly isolate themselves from the rest of Linux and not have to be rewritten and recompiled quite as often. It would also open the door to modules being implemented in userspace and in different (safe!) programming languages. Perhaps, one day, we can apt-get install the new wireless driver, and know that it doesn't contain any buffer overrun vulnerabilities...

3. Parallelism. I believe we are at the brink of a wave of innovation in the realm of parallel computing. A lot of research has been done already, but most programming is still stuck in the sequential imperative paradigm. As far as I know, Linux does a pretty decent job, compared to the competition, when it comes to parallelism, but there is still a lot of room for improvement, too. I expect most of the gain to actually come from applications and not from the kernel, but the kernel can help, too...by better exploiting parallelism itself, and also by providing some necessary support and information to userland. Computers with multiple cores will soon be everywhere...let's take advantage of this to the maximum extent feasible!

Explanation requested

[kanweg]

I understand that to resize, you need to need to remove 1 pixel over all the height (or width) of a picture. I understand that doing that as a row/column is not a good idea, and that there is nothing wrong with doing it as a path. I also get that there are paths that are more suitable for removal than others.

Now, how to determine the paths? The authors are talking about energy levels, which I don't get, and gradients. What do they mean by that, and how do they determine it? They talk about an 8-point connection (which, I presume, is the 8 pixels surrounding a non-edge pixel). Are they looking for a path having the least amount of intensity difference? (If so, wouldn't a picture with a lightning bolt be very susceptible to have the bolt removed?). My problem with that is that it is very likely that part of a path (say, the lower half of a vertical path) that is optimum, is also the optimum path for another path starting somewhere else. So, like a capital Y, two starting points (the top tips of the Y), end up following the same path. Now, if I remove one path to shrink the picture, the second path is no longer available (part of it is MIA). So, paths are not allowed to overlap, I guess. That would mean that not too many paths can be available on a picture, yet the pictures they show displaying the paths are fully coloured.

To explain how you run out of paths soon: Suppose it does work like you cannot cross paths (or have overlap), once you have two paths and one crossing that, you run out of paths. Example, the capitla N, with two vertical paths and one going from topleft to bottom right. For none of the starting pixels at the top between the vertical legs of the N, a path is available anymore (once the \ path is established). So, should there be an algorithm (weighting factor) to make sure we have enough paths by detecting such nasty inclined paths?

Bert

Drivers

[xophos]

I would put all those resources into driver development. Especially the GPL graphics driver Projects for NVIDIA and ATI cards.
But any other widespread comodity hardware that needs to use binary crutches at the moment to be useful under Linux needs help too.
I think FOSS Comunity has already shown that Applications and Desktop Experience need no extra help.
With driver development the Situation is a little different.
It's hard to do in your spare time.
There are not so many People who have the needed skills.
Reverse engineering takes loads time and resources.
The Developers need to have the Hardware to make drivers for it.

Areas of focus

[Anonymous Coward]

Slanted towards gnome, but nonetheless.
1. OpenOffice base: Make it as good as ms-access, as easy to export to the web, and as versatile of a reporting tool.
2. Get Out of tree drivers in tree, especially in the video area. LIRC should be in mainline (Or something that handles all of it's drivers) same for spca5xx.
3. Video and voice IM chatting (There's no reason for this to be missing, as all of the support is coded, it's only integration work).
4. Flash reverse engineering: This is especially critical for 64 bit and ppc.
5. Webkit everywhere. Gnome uses 3 html libraries, kde still doesn't use it. From what I understand embedding mozilla is a pain, as you have to send everything beyond the simple stuff with xpcom. Everyone seems to think webkit takes care of it. I'd just be happy with 3 less libraries.
6. DVD/VCD playing in totem for unencrypted discs (With menu support) Right now I have to install xine for dvds only. The elements are in gstreamer (though not maintained) This is just another duplication.
7. Integrate openmoko with linux as much as windows mobile is with xp (Playlist syncing, drive mounting, contact/dates/email/bookmark syncing)
8. Remove Bonobo from gnome (From what I see nobody uses it as it was intended and there are 3 people who understand it, and with this status it does the opposite of what it's supposed to).
9. iDVD: Make qdvdauthor or dvdstyler as easy as idvd.
10. garageband: make jokosher or the beast as easy as garageband.
11: Easy: Give me different wallpapers on different desktops.
12: Make the gnome dock cool again, give me avant window navigator or gimmie by default.
13: Integrate im/irc/email/contacts. Make it work. Right now telepathy has 80% coverage of protocols, and crashes for me so I can't even try the integration w/evolution.

filesystem, init, xwindows, standard GUI conf tool

[MLS100]

1. Replace the ancient and retarded POSIX filesystem.

This means a filesystem with organized folder hierarchy.

Jesus this is something that has been sorely missed for ages. If I install Firefox, the files installed by the package are strewn all over my system. Wheres the config file for that application I just installed? Is it in /etc? No. /etc/appname? No. /opt? No. /opt/appname? No... locate appname.conf, hmm nothing.. What about the actual application itself? Is it in /bin? /sbin? /usr/bin? /usr/sbin? /opt? /opt/appname? /usr/local/bin? What about docs? What about all the other crap that came with it, do I have any chance of finding these files?

How about /applications/appname, always. Wheres the config files for Appname? In /applications/appname/config of course! What about docs, in /applications/appname/docs of course! Want to remove appname? rm -r /applications/appname, tada! How do I know what is installed? ls /applications . What about that app that says it needs whatever library, you can't have the system searching through all those app folders for shared libraries right?

Right, so lets use these fancy little things called symbolic links, and link them in a shared library directory, but keep the files themselves in the /applications/appname/libraries directory where they belong, but an easy list can be checked with ls /applications/libraries .

By golly Holmes, this means that all those man hours spent on fancy packages and management thereof were a total waste when all you needed was a person with a brain setting up the folder hierarchy?

Indeed, Watson, indeed... and I'll go you one further, this setup works no matter the distribution, because blimey, they ALL support folders! Who knew?!

2. Init

Lets face it, init needs a major rework. It is slow, clunky, and annoying to manage.

3. XWindows

I don't think much is needed to be said here that hasn't been beaten to death already, XWindows needs to be more than graphics slapped on top of unix to be a viable candidate for GUI management.

4. Standardized GUI config tools.

This means that Fedora control center looks like Ubuntu control center looks like every other control center so if someone needs to find something, it is always in the same spot and configurable. Pretty much every single conf file I've seen is an easy GUI conversion, varname=(true/false) -> checkbox/dropdown, varname=(one of a few available values) -> dropdown, varname=(some string) -> input box, and so on. Define sections as tabs instead, and bam you have organization.

--

The bottom line is this, this is the 21st century, you've put thousands of coats of paint on your 1974 Unixmobile to make it look new, but on the inside it is still running the same decades old technology it has always run.

This whole post assumes Linux even wants desktop share, if it doesn't then by all means keep working on getting 10 more points on Bullshitmark 2007.

OpenOffice,Samba,Wireless

[baggins2001]

OpenOffice still needs file sharing(multiple people can access and edit the file at the same time). In many offices this is a show stopper.
Samba, they have done a great job, but I would throw my support into making sure that Samba4 is integrated or can be integrated with LDAP so that management of growing and more complex infrastructures can be handled. Also that it makes remote management of networks easier.
Wireless, wireless is like the albatross hanging around the neck of most linux users and advocates. You can't just go and get any wireless card and integration into a secure environment hasn't been easy. How do you integrate your LDAP or your Radius with some of these wireless authentication products? Radius is so flexible it's mind boggling. LDAP is just complex. I would first take these wireless products and show you how to set up the servers and clients so that they are secure. Then how to group users so that you can limit/give them permissions. I have yet to see a linux wireless card that can interact with AP and jump channel to a less active AP in the same area. (I would think it is out there somewhere I just haven't seen it).
3D effects are really cool, but >80% of the users don't know how to use them and don't care to.
But, hey 6 years ago I thought virtual desktops were just neat. 6 months later I don't know how I lived without them. Still drives me nuts on a windows system when I want another desktop.

Sounds like "no sun4m" bmc's view on Solaris....

[sethstorm]

First, I'd give up the holy war some kernel developers are having with proprietary drivers. Drivers are proprietary because some companies want to release proprietary drivers. Live with it. Stop fighting it and try to make everyone's experience better.

Then you have the HCL problem ala sun4m/sbus and opensolaris allowed to rear its head. Tons of Sparcstations, little documentation. They can have 'em if they want to provide documentation to the point that'd make even Theo's group happy on how to write their own for any other platform. Until then, that'd make you look like Sun "Sparcstations don't exist" Microsystems.


95%+ of Linux users today (a much MUCH different demographic than from 10 years ago) do not care about open drivers as much as working hardware. If you aren't going to help make people's experience with the OS better, then in the words of Ludacris: "move, get out the way, get out the way".

That includes those who get in the way of hardware support. Thus the holy wars will continue.

refocus development with connectivity

[idanity]

ability to setup wireless connections w/laptops (at install) i cannot even remember how many times ive used a resque (linux based) live cd to help re-claim friends files (before they re-installed windows). and only 1 time i was able to connect right away (to the internet). now, if i was always able to connect to the internet, easily and fast, i would have converted all of those friends to linux based os's. however, many times i could NOT ever connect to any network. that is the biggest drawback (even now)as soo many use the pc's just for email. so, in my order of importance, connectivity-#1 connecting even at the install, or on a live cd is so great (when it works) it should be fast, easy and powerful enough to just be able to open a browser and have it work.

retarded POSIX filesystem - /opt

[krischik]

For your first suggestiong: I belive that was what /opt was about. Only distros move away from /opt again. A few releases about SuSE hat /opt/gnome for gnome - now it is gone angain and gnome move to /usr. Currently KDE 3 installes in /opt/kde3 - but the KDE 4 beta installes to /usr.

Personaly I think it is a big mistake. /opt/application is a much cleaner system then cluttered down /usr

Martin

Re:The defacto desktop distro for the general publ

[JoshJ]

You have to add the Commercial repository to /etc/apt/sources.list (there's a way to do this via the GUI in Feisty).

Performance Squad the desktop

[schwaang]

I'd like to see a Performance Squad attack all the desktop apps and their underlying components.

Not the kernel, but kernel hackers do know a ton about how to get good performance, so if they all took time out from the kernel to make the rest of the desktop snappy that would be just fine with me.

Of course I've seen some efforts at this over the years. Dave Jones' perennial "why userspace sucks" talk, some work by Robert Love, some other GNOME folks looking at memory usage, the recent Intel tool looking at CPU-wakeups eating battery life on laptops, and lots of other pieces of the puzzle.

It would be great if the basics of performance "best practices" would become widely known by desktop app programmers again. Instead we're falling into Microsoft's habit of being lazy about performance and expecting Moore's Law (increasing CPU speed and cheaper RAM) to bail us out.

Now my girlfriend's answer would be different: OpenOffice still sucks too much (feature-wise), and it's keeping her from switching from Windows.

Zooming Interface Paradigm

[Anonymous Coward]

What is needed is a Free operating system that is clearly ahead of all the others. The traditional desktop/WIMP metaphors need to be ditched and a lot of resources need to be invested in building a real zooming user interface in its stead.

improvements

[Anonymous Coward]

I would put resources to make Linux more user friendly to the average joe/jane like grandma and grampa.

Basically give it the things that would make it better competed with and replace windows.

A standardized develpoment environment for video/audop like directX so more game companies would develop for linux instead of just windows

Better desktop and GUI support so that you never need to use the command line unless you want to.

Better hardware support, especially video and sound cards, but everything else too.

A crippled interface so someone with no computer experience can mess it up, but yet make it so someone with computer experience cna un-cripple it unlike windos where MS wants to tell you what you can and can not do and makes it very difficult to get it to do otherwise.

make it so that it is idiot friendly (like windows) but yet so a non-idiot can get work done too (like it is now)

The only way linux is gonna be able to really compete with windows on the desktop is by being more like windows, but without the hassles of windows.

I would be using linux now if the games i like would work on windows both now and in the future without the hassles of trying to kludge it with cedega/wine.
Unfortunitely very few games support linux natively.
I have though at least played with linux and can use it, and will eventually put it on a second computer when i get one so i can do internet stuff while i game.
Though i prefer the GUI to command line, after so many years being stuck with windows GUI is more intuitive for me instead of having to remember a ton of commands and thier switches. It was hard enough for me to learn dos.

My sister cant use linux due to being so computer illiterate that even windows is very difficult for her.

Anther friened will only use windows ME beacuse that is the one she figured out how to use and windows changed too much that she doesnt know how to use XP let alone vista.
She would be totally lost with linux

Linux needs to be intuative and usable to people like my sister and my friend.

Um... what about usability?

[c0d3h4x0r]

Linux-based systems have awful usability. They are unnecessarily convoluted and complex, riddled with inconsistencies in both function and appearance. They are geared for success only in the most common cases, but don't provide adequate failure handling and resolution steps for when things don't go as planned.

It doesn't matter how technically great something is if it's stupidly hard to use, wastes users' valuable time and energy, and is too easy to accidentally foul up if you aren't a trained expert. And making something properly usable does NOT mean dumbing it down so that it's obnoxious for experts; good usability makes the experience better for everyone by saving people time and annoyance. It's about time the Linux community grokked that and started caring about the poor users instead of telling them to RTFM. If a piece of software requires reading a manual to understand it, rather than the user being able to explore and discover and figure it out intuitively, then the fault lies with the design of the thing, not with its user.

So if I were king, I would immediately require ALL people working on Linux to read all the standard classic books on usability (The Design of Everyday Things, etc). Then I would transfer 75% of all Linux-related manpower and thinking to usability improvements. (The other 25% I would allocate to improving device support/drivers).

Re:Um... what about usability?

[Anonymous Coward]

Hear, hear! Don't forget The Humane Interface . It should be a requirement for any programmer or designer.

I would also make sure Linux reads these books himself, as he seems to have the impression that thinking about the interface is "fluffy".

ioslaves

[SavvyPlayer]

Kioslaves are a beautiful feature of KDE, but to be universally effective (less http://ask.slashdot.org/) they must be implemented at the kernel level.

User testing

[RobertinXinyang]

Putting normal people, not geeks, in front of the computer and asking the to preform tasks. Then record what they do, and try to do. Then fix the things that frustrate them, change the things that they try to do in the wrong order so that the order the computer wants is th eorder that the typical user tries.

Repeat until using linux is easy

My suggestions

[Lord Kano]

None of these would be simple, easy or cheap but IMHO they would vastly improve the user experience.

• Develop a new GUI system that is a bit less arcane and easier to develop for.
• This relates to the first one, but get rid of library hell. It's such a pain in the ass to have to download three or four sets of libraries just to make one app.
• Find a way to eliminate dependancy hell. URPMI made great strides in that direction but it's an even bigger pain in the ass when you have to install 30 dependancy packages just to install the one you want.

LK

Change it's user base's attitude

[SameBrian]

Anything and everything are possible in Linux. People just whine too much. It's not meant to allow people to single click and stay out of the console. It's about the user being able to mess around and do what exactly they want. The documentation is also done very well, so that's not a problem. If you don't believe me, try looking up a problem in Windows (or, God forbid, using the system's help) that's a bit more complicated than "my mouse cursor won't move".

Comment removed

[account_deleted]

Comment removed based on user account deletion

By making undocumented HW illegal

[aim2future]

I have hard to see any significant improvements needed to be done on Linux itself, which is not already being in development. It is the world around which need to improve. If I had the resources I should lobby politicians over the world to make it illegal to sell hardware which is not documented.

It is very annoying when trying to buy hardware and find that not even the chipset is specified. Such products should just be put in the trash bin directly, not to be sold. Even worse, manufacturers that change chipsets to something incompatible and preserves the old product name, this is to deceive the customer and should definitely be illegal as other types of frauds are.

Of course I prefer HW whis already on the package and product description at the site can tell that this HW is Linux supported. The chipset used should anyway be declared and it should be well documented.

We (our company) are working on a system to at least enhance the commercial aspect, that is as a buyer your have incentive to prefer products which are well documented. It should imply bad business if you try to fool the customers.

As a customer you should ask yourselves: do I really dare to buy this product?, as it lacks documentation; no, I go for this product instead, which is proven to work and is well documented. HW manufacturers who don't document their stuff should either go bankrupt or be jailed for fraud.

3D

[Ox0065]

& I don't mean hardware drivers. Be forewarned: I intend to have a good hard whine here. I've been looking for a formZ replacement since '97. It would make me so happy.

a professionally usable 3d modelling package for something bigger than a little green alien. I mean buildings, streets, cities, the furniture in them. DO NOT SAY BLENDER. Programs like blender, wings, have a long way to go to match even the most basic of modellers like sketch-up for use in anything other than little green aliens. Its not the underlying modelling that's the problem. Its the data input method and the 90s four pane character design work method, not that there's anything wrong with that...

Its just that, with all the other modes in blender, why not a level design mode? It would need its own suite of input tools, but point, line, face, polygon couldn't be that hard to rewrite for interactive input. move, rotate, scale, & their extension, array couldn't be that hard once you have sane data entry. Its just about projecting the cursor in the paper/screen plane out in perspective until it hits the active reference plane. click, thats where I want point two...

Something that allows modelling in context (points entered off elements already in the model); geometry entry modern enough to model in a single perspective window, moving on from that early nineties four pane view. Something that allows geometry to be entered in the model's coordinate system, not the %$&$ing viewports!!!!!! SO YOU CAN WORK IN THE MODEL FROM SOMEWHERE WHERE YOU CAN SEE WHAT YOU'RE DOING!!! and how it relates to what's going on around it. Instead of turning every move operation into three move operations in three different windows, then a change of view to see that its vaguely right, then another three moves to get it kinda sorta right... ...NO! one move, its where you wanted it, see if you like it, keep it or develop it. or no moves, no stretches, no rotations, just make it where you want it in the first place.

Once upon a time I laughed when I saw people working in four pane view without snaps. They seemed to genuinely feel they were working efficiently. There is a difference between precise and accurate. modelling quickly and precisely makes editable models. Models of inorganic elements made without snaps aren't editable. Boolean operations become a nightmare. File sizes & polygon counts blow out. Its just like that. Try the alternative if you disagree.

A pseudo-sorted wishlist

[Sean Riordan]

Hareware Support: Wireless, whether 802.11, GPRS, EDGE, or SuperFooLink is not what it could be. 3D driver recompile shenanigans are not fun. Enough with the hibernation/suspend issues.

More seamless integration of emulation environments. Wine, Crossover, whatever.

Printing.

Optional Integration of KDE widgets ala Karamba, but without the resource suicide.

One a more application side of things:

Better OO interoperability

Better support for viewing/manipulating high resolution Jpeg2000 and NITF files

Open replacements with a decent level of compatibility for:

STK
The Linux version was abandoned several releases ago. Combined with
the ever increasing expense when you add modules to do anything useful
and this is a great candidate for a profitable open source app.

Labview
The NI lock in might be impossible to get around, but almost anything
would have to be better.

Visio
Currently number one reason to stay current with Crossover Office.


And on a totally not Linux note:

Better partition management for FreeBSD if for no other reason than to rescue people (ie me) from piss poor planning at build time.

Then do it

[SmallFurryCreature]

What you are talking about is actually the hardest thing of all to do, and the people who can do it, have no need of it.

So apart from the enormous technical difficulties, you are faced with the human problem of getting people to do something for free that does not benefit them.

First, the technical problem.

Getting a piece of software to install on a piece of hardware without the user knowing anything about either requires someone, somewhere to predict ALL possible situations. This is impossible, there will always be some odd piece of hardware that the hardware detection team never heard about, some combo nobody has tested. Even if by some herculian effort you managed to do it, some company will release a new piece of hardware and you can start all over again testing every combination and its effects.

This is extremely noticable with "brandname" PC's be they laptops or even simple desktops, every time a new model launches by Dell/HP/Lenevo you find that they added a new piece of hardware but changed it subtly and the linux community can now try to figure out what it was they did. I never have problems building linux machines with white-boxes that I put together myself, but dread having to put it on a laptop or small-form factor pc.

Then there is the matter of choice, I could easily give you a piece with MY choice of hardware and MY choice of software and MY configuration choices and you wouldn't have any problems with it. But who am I to say how large your home partition should be? How you should log things.

Linux already is much better then windows in this respect as most installers are proper GUI's with proper help, explaining everything that the user is asked to do but at some point there just isn't anything that can be done to make things easier. The user has got to engage his brain, and think for himself.

If you want to see just how easy linux can be, just try a livecd. That is something windows or OS/X doesn't do, and it is very good indeed BUT it cannot account for every situation, livecd's will fail and the user will be left in the dark.

The second problem:

The people that could write the distro that could do all the hardware detection itself, don't need it. That is the eternal volunteer projects problem, the people of SAMBA don't need a simple way to configure it, they KNOW it and to them the configuration files hold no secrets. Why should they spend their (free) time on something they don't need? Hint, documentation and writing configuration tools is the most boring part of development.

Yes it is a problem, BUT linux is not a company or even a movement, it is a bunch of people who do things they like and share them. If I am singing in the shower and you happen to like it, that is fine, but I won't do requests. If you want that, pay me, except I got a good job, and I like singing the songs I like so you would have to pay me a LOT especially since I hate the song you requested.

That is one of the weaknesses of linux, it can't be helped. MS and Apple can just pay some developed to do boring work, linux by and large cannot.

It doesn't even help that much if somebody volunteers to do say the documentation, the developer still would need to spend time talking to this person, time he isn't spending on something he likes.

This is the nature of linux, it is made by people who LIKE what they are doing, but it misses those bits people don't like doing. Only money can change that, so how much did you pay for your linux box?

Re:Then do it

[David Gerard]

It'll be solved as hardware manufacturers feel a need to target Linux.

Dell is already asking its suppliers for hardware that has a proper open source Linux driver rather than a binary blob. Expect this situation only to improve.

One install method for 3rd-party software

[koinu]

This is really what we need. So that commercial product can be easily installed.

You cannot seriously replace something like emerge by apt or RPM. Not only me would be annoyed, but everyone in this world.

I don't believe it is difficult to make a simple installer/deinstaller like MS-Windows has. One problem might be the toolkit, but developers have often made a 2-layer design that abstracts the GUI from command-line utils. This is a principle that I like. This way you can have your Gtk+ or Qt ontop of the real application.

I would also suggest a more intelligent installer that is focused on games. It should analyse hardware/software prerequisites with simple mechanisms (e.g. a simple configuration file) and have an interface that forwards performance settings to the installed application (come on, we have procfs!). Also rating (age, violence or nudity should be indicated by simple symbols) and so on.

More ideas are welcome. :)

Some ideas

[Anonymous Coward]

Better hardware support is always good.
I want a lot less configuration of hardware, when you plug something in, it should just work.
I also want to see more external libs, like libgphoto and such, be built directly into the kernel so you never have to bother with installing those libs. All you need is the (gui)application to execute commands to do stuff with the hardware
It is also very wrong to have stuff like hal, udev etc. outside of the kernel. That are packages that you install separately.

Everything that is even remotely connected to hardware manipulation should go straight into the kernel.

My Wishes

[g1zmo]

A full Exchange client. Then I could completely ditch Windows on my desktop at work.

Native Unix versions of all of my favorite games, released at or near the same time as the Windows counterparts. I don't care about your favorite games, I just want the ones I play to work. :) Maybe I'm just really unlucky at picking them, but I've never gotten a single game that I like to play to work under Wine or Cedega. Maybe Madden '03 sorta works under every third version of one of them, but I want to play 2008. Company of Heroes? No way. MTGO? Last time I tried none of the text would show up onscreen. The list goes on...

Ndiswrappers

[hairyfeet]

I would dedicate resources to build more Ndis-style wrappers, starting with printers. The only major problem that Linux has now is lack of drivers. With Ndis-style wrappers we would be able to buy any piece of hardware or take any computer and make it "just work".

While I fully support Linux native drivers (like my Broadcom 4318 which is running on bcm43xx) there is simply too much hardware out there to write a driver for every one.By having wrappers we would be able to switch more folks over to Linux.I know I personally have had to tell folks that their Lexmark printer/scanner/fax combos just won't work in Linux, so they ended up sticking with Windows.With an Ndiswrapper for printers it wouldn't matter what type they had,it would "just work", which is what most non-tech folks care about,anyway. Thats my .02.

Tibetan characters

[jjohn_h]

Let the keyboard behave the same in text mode and X11.

Let the default keymap be the same in text mode and X11.

Allow all those combos (ctrl-home etc.) to be identifiable by default in text mode.

Get a new maintainer for keyboard/keymap issues.

Let Andrew Brouwer be happy with Tibetan characters in Unicode.

Scribus

[Micah]

Have you used Scribus for DTP? I suppose one could find ways to criticize it, but it *is* pretty dang good.

Wide-spread use of version control on desktop

[Anonymous Coward]

One of the better ideas of Mediawiki/Wikipedia is the easy integrated revision control system. The major distros supply several revision control systems - GNU Arch, Subversion, Git, CVS, RCS - but do they offer it to the user like Mediawiki does? No. Wouldn't it be helpful if you as the system administrator out of the box could see which configuration files had changed? Wouldn't it be helpful if users of OpenOffice, GIMP, Inkscape, Blender 3D, Scribus, Emacs, Vi/Vim, Kate or whatever upon first use was offered to choose which of the auto-detected revision control systems to use, and then offered a checkbox to turn on revision control upon saving of new files? YES, this would be helpful both at home and in professional environment.

W3

[SL Baur]

Linux development does not need "refocusing". Most of us are volunteers and work on stuff that is either useful to us or fun or both. Bad term.

When I was in a position to donate substantial personal time and resources into Linux Development, I focused on XEmacs because it was almost, but not quite, the perfect development tool, so I lead the fixing of it with help of countless other volunteers.

To answer your question though, I'd have to say it would be a web browser (ie fixing W3 in XEmacs). All web browsers suck, with the exception of Lynx, which has limited but most useful function to some and I would not expect my wife to be able to deal with it, alas.

Opera is decent, but doesn't run anywhere useful. Firefox is marginally acceptible, but only marginally. Everything else, just plain sucks and (with respect to Microsoft Windows where it is considered a user shell) is substantially a regression from proprietary code I wrote 20 years ago.

So, I'd send those Kenyan monkeys after Mrs. Perry until William agreed to head a team of programmers working on W3. XEmacs with a true world-class browser/local file manager would be heaven, in my opinion.

- steve:*:500:500:Steven L. Baur:/home/steve:/bin/xemacs

Less "fluff" - Back to roots!

[Anonymous Coward]

I would say, less (much less) GUI crud.

We're loosing our UNIX heritage and I think that is a very sad thing.

I was real "impressed" with the way thunderbird deleted my mail spool, wanted to implement its own rule parser, pop client, etc..

I like the way package managers get all fouled up when you compile stuff yourself or install things from other package managers. I especially like the way people insist you use a package manager.

I think it's great that you can't install something without downloading 20 other packages, each in turn wanting
20 other packages for optional features you'll never actually need. (LDAP for CUPS?? just so I can print 1 document a year??)

If I wanted windows, wouldn't I just run windows?

Oh, and a good, decent standard 'batch' implementation. (Some day, I just might code that myself)

Linux will never beat windows as far as the desktop, much as it may surprise us, it's not about user friendly or any of that, most people simply do not care.

Installing software is a hassle, most people will avoid it, therefore, most people will just use windows. (even if windows is harder to use, and yes, I do think windows is harder than a good CLI)

Fixing userspace

[stdazi]

Linux would be way better if somoene could fix this kind of issues showing "Why userspace sucks" : http://mirror.linux.org.au/pub/linux.conf.au/2007/ video/talks/38.pdf

It wouldn't take a lot...

[CaptainTux]

It wouldn't take a lot to make Linux more acceptable to the masses. Personally, I'd focus on two areas:

1) creating a unified software installation/removal system.

2) Better hardware support.

Users should never have to manually install or uninstall software and new hardware should work with little or no "tinkering".

no no No no nO No NO NO NO!

[spagetti_code]

That is completely the wrong approach.

You need to ask yourself - what will make more of a difference to the users?
Q: Is it performance?
A: Nope. Performnce is arguably already very good. You might be able to wring another couple of percent, but thats not going to generate any excitement.

Q: is it reliability?
A: Nope. Linux is reliable. Its stable. There are certainly actions we could take to meet needs in the high-end corporate/daa center world, but for greater adoption, thats not it.

Q: Is it hardware support?
A: Yes! (well, thats a part).

Let me give you a hint.
The groups we need to have onside are
- the partners. They will sell it and install it and get it out there
- the users. They will buy it. Both home and corporate.

And what to they want?

Here's my pick, and I'm sure there are a few others:
SIMPLE hardware compatibility
I want to plug my mobile phone in (without having to figure out multisync)
I want to plug in my webcam
I want to use ANY video card without having to look up some compatibility list

Seamless Inteoperability
Open and share documents

Simple UI
Ubuntu were the first to cotton on to this. Users don't want a complex menu structure with admin functions spread all over the place.
This is the one area that I would say several Linux distros are really on the ball.

Avialability Of Core Apps
Photoshop (sorry, gimp doens't cut it yet)
Games

and so on...

Give me Office in Linux for god's sake!

[slashdotmsiriv]

I cannot believe that nobody mentioned it yet.

The only reason i reboot to linux is Powerpoint.

Give me a decent MS Office replacement and I will never load those dlls again.

Really, why is it so difficult to make OpenOffice actually functional?

a unified distribution base

[drfrog]

i think the linux software base was a good start, but there is too much fragmentation between distributions

it would be nice if there was less diversity at the core, that way something that runs on ubuntu would be able to be run on suse or slackware

every distribution is trying to specialize, and thats a good thing, but when there is too much it fragments the communities efforts at large

I would start over with a different set of tools.

[master_p]

If I had a team of developers like that, I would write a new O/S from scratch, in a language other than C, and I would do away with many of the traditional concepts like filesystems/processes/drivers etc.

The OSS world lacks visionaries, people who are willing to drive the state of the art forward. Linux is a nice try for a non-commercial Unix, useful in many cases, but it will not drive computing in the 21st century. Neither Windows will do, but at least I can not expect Microsoft to dump their profit-making business easily.

Universal Installation of Applications

[TheSoggyCow]

Linux really needs a simple, standard, works all around, way to install (or uninstall) any Application or Driver. I think this is what is truly holding back Linux from becoming a mainstream OS like Windows or OSX.

Of course better graphics diver support would be awesome, but thats technically not Linux specific and a lot of things would have to change for this to happen.

And a business/enterprise standardized method for management that would allow hundreds of computers to be simple and effectively updated, backed up, etc. I know there are some solutions out there but one that would work not matter what would be perfect.

I guess all Linux needs is a little standardization for it to really take off. Right now, its the tweaker's dream, configurable in so many ways your head spins. The mass public is looking for something that is simple and recognizable. Something sadly less configurable. Someone push Linux in this direction and Microsoft and Apple won't have a chance.

My 2 cents....

Audio

[mechanyx]

Audio.

Please read Lennart Poettering's 2007 Linux Symposium paper Cleaning up the Linux audio desktop mess.

GUI software performance and memory use

[Bloater]

A lot of GUI software these days is *much* slower and uses *much* more memory than it needs to to do what it does (in the case of memory often 10 times more is used than necessary). Speed up GUI stuff and use less memory - particularly evolution (which also needs a vast amount of fixing) and firefox but also GNOME in general.

Two things: Hardware-centric, and Source On Board.

[torpor]

There are two things that I would do for Linux right now:

1. Focus the distributions on hardware packaging. Yes, not just Ubuntu laptops, or RedHat servers, but embedded systems (such as OpenMoko effort) with complete, full, 100% included Linux on board. You buy a system, it has Linux onboard and ready to go. This is already happening everywhere, but I would push it further and further into other realms of hardware. Until recently, for example, I worked for a major synthesizer manufacturer (musical instruments) with the purpose of putting Linux into digital musical instruments - I would continue this effort, for more and more hardware, and I would put Linux in the box from the start - not as an afterthought or a bring-up/port, but as part of the effort to get the hardware out there.

2. Full Source distributions. This means, a fully populated /usr/src tree with *everything* needed to maintain updates for each individual package - *without* the use of package management. I know this would seem by some to be a step backwards, but what I would do if I had the time would be to develop a completely Source-on-board distribution that has everything a person needs already unpacked and ready to go in order to be able to make local source changes to their system, as well as contribute to each and every dir in the /usr/src tree - using whatever source management systems are the native for each package. So if dhcpd has a subversion server, /usr/src/dhcpd would be fully set up to be able to pull from that tree, using the tools of the primary programmers for dhcpd, instead of replacing them with Yet Another Package Management Nightmare. Too many distro's are not distro's, but in fact are source-code management projects - this needs to be brought much closer to the Linux core filesystem, in my opinion. Its just far too difficult to find the source for a package, unpack it, patch it, etc. - all of this should just be onboard from the get-go.

Perhaps not revolutionary idea's, but as a Linux user since the day of the minix-list post, I feel that there is too much focus on distributions-as-a-fix-for-people-unable-to-grok-s ource-management and not enough focus on having it all onboard in the first place, from first boot onwards ..

Make sure the desktop is as good as Windows ...

[golodh]

I'm sure I'll be blasted for this, but to me the user experience of the Linux desktop (specifically the KDE desktop under Open Suse 10.2 that I tend to use) simply isn't as good as say Windows 2000 (on the same hardware), and to me therefore needs more work.

Why?

- Fonts: the fonts under Windows are much better. Crisper and more readable at small fontsizes, thereby giving the desktop a more compact and more readable appearance. And yes, that's after you download and install the optional Truetype fonts.

- Polish: the look of individual windows and the desktop as a whole is more "polished" under Windows. Despite all the work that has been put into KDE. I don't mean tomfoolery like transparent floppy windows, I mean

- Responsiveness: under Windows 2000 the responsiveness of the system to user input is much better when the system is opening a new application. When I e.g. start an instance of Mozilla under Linux, the system responds sluggishly while it's loading. Windows remains responsive and loads the application quicker. I don't mind that it tends to take longer under Linux to load the thing, but I hate being forced to sit still and watch until it's up.

Oh yes, and please don't jump on me to tell me that I should try GNOME; I tried it and discarded it.

Re:Make sure the desktop is as good as Windows ...

[The Master Control P]

I've never understood why people don't like Linux's fonts; I never even noticed when I switched. I mean, I'm using 12-point menus/titles at the moment. I could easily go to 10, probably 9 or 8, but why? I won't save *that* much space or anything. Maybe it's because I have a large desktop (1920x1200) and I'm falling into the same syndrome as developers who figure there's no point writing tight, efficient code because faster CPUs and more ram will cover their laziness.

And as far as remaining responsive under load, I have the opposite experience. I've come to expect Linux & KDE to keep everything smooth unless there two interactive CPU hogs running (e.g. games glitch if I'm compiling at the same time) or it runs out of physical memory (e.g. opening original haha-guy thread from Fark.com). Other than that, I've come to expect windows to redraw and respond instantly. Amarok once went nuts and was eating 80% of my CPU until I fixed it. Until I saw something eating 4/5 of my CPU, I had no idea how terribly wrong things were going; Amarok even kept playing as it went berserk trying to parse an 80MB collection database file. Windows makes me crazy, because things seem to seize up and then unstick on their own schedule at times. For example, Opera on 2K will completely lock up on me until it loads Acrobat for a PDF.

As I see it, KDE + Amarok (mp3 jukebox) + Mplayer (video) + Konqueror + Kmail (email) + Gaim (IM) + OpenOffice + Akregator (RSS feeds) + [cinepaint | krita | gimp] is pretty much a complete desktop. Neglecting of course the other hundred apps you don't use every day, and the thousand others that act as backends...

CLI Update!

[samfff]

I think Linux should have a totally different interface to Windows or OSX. The target audience for Linux are server administrators and power users and I think the interface designers should realise this. I really don't think that linux will ever beat Windows or Mac for usability, and maybe it doesn't need to. If Linux rewarded you for learning the bash shell then it would be a really interesting environment. As it is I can switch into the terminal to try and get some things done, and if I typeset my latex document I then have to switch back to the GUI to view the pdf. Linux is confused about whether it should be using a GUI to do something or whether the CLI is the better way. Window Managers for Linux all seem to be relatively similar to Windows or Mac, and there doesn't seem to much reason to use linux for a DIFFERENT or better experience. I think this is mainly due to the many different developers involved and their total lack of interest in consistency for their users. The idea of using the CLI is still basically an essential of the latex system due to compiling software. It should be developed to take over the interface, not be hidden behind an inadequate GUI. This would refocus the idea of linux into a straightforward, techie, scripting style interface. However, the CLI needs to be a lot more exciting than the crappy CLI we have now. They need to do things like allowing usage images and media, using search abilities, fixing file heirarchy addressing (as it makes CLIs really difficult) and doing spotlight like smart folders. An CLI interface of this type would need a really good built in text editor that is not quite as obscure as emacs or vi. Linux should embrace its techiness, but not for techiness' sake.

Focusing Linux

[Anonymous Coward]

Just focus on the GUI.

I love that linux can be tweaked by users, but many beginners don't know where to start. Please focus on creating a new, cleaner and simpler desktop environment. Also, please encourage the use of more user friendly download packages and installers. I guess what I really want is an OS that looks and behaves in a similar way to Mac OS X, but still has the open source advantages of linux. All linux distros I have used in the past use mainly KDE or Gnome desktop environments, while some of the features of these environments are pretty good, they generally feel and look a bit like a tweaked version of Windows 95. It would be really nice to see a linux rival of OS X, with a much more simple interface and lots of eye candy. I think there is one in the works but it doesn't look to be shaping up too well. http://www.etoile-project.org/etoile/mediawiki/ind ex.php?title=Main_Page

My wish

[Anonymous Coward]

My wish: I want a coherent, integrated, well thought out desktop OS. I hate the frankenstein - lego puzzle feeling I get when I use any Linux distribution. Make the pieces fit together seamlessly. I do not want to know where one ends and other starts.

Not projects, but tasks

[jsebrech]

People have forgotten that the purpose of an OS (including linux) is letting the user perform tasks. This is a cross-project discipline.

Suppose a user wanted to record some family video, edit it into a movie with some nice graphics and credits thrown in, and then distribute it on the internet on a site of their own (say "simpsonsfamilyvideo.com") so they could tell the address for the site to friends and family to check out. Suppose that this user was your mom. How much explanation do you think she would need to do this? Do you think she would just give up halfway and make her geek son do it?

The reality is that linux doesn't need any more features. It does everything a user needs. What it needs is integration, polish, documentation and task-orientation. Starting from a common task that users do, and then thinking "how can we improve that?" is the only way to move forward from here on out, and doing that means that you'll have to mess with dozens of projects, instead of focussing on just two or three.

Better Wireless Hardware Support

[Roadkilled]

better wireless hardware support specifically wireless cards that has some variations supported and some not at all I think that making some brand names completely Linux compatible will benefit to both the User and the Manufacturer of the product and will encourage more companies to walk the path to full open source Linux compatibility.

How to develop

[Frozen Void]

1.Don't change the command line oriented/legacy architectures(.config /dev ) whatever. Power users need that.

2.Develop an automatic,user-friendly GUI interfaces on top of that.

3.Test the said interfaces with least computer savvy users, correct any problems,test again
until its works out of the box.

4.Each extra parameter,text file change, compile and installation of drivers brings user experience into discomfort zone.

5.Make the computer work for the users, not users working for a computer to work.
Steep learning curves and complex syntax limit
usefulness, and it not justified by "power" user gains by studying the system. Rather its a system designed to be hard from the start.
No compiling,debugging,configuration,reliance on user input.

6.If done right the system would be better then Windows. Which has numerous flaws in its "user friendly design" like interface labyrinths.

Make it like Tenebrae.

[crhylove]

I'm actually serious though. I think our OS should be completely 3d, with real time lighting and everything. The 2d desktop is so 1994. I think that compiz is just the beginning. I'm looking forward to the NEXT generation of 3d interactive desktop. Our kids will literally run circles around us, just as we did to our parents on computers.

rhY

Network filesystem browsing

[Anonymous Coward]

There's one thing that seems to be invented time and time again. I'd like to see some infrastructure for providing something like fuse-smb in the filesystem layer that userland daemons could plug into to support various protocols (I mean protocols that actively broadcast their existence). This would make it possible to browse networks from the console by simply traversing the directory tree.

At the moment network browsing functionality is in the VFS layers of different GUI systems. This is just plain stupid as this is fundamental functionality on todays desktops and should be provided at a lower level. It also means that every GUI is reinventing the wheel. Also there are tons of standalone browsers that implement the functionality meaning lots of wasted time duplicating what others have done.

This is the one single thing I hear complaints about from newbie users who use linux at home. The problem being that their favorite GUI system might not be Gnome or KDE which means that they need to use a network browser not part of their GUI. This means that the program they are using is not the standard file manager for the GUI, so drag'n drop of files wont work. By moving it down we would get it supported by the standard file managers.

This wouldn't even require lots of resources. Just more than I have myself as a math grad. student...

Get rid of X, make the kernel fun for hardware

[Anonymous Coward]

I'd remove X11 and replace it with a standardised graphical interface that does not involve high levels of networking and instead focuses on the desktop side of things, this would of course be for desktop users not servers. Instead of complicated config files the gui would simply find a graphics driver and work with the kernel framebuffer driver which would do almost nothing because graphics would only really be needed for the gui.... this is desktop emphasis in my idea. Each gfx driver in the gui could have a config file for simple setting changes if required. Graphical customisation would all sit on top of one simple, well-defined standard graphical interface instead of all the rubbish that currently sits between users and their gui.

I'd change drivers to a binary model that used a well-defined non-changing interface where anyone could make drivers without being a kernel hacker and could use them in any kernel simply by compiling them and sticking them in a driver directory. No modprobe or hotplug rubbish...

I'd also switch to a model where the kernel changes very little and drivers are the key to everything where the baseline is not constantly moving and you don't need to recompile on a weekly basis if your graphics driver is out of date... just get the latest version from the web.

Package management on linux distros is also far too complex for the average user. Again i am talking desktop users, just my opinion so no flame wars please..... I am inspired by such projects as www.pc-bsd.org and www.syllable.org (The Syllable Operating System)

I'd love to see more people interested in such projects.... not to say linux is not always moving forward but it has so many things holding it back on the desktop that i have seen since starting using it in 2003/2004.

I'm not quite sure

[yoprst]

Linux distros need to strive to be better than Windows.
Back in late 80 - 90 MS software and PCs used to be worse, but cheaper (Actually, since late 90 PCs were better and cheaper, but MS software still was crap). So they won. Now Linux is worse, but cheaper. I can't see how it may fail to win, eventually.

Better Beowulf clusters

[rwa2]

No, seriously... especially since this is what Linux is best known for.

Specifically, I'd like to see more mainstreaming of some of the features from mosix / openmosix. I played with mosix clusters (deployed from a master NFS server using debian's diskless-image package) quite a few years back, and I've been yearning to set up more.

Some of the features currently possible (or that I'd like to see):

transparent process migration CPU-intensive processes automatically run on nodes with free CPU cycles load balancing / optimization Network or disk I/O intensive processes run on the node nearest to where the data is stored high availability & fault tolerance your cluster can still be usable through individual node failures or even during upgrades(!) scalability Simply add compute / storage / graphics nodes to the system... as you and your friends pool resources into the system, it just gets more powerful So if you had a computer, you could just plug in to the network and netboot to get the current kernel and mount an NFS-root partition and join the cluster. You could set up any authentication necessary for CPU sharing and offer up any local disk partitions that could be automatically allocated to the network RAID. You might have mobile nodes (Laptops or maybe even PDA thin clients) that would run apps off the cluster, and then cleanly migrate its processes and hoard its files back to itself before cleanly detaching and continuing to run on its own. *NIX and Linux in particular already has a lot of good projects that handle the individual pieces you'd need to assemble a system that would work almost completely over the network like this (booting, parallel global network filesystems, even good 3D OpenGL over the network), it's just a matter of assembling, integrating, and filling whatever remaining holes to make this seamless.

hotUNplugging

[arghileh]

Windows used to bluescreen if the floppy disk you were writing to disappeared, but it does not anymore. It yells at you but it is ok with it for the most part. Linux on the other hand really does not like it when hardware disappears without its knowledge; this is most apparent on mounted filesystems. One part of robustness is how we deal with failure and having a USB drive yanked or reaching a bad part on disk or in a file system should not cause the kernel to spit out thousands of error messages, one is enough, and deal with it already; don't lock up on me, if the file is inaccessible say so and move on.

design

[sentientbrendan]

People tend to focus on individual features that linux is missing most, but as a developer I have to say that what linux is most lacking is overall design.

Unlike other operating systems, linux lacks any kind of architect responsible for the overall design. Instead there are a bunch of competing groups providing specific components, and no one really responsible for orchestrating how they all fit together. This means that a lot of the underlying facilities in linux are either outdated (left over from early unices), munged together hacks, or are good systems, but aren't used by much of the system.

Some examples of areas that need improvement:
daemons: The model for daemons used now is archaic and needs to be done away with. There needs to be a central place for registering and controlling all services. There needs to be a single system level API to support writing services. All this nonsense aboutlocking files to determine if another daemond is already running, and writing your PID out to a file needs to be done away with. Finally, someone has to actually take existing daemons like httpd, ftpd, etc and add support for the new model to them.

program configuration:
Right now all the system level programs and even some user level programs use their own custom flat file format for even trivial configuration details. Gnome has a pretty nice configuration system used by most gtk apps called gconf. This or something like it should be used by everything.
At the very least someone should introduce the X.org guys to something called "xml" for use in xorg.conf.

C++ support:
Generally, linux is a crappy place to write c++ apps. The c++ abi on linux has broken numerous times, which means that people are afraid to write c++ on linux and a lot of libraries only have c interfaces. Even libraries that are written in c++ often only export c++ interfaces because of abi issues.

Build systems:
Currently, most open source projects use something called "autotools". Does the following looks familiar? ./configure
make
sudo make install

That configure script and the makefile are generated by autotools. From a users perspective, it aint too bad, and everyone's familiar with the install process. From a developers perspective, autotools is pure hell.

Autotools is essentially a really bad attempt at providing tools for writing crossplatform software on top of make. It is overly complicated, slow, most people don't really understand how it works, and it encourages developers to write heavily "ifdeffed" and totally unreadable code. Furthermore, despite the fact that everyone uses it, I don't think it's actively being developed anymore (kind of like cvs...).

Essentially, someone needs to identify a new standard for c and c++ build systems and talk a bunch of the existing open source projects to convert (which would be difficult) or at least make all *new* project under the new build system.

X:
X is a major piece of linux architecture that most people will agree sucks. X is hard to configure and doesn't do autodetection of hardware configuration right even on supposedly good drivers. However, autodetection of good hardware configuration may be still a bit ambitious. The biggest bang for the least effort would probably be to just make X more configurable by doing away with xorg.conf and replacing it by either plugging into gconf or just using some kind of xml file. Once this is done, it becomes trivial to write GUI's to configure x.

Oh, also, I shouldn't have to restart my computer to change my resolution.

Sound:
Sound support sucks on linux. However, all of the existing (non-driver related) linux sound problems would just go away if someone made /dev/dsp multiplex itself...

So really, what I think that linux needs is high level direction in terms of architecture. That doesn't really mean getting people to write some new piece of code so much as having someone step up and say "this is the right way to do things" and to talk to the existing projects and convince the maintainers that these approaches are the way to go and get their support.

Improved support for multiple monitors

[Anonymous Coward]

Multiple monitors just work on windows. To get the fine touches you need extra software but otherwise they simply work.

Multiple monitors never just work on linux. Sometimes they do work mostly, sometimes they do work barely, and mostly they just plain don't work unless you know a guru.

all i can say

[kaizokuace]

is that from a desktop type user's perspective people use their computers for what? To access media (video, music, games, content creation software, internets, word processing and whatever else is in office, money management, etc.) most of those are just apps that we use. Things like games and video and sound are hardware and driver issues also. Clean those up and make the OS easy to configure and customize and the average, even slightly above average user will be happy. When we all say we want linux to be a desktop os too then you have to think of the desktop user's needs. Get those needs fulfilled and you got yourself a platform. Also standardization between distros would help. Hell I wouldnt mind if there was just one distro. Anyway apps win the platform. If everything runs on linux then it is the platform of choice. Not everything I use runs on linux but everything I use DOES run in windows. I just dont like how much of a pig windows is so I only boot into it to use specific apps. Games will come to linux when there is enough market saturation to intice devs to make linux ports of their games. There isnt even enough mac users to get that to happen for osx! Maybe if some investors put forth the money to get games developed for linux then that could start a movement in the games industry. All that is needed is the right kind of metrics to show that enough users are playing games in linux. That and a nice dev environment like visual studio. I dont know everything in the world but this is the way I see things. Another comment earlier mentioned a linux website to help people get into linux. I think this is a good idea. A big problem with people not switching is that nobody that isnt totally cool and computer savy knows what linux even is! Advertisment would help a lot.

games..

[amberp]

yes, games! If i would have resources i would have used them to develop and port more games on Linux. I personally feel that good games are very important to attract a bigger number of people to use Linux.

a decent GUI and its decent interface builder

[Anonymous Coward]

The Linux kernel is already a masterpiece wrt modularity, stability and portability. It's the user side we must concentrate on.
What a lot of developers need is a well designed and fast GUI library with its interface builder to help the creation or painless porting of desktop applications.
I've tried almost every system out there, GTK+Glade, Qt+Kdevelop+designer, Fox, wxWindows->widgets etc, and they all have usability problems to the point none of them could actually compete even with the old Delphi2 (that's over 10 years ago, folks!) when it comes to creating a working gui without shooting yourself in the foot. GUI creation is a 95% visual and psychological and 5% technical issue. If a GUI library and its builder are so badly conceived I've to take brain time usually
allocated for algorythms and more technical problems for it, then there's a problem somewhere.

In my opinion this is the way to go. What Borland attempted and failed, those folks did successfully: native code both on Linux and Windows (no Wine), a functional GUI builder and a pretty darn good library, clone of the original VCL from Borland. The people at Trolltech should seriously take some ideas from this project.
Too bad it's not for C or C++, which isn't a problem when creating apps from scratch because Object Pascal is a pretty decent, fast and complete language to write software on (forget the Pascal you've learned at school), but that doesn't help to port a lot of software written in C and C++. Nevertheless, from a GUI point of view this is the way to go.

TODO list for a team of highly skilled devs

[Anonymous Coward]

Create a program which one can run on a machine to capture all relevant hardware data and decode that hardware configuration into a virtual machine which can be reused anywhere. If such a program existed, then all Linux users could encode their hardware into virtual machines and submit them to a central site which maintained a gigantic hardware test suite. Once a new kernel and/or distro iso was prepared one could run that ISO through all those virtual machines and get back a success rate. Once the iso booted inside each virtual machine it would do some basic tests like test wifi, play some sound snippet, play some video to check hardware decoding of video etc etc. At the end of the test suite run, the rate would say something like "this new ISO image of distro XYZ booted successfully on 9456 out of 11456 hardware configurations".

This would give you a tool to compare different iso distro images with each other and guard against regressions and it would also make it easier to track progress.

Documentation

[smchris]

Fuller: examples, exceptions

Enterprise management

[AYeomans]

I suppose I should have expected this from a Sunday Slashdot post. *All* the issues in the comments are niggles about personal desktops. Nothing on the issues on how to run a large-scale enterprise deployment.

Sure, Linux has all the hooks available for enterprise-wide administration, directory-based authentication, centralised logging, patch management, backup, etc. And a wonderful choice for most of these - but no simple plug-and-go option as easy to initially deploy as Active Directory. No, AD isn't perfect - but once it's deployed, it's there for good.

Linux Development

[TheCoroner]

What has prevented me from switching wholesale to Linux is peripheral hardware driver support. One strategy may be to emulate driver support from Windows. I realize this introduces a host of issues, but it would effectively eliminate the reliance on the vendors themselves if Linux could integrate hardware in the this manner.

Manipedia

[Anonymous Coward]

If there were a wikipedia style man page website from which all the distributions kept their man pages up to date, perhaps we could begin to alleviate the supposed learning curve when using linux software. Honestly though, I think that the only real problem with man pages is the lack of examples to help guide the way. Thinking that a more eloquently written man page description is going to make 'ls -lah' any more easily understood is slightly ridiculous.

Rewrite All Audio Code

[Apreche]

OSS was crap and alsa was supposed to save us. Guess what? Alsa sucks too. We've reached a point in Linux where if you want to play stereo sound out of your sound card, it mostly works. However, more and more people are using on-board sound cards, and if you buy a new motherboard the chances are that the sound card will not work at all. Having your sound input, like your microphone or your line-in work is seemingly up to pure chance. If you have somethign weird like a microphone built into your laptop, just pretend it isn't even there. If you're planning on surround sound, just kill yourself.

Playing and recording audio is one of the most basic functions of a computer. The situation back in the DOS days where you had to tell every game you installed that you had a sound blaster is better than the situation in Linux now. Have you looked at an alsa configuration file? Why do I need to setup extra dbus stuff to be able to play audio from more than one source at a time?

Playing audio is basic desktop functionality. It's well understood. Everyone who uses Linux on their desktop wants audio to just work. For many people, it doesn't. Fix it. Fix it now.

kernel outcasts

[Anonymous Coward]

I would also put up a web site with alternative kernels. www.altkernels.org or something with kernels that use patches that never made it into the release. I'd love to try a gaming server and benchmark all the different schedulers. I'd love to download a pre-compiled binary with all the modules relative to a specific distribution with a certain realtime patch. Sure there would be a lot of packages but at lease people with kernel patches that know will never make it into the kernel will have a place to gather... the outcasts I suppose, but I wonder how many kernel patches don't get written because they know that there's no way that their mod will get mainstream and not enough visibility will let it be tested.

The big issue is that there would have to be compiled kernels for all the distributions, their versions with modules, so that it's easy for everyone to use, not just kernel hackers, but those who want to try new features and want to do it without a crapload of command line and debugging output for the next command line.

I know that the distributions could do this but that's a segmentation that would not let this work.

Gather the outcasts and let them live.

on a side note, another change I would do is get all the scripts that are part of the lsb put to binary, and have them well documented. other standard scripts that do a lot of work on the system could be put to binary, it would speed up the system quite a bit. sure there's not much difference with one script but when 40 scripts end up getting executed for one action, there would be a real interactive difference. I know this has nothing to do with the kernel but whatever.

Dan9999 (sorry for the AC, too early in the morning to get my pw emailed to me and log in)

OpenGL

[Anonymous Coward]

The one thing I would like to see getting focus is OpenGL. Until OpenGL is considered a true competitor to DirectX games companies (with the notable exception of Id) will not use it. If OpenGL became the standard then porting of games becomes much easier and one of the major remaining stumbling blocks for Linux is removed.

Refocus Linux Development

[duduhead]

I would love to see Linux in general use as a desktop replacement for MS. In order for that to happen it must be able to install itself taking advantage of all the available hardware in my system, read Word and Excel files and enable me to edit them and republish, and I also need to be able to run Photoshop and Premier. Until that happens it's a great server.

How installation of patented software could work

[tepples]

To be able install some form of linux right off the bat, have it autodetect all my hardware and just WORK. This requires makers of hardware to "just WORK" with distribution maintainers, and many makers of hardware have political reasons for not doing so. Some users who have replied to this story claim that Dell is about to deprioritize hardware makers who refuse to cooperate.

To install programs simply by going to some kind of "Install New Software" select the program (possibly from some ordered list of say, Word Processors, Graphics Packages, Games etc etc) and have it download and install, with a password required for install. Are you asking for something like Windows Marketplace?

Some kind of plug-and-play like utility for new hardware. I shouldn't have to mount a USB drive from the command line, it should just be recognised. True, it'd be straightforward for a process to poll the bus every few seconds to find drives to mount in /mnt, and some distributions already do this. But what about unmounting? Or are you requesting that the implementation of FAT32 keep the file system in a consistent state even when the user yanks the USB data cable in the middle of an operation? Is this even possible? Even Windows has the "safely remove hardware" icon in the taskbar notification area.

I know this is a licensing thing, but I'd really like to get standard formats to work (mp3, jpeg, wmv, etc) without having to poke around.

Here's how it could work: Each package contains a list of patents and expiry dates, one list for each country. When the user tries to install a patented package from a repository, the package manager geolocates the user's IP address to find a country and places the the package on a queue to install. The PC synchronizes with an NTP server once a week. After each synchronization, the package manager wakes up, checks the queue, and installs those packages whose PatentExpiry date for your country is before the current date. At any point, the user can open the package manager and see the patent numbers that are blocking installation of any given package. Sure, this would take several years to install packages that rely on recently invented process, but it's the only way to circumvent royalty requirements in a free software distribution.

Or you could just use Linspire's CNR and pay for your proprietary software.

Penguins.

[Oktober Sunset]

There is only one thing I would change. More penguins! Penguins everywhere, a penguin at the command line, penguins on the desktop, printer test pages would just be a picture of a penguin. Ascii art of penguins inserted in code comments. when thinks are erased, they are overwritten by a pic of a penguin first. Anyone makes a CD based distro that is less than 700 mb, the rest would be padded out by penguin pictures, DVD distros would be padded out with penguin movies. It would use your webcam to search the room looking for penguins and if there wasn't enough, it would threaten to delete all your data unless you stuck up penguin poster immediately. It would steal all your passwords and post pics of penguins on your blog, and email all your friends with penguins, it waould also order penguin soft toys and penguin accessories with your Ebay account. It would play penguin calls and songs about penguins all the time, especially at night, when it would wake you up at 4 am with penguin calls.

Eventually, it would contact online plastic surgeons to arrange a penguin conversion operation to turn you into a giant penguin.

To put Microsoft in the hurt locker ...

[Elias Israel]

Focus on two things:

First and foremost, a real, FREE replacement for Exchange that not only duplicates its functionality, but beats it hands down and offers dramatically lower TCO. Exchange is a pig that eats server machines readily and greedily. It's also the main anchor that keeps a lot of other MS technology locked into the IT departments of many companies. Take it out.

Second, further extend and enhance OpenOffice so that it also beats MS Office on features and ease of use. If the execrable OOXML becomes a standard, implement it fully and relentlessly. Publish the resulting code as the "missing parts" of their presumed spec. Deny them the ability to claim that their "standard" is really only implemented properly by their own products.

Open Mosix

[flyneye]

I would without regret or another second to think,reopen Open Mosix,and with a fleet of agreeable developers,finish and include it with the kernel.(yup,tools be done too)
The world needs to be able to have home clusters and/or a use for older computers.
Microsoft will never implement this wonderful feature.We need to breath some life back into it.
Hands up,who misses Dynebolic 1.4.1? Clusterknoppix?Quantian,for math and science? Who wants to use the power of their network to composite video or run blender?Who wants audio effects rendered on large files in mere seconds? calculate prime numbers with the speed of ALL your processors in the house?
Yup,thats what I'd do to refocus Linux,merge Open Mosix.

Not just devs....

[Toreo asesino]

There's one thing that makes OSX and Windows so much nicer on the GUI side than any Linux window manager:

Apple and (especially) Microsoft spend billions on finding out how the user interacts with thier software, and how to make it easier and more natural. GNOME and KDE don't and it shows in comparison.

Developers are definately vital, but unfortunately, it's a bit more complicated than that.

I'd pay (competitions)

[shdowhawk]

I think that to be able to "fix"/"improve" linux... we have to fix the way that we develop for linux. I love the open source community... but lets be honest. It's pretty useless at times. It seems like a lot of projects get 70-80% done ... and then interest dies in it... or when it comes to the really hard / boring things like driver development ... no one wants to do it. So ... I feel that to "fix/improve linux" we need to change the way that things are developed. How can this be done? Pay the coders!

Why is it that OS X / Windows is "better" and "winning"? .... easy... $$ = motivation. Their employees are getting paid more. Hell, some open source developers aren't paid at ALL. (damn but we do love you all for doing it anyways!)

My solution to all of this, is to pay the "best" programmer(s)/group(s) to get individual pieces done. This can be done in the following way:

Step 1. Set up a foundation or an organization of some kind that has the financial responsibility to hold on to a chunk of money. (This could be the official linux dev. Group, but preferably NOT a distro, as they will be biased. This has to be unbiased towards any distro)

Step 2. Set up ye olde competition of the week / month. Think X Prize (maybe call it the linuX Prize? hehe).... the $ reward will be voted on by the developers of the linux kernel as WELL as the whole open source community. Things like wifi and sound (oss) would be weighted higher than say, getting a good pdf reading system to work. BUT! Still give a prize for the PDF thing. Say... 30-40K for the top winner of the Wifi competition, and say ... 5-10K for the PDF one? Then maybe give something to the 2nd and 3rd place winners too. Hell, maybe not even put a cap on it... donate X for this cause... there might be such a huge call for something like wifi drivers that over 100K+ might be raised. Hellz yes!

Step 3. Generate money. This can be done in a ton of ways. Example: Community support. I'd for example be HAPPY to give 20$ into a pot for someone to come up with a good working easy to use and "out of the box" wifi driver / recognition set. Then add in sponsors. I'm sure we can generate lots of money to keep this going.

Step 4. Advertise. Make a single website. Professional looking, lots of feedback on LINUX. Explain why linux should be used and it's major features (do NOT give specifics on the main pages or you will scare people away. Make the specifics easy to find as a separate tab). This is NOT for a specific distro, but for linux in general. That being said... have a wiki like set of info for each distro... but don't overload it with details (distro watch is very cluttered). Show the BASIC pro's and cons... and have users put their ratings on it. That way when a person decides they want to use "linux" ... they can easily find/see that the most popular or best rated disto is X (this caused me a huge headache when I first got into it). Also add daily / weekly / monthly (potentially all 3?) user polls. Example: "What is the next big thing we should compete for? Option 1,2,3,4,5 or 6)"

Step 5. Reap the benefits. If we have 10 people/groups submit what they think will win them the best prize, we'd have 10 pieces of code that (arguably) will all work. Even if we choose the "best" one, we (the community) might rip them all apart, and take the best aspects from all entries to make a new complete even BETTER version!

Rules for code: a. GOOD documentation for the code. This means "readme's" as WELL as good IN code documentation. b. Portability / Expandability. Is it generic enough to work for everything / most everything? (example: wifi drivers that recognize at least 80% of wifi devices?) c. Speed/optimization. Portability is nice... but if it slows everything else down to a crawl.. then it's really not that useful. d. 3rd party integration. How easy is it to take this code and build on top of it.

... Motivation is the key to making linux better. What better way to get people motivated than to reward them for their hard work with some tangible materials. $$$

consistency

[edis]

"what application or kernel area would you attempt to improve" - as long, as you can relate "linux experience" to be amongst those named, this would be most important. Watching it for over 10 years by now, it is hard to miss disappointment, that is slowly replacing slightly bigger hopes for OS, had before. OS/X is there by that time, and it is willingly accepted as consistent Unix-allowing, rarely disappointing, rather well polished tool. Not sure, there are still as many enthusiasts over there to test next version of next distro for fun. Maybe me myself am getting older, but it was a while, since I was willing dive into that experience once again - too predictible, too little to be found, too disappointing to discover new headaches.

If distributed opensource development has next big issue to overcome, I would say it is consistency amongst distros, in user interface, in quality of applications deployed, wherever you see. That's difference, commercial software is able and willing to varant - deliver not promise of capable product (and multitudes of promises), but one of usable scope, finished in very certain extent, even if that value is being asked money for.

Linux/opensource development has to get focuses. It is not easy to do being loosely distributed by origin, but there were such efforts at the earlier stages (like LSB, LI) - either they should function in new quality, or other efforts should find place and bring in new sense into Linux development.

Linux vs. Winblows

[Ex-MislTech]

Unless your using Ubuntu or similar distro, your gonna use the CLI a fair bit.

You can try using YUM, etc etc, but eventually your gonna get down to some
nitty gritty CLI, its the nature of Linux's roots.

More GUI will be added with time, but a lot of the Linux folks
do not have making Linux just like windows as their priority one task.

Linux would be best served by Distro's that best serve Enterprise tasks,
gamers tasks, common SOHO biz users.

Some distro's lean this way a bit, but the names chosen for a lot of
the apps are a tad less intuitive than what M$ chose.

Office is obvious, and so is Open Office.

GIMP makes ppl think of pulp fiction, lol.

A few other apps are likewise.

But in the long run what matters, is things like
the netcraft stats, Linux/BSD/*nix is kicking the crap
outta M$.

Nuff said...

Stable APIs!

[king_ramen]

Lock down the APIs so verdors and closed source projects can link to the kernel. Things like FUSE, NDISwrapper, etc. should be "the way" ALL layers are done. The kernel is robust enough now to lock down the interface and allow the hardware and internals to grow independently of each other.

Re:start over from the gound up

[Ex-MislTech]

If you are looking for market driven GUI paradise then open up and say "Ahhh"
and get you big mouthful of Fistya^H^H^H^H^H^H...Vista, and stop using Linux.

Oh ...and STFU...

Duh - make it install with no fear factor

[cjonslashdot]

To technical users, Linux is already easy to install. But to the average person, who spends most of their time in things other than the ins and outs of computers, Linux installation is intimidating. The very thought that one might have to edit a file in a text editor is a non-starter. For mass adoption, Linux needs to install at the click of a mouse (from a running XP or Vista window), with minimal risk of problems along the way. This needs to be the case for the broad range of systems, from laptops through desktops, and for nearly all types of displays, sound cards, network cards, and so on. If this cannot be achieved, Linux will not achieve mainstream status as a desktop system. Period. That is the challenge, and rather than tinkerking with niceties, this is where effort should focus, IMHO.

Re:Thorough context sensitive help...

[ChrisMaple]

Complete, accurate, current, non-ambiguous, hierarchical documentation, clearly written. With examples for difficult stuff. I've wasted hours creating simple bash shell scripts that would have been created in minutes if documentation were better. (Bash is a particularly difficult case. There are many diverse things going on, which makes the documentation file huge and hard to organize.) For complex stuff, man files are too hard to search for relevant data. Info files are hard to navigate, and (in my experience) searches aren't possible from one section to another. HTML or other similar formats have the capability for good documentation, including hyperlinks for table of contents and index.

No thing

[Anonymous Coward]

Teh Lunix is prefection, by design. It's impossible to improve, since it's always perfectly flawlessly perfect, and MS is an evil conspiracy which is hiding this perfectly flawless perfection from teh world.

And death to anyone who says otherwise: they just hate teh Lunix, and are getting paid by (or work for, or own the stock of) teh MiKKKr0$$$l0th!!!

Re:How installation of patented software could wor

[VE3MTM]

True, it'd be straightforward for a process to poll the bus every few seconds to find drives to mount in /mnt, and some distributions already do this. But what about unmounting? Or are you requesting that the implementation of FAT32 keep the file system in a consistent state even when the user yanks the USB data cable in the middle of an operation? Is this even possible? Even Windows has the "safely remove hardware" icon in the taskbar notification area.

That's actually possible in Linux, with any filesystem, if you supply the "sync" mount option. You'll notice a significant performance drop when you use it, because programs will freeze when they access it when otherwise they wouldn't. Once they finish, though, the data will be on the device.

Games

[ThatsNotPudding]

Establish and fully fund a gaming house to create truly fabulous games that only run in native Linux (on both PCs and reformatted consoles). Hire the best in business (as Kane hired the entire staff of his rival newspaper). Establish full ride CS/Gaming scholarship competitions at the High School and Junior High School levels. Honestly, with the exception of timid, hide-bound corporations, the only thing keeping Windows relevant now -is- the gaming platform. Yes, massive company installs of Office and Exchange matter in the short term. But even the most corpulent, clueless institutions must have a constant influx of new, young employees. If those young ones have a platform (Linux) that finally satisfies all their personal computing needs (read: games), then those that rise up through the company ranks will make sure the MS dominance slowly withers away. I think this is what makes Ballmer wake up screaming in a cold sweat - and why MS is willing to loose so much money on the Xbox 360 and even do the right thing regarding the red ring of death.

spontaneous order.

[anwyn]

The Linux kernel development process is a spontaneous order. No government or "Central Committee" established it.

History has proved time and time again that when ever a "Central Committee" tries to interfear with a spontaneous order, the result is always bad.

Those who sit at ease making "bright idea" suggestions on how the kernel development process should be improved are really those who in their minds promote themselves to be members of a controlling "central committee".

The kernel development process can not be improved by those who sit.

The real way to improve the process is to git (pun intended) involved. Make a patch suggestion! Or if you are a rebel try to fork the process! Do something!

Chair warmers will accomplish nothing.

EVERY computer has a command line!

[Anonymous Coward]

EVERY computer! has! a! command! line!

MacIntosh adopted Bash. Windows XP has the Windows PowerShell. Go look it up! EVERY computer ever invented or to be invented uses a command line, no matter how cleverly disguised or hidden it is. Every time you use the Internet, you're using a command line. The URL is a command line. Google's search form is a command line. The thingie you type in AOL keywords is a command line. Chat programs are command lines. This comment form is a command line.

If it has buttons and a readout, it's a command line.

You don't want to deal with command lines? Become Amish, and never use anything containing a microchip again, including a microwave oven, a clock-radio, cell phones, ATMs, and soda vending machines.

Printing still not right

[migloo]

Agree with suggestions #1(wireless failures) and #2(sound conflicts)
I would add that printing still does not work seamlessly.
CUPS is (sometimes) a disaster that should be replaced by something simple and efficient.

Re:Focus on education and after this, documentatio

[ChrisMaple]

FWIW, vim's bultin ":help" has had a good section on search-and-replace for at least 5 years. regex is well documented in many places, and applies to vim, sed, grep and other tools. In my opinion, no video needed here.

There are great gaping holes in documentation for gnome and kde tools and UI. There are places where videos might help, but who would want to provide them, especially for free?

package management

[quitte]

What all linux distros miss is a mechanism that allows to install software as a user. and by that I don't mean system wide but in your home directory or in a group wide directory. as any software that doesn'T need to run as root can be compiled and installed in $HOME anyways the package management should support that. Of course I don't want a mini distro to be used in your home directory but instead packages that can be installed both sysem wide or in your home directory with intelligent dependency resolving. While it is a little tricky to automatically remove a users own package once it appears as a sysem wide directory, or reinstall it once the system wide version disappears, I think it is doable.

Reorganize the UNIX file system!

[Theovon]

Plenty disagree, but just as many agree. The UNIX file system organization that splits a single app across multiple different places (/bin, /usr, /etc) is unintuitive and needs to go away. Moreover, how each app is organized is not necessary the same as every other. Plenty of people learn how things are organized, but having to memorize these arbitrary things increases the cognitive load for experts and makes it harder to learn for novices. Just starting by putting each app in its own self-contained directory under /Programs or /Applications like GoboLinux or MacOS would be a good start in cleaning up this horrible mess. Just think: If we did this, these unreliable application managers like Adept would no longer be needed!

wine and installation

[Cyko_01]

as a linux noob I found that installing programs that were not listed in the default repositories to be very difficult. I think an MSI type installer would be extremely useful. I was so happy to discover autopackage and I would LOVE to see it more widely used for non-repo apps.

The other thing that I think needs improvement is WINE and game compatibility. Lets face it, commercial games are not being developed for linux. we need a way to make them work as well as they do on windows so that "linux has no good games" can no longer be used as an excuse not to switch. we should be able to reply "linux runs almost all the games that windows does"

Re:Three things. BETTER GUI FOR SECURITY... apk

[Anonymous Coward]

"In other words, something that improves KDE, Gnome, X, etc. is a perfectly fine answer to this question." - by EvanED (569694) on Sunday August 26, @12:50AM (#20359235) I agree: Where you have to concentrate most would be where "the rubber meets the road" & in the GUI interface where users "meet" the OS basically.

One thing I've noted, when I challenged the *NIX crowd here to a multiplatform test of security in the CIS TOOL (noted by COMPUTERWORLD & SANS, both of whom are often cited on slashdot & thus, respected to a decent degree)?

Some folks here stated (more-or-less) that working with SeLinux & its MAC (mandatory access control) kernel hook addon layers & configuring it is "complex" & probably a bitch to do, and to secure your rig with "layered security" thru & thru, you need to work with SeLinux imo, above & beyond say, NetConfig/IPTables + chown/chroot/chmod stuff.

E.G.-> HARDENING LINUX

http://it.slashdot.org/comments.pl?sid=267599&cid= 20203061

Security matters, online nowadays especially. Creating a "nice, easy to use interface" to this toolset would be a GOOD start, imo @ least, & yes, apparently those of others who are actually into Linux distros that have SeLinux "baked in" from the start (UBUNTU/KUBUNTU for example, do).

Working with ACL's for the registry &/or filesystem in Windows NT-based OS' have such an interface, & it makes it easy enough to work with via MMC.exe snapins for policies, & also via Explorer.exe + Regedit.exe, adding/removing users & altering their rights easily enough (sometimes it's not so easy even there, due to cascading & inherited rights on groups especially imo, not so much for individual users), via those tools.

If Windows & Microsoft can do it, for ACL control? So can LINUX for its version of the same in MAC!

APK

P.S.-> By the way, I actually respect how far Linux has come, since when I first tried it back in Slackware 1.x back in 1993, because it has come a long ways in many respects (mostly imo, plug & play, device support (the TOUGH area imo, for Linux, because everyone's "looking to sell their hardware" & thus, oem's of hardware HAVE to code for Windows, & Linux is an 'afterthought' unfortunately), & kernel level threadwork (as opposed to the original model of round-robining thru usermode threads into a single kernel thread instance) so Linux could be enterprise-ready for SMP)...

Still, when it all comes right down to it?

Folks "connect to an OS" via its GUI interface shell!

& this is where development needs doing, like for SeLinux (the example I use) bearing distros of Linux, to make it simple(r) to work with...

Now, admittedly, it's been a LONG while since I loaded a Linux (Redhat 6.0 here in fact), so I am not sure what the interface to SeLinux looks & works like, but from what I read (mainly from a fellow named SanityInAnarchy a /. user here who complained SeLinux is "intricate & complex") it apparently is, & needs to be made simpler! apk

Re:Focus on education and after this, documentatio

[Anonymous Coward]

I just want to say that in regard to Vim, there is an excellent manual accessible by entering ":help" (without the quotes) in Normal mode.

For help on substitution, enter ":help find-replace". The section on substitution can also be read online.

There is also a talk given by the author of Vim, called 7 Habits For Effective Text Editing 2.0 available on Google Video.

Hell What !? Calling out to Devs

[codermaniac]

Do NOT follow any of the guide lines here, just don't care, I don't. Do what you like and do it as you like. Keep programming as a Joy, solve problems that itches you, we don't need guidelines to what to do, we do what we love. Its open source people, you have the code, change it if you want to what you want.

Sell it to microsoft

[Anonymous Coward]

Just like they did to MS Works, MS Linux could be the piece of shit budget OS that it is, but better developed so that it can be good. like windows... you guys know... with the working and compatibility and such.

Re:Sell it to microsoft

[JoeCommodore]

And then in a year or so they will "talk the talk" and eventually release some half-assed upgrade (fixing all those "Windows incompatibilities" in the original - as in windows-izing the APIs, retooling the standards to MS ones, etc.) and then later report due to "lack of interest" for the other platform versions drop all support for Non-Windows versions.

Then the next release they will tout it as their innovation in technology, slap on a new sparkly logo, maybe a new name (or just add "Visual", "Vista", a "#" sign or some other McThing to the name).

Then as they release newer versions - the extend it by duct-taping .Net (or some other new MS interoperability API) to it (and not to well they lost, fired or re-classified the original core programmers by now) and, futzing it so it only is cross-compatible with MS Access, Office 2007 or IE tags, etc.

Lastly some other (non-MS) company will innovate in some other market that they don't have thier hand in and they will drop it like an old fish and start to leak FUD about it all over to encourage developers to switch to their "new" competitive technology.

Been there done that, got the big heavy boxes.

Offline Software install and updates

[TechHSV]

I have a network that can't touch the Internet. So it takes for ever to install software because I have to track down all dependencies. If a program needs a dependency it should tell me what it is and where to get it. I shouldn't have to search all over rpmfind or other search engines to install one piece of software. Don't even get me started on updating the OS. Someone needs to package it all together and make it a simple download or have something similar to WSUS. Red Hat has one, but the offline version costs a ton.

Re:Offline Software install and updates

[Ash-Fox]

I have a network that can't touch the Internet. So it takes for ever to install software because I have to track down all dependencies. If a program needs a dependency it should tell me what it is and where to get it. I shouldn't have to search all over rpmfind or other search engines to install one piece of software.

So, why can't you just mirror the repository for the distribution you're using on this closed network?

It wouldn't be too hard to download it to a portable harddrive.

HW mfrs who won't play ball with anyone but M$

[tepples]

10 LET M$ = "Microsoft": REM Slashdot limits the length of a comment's subject

Here's my pick, and I'm sure there are a few others:
SIMPLE hardware compatibility
I want to plug my mobile phone in (without having to figure out multisync) What is the text of the letter that you have written to your mobile phone manufacturer and network operator, and what is the text of the response?

I want to plug in my webcam What is the text of the letter that you have written to your webcam manufacturer, and what is the text of the response?

I want to use ANY video card without having to look up some compatibility list

What is the text of the letter that you have written to your video card manufacturer, and what is the text of the response?

The point is that these issues are often out of the control of the maintainers of any Linux distribution. If a hardware manufacturer won't play ball with anyone but Microsoft, what do you suggest doing?

Seamless Inteoperability
Open and share documents Please be more specific: What documents? Share in what way?

Avialability Of Core Apps
Photoshop (sorry, gimp doens't cut it yet) Please be more specific: What does Photoshop Elements do that GIMPshop does not? Or which set of users are you thinking of who needs those features that are in Photoshop but not in Photoshop Elements?

Games Please be more specific: Doesn't a typical distribution of GNU/Linux come with several games?

Re:HW mfrs who won't play ball with anyone but M$

[spagetti_code]

[Note: I run linux at home (both home PC and MythTV). I usually run Xp at work, but switched to Ubuntu for 6 months, now to XP Tablet]

The first reponses are classic examples from the linux fraternity... Got a problem? here's the source code. Now go away. I'll skip those.

Seamless Inteoperability Open and share documents

Please be more specific: What documents? Share in what way?

Easy... I want to read and write docx and doc file properly. For all their uglyness, they are pervasive. When I ran ubuntu on my desktop, these were the biggest pain by far.

Avialability Of Core Apps Photoshop (sorry, gimp doens't cut it yet)

Please be more specific: What does Photoshop Elements do that GIMPshop does not? Or which set of users are you thinking of who needs those features that are in Photoshop but not in Photoshop Elements?

See here

Games

Please be more specific: Doesn't a typical distribution of GNU/Linux come with several games?

You're joking, right? (Sadly I think you're not)

If you want to see linux make headway, dont make it the users problem. Its a BUSINESS challenge. We've got a *great* technology (Linux) - how can we see broader adoption and more revenue. What do the *users* want (hint - its not to be reading through USB driver help sites wondering why the printer doesn't work).

I want to see linux improve, I want to see linux adoption increase. For two reasons:

1. Linux will get a broader adoption driving more business, growth and better products.
2. Microsoft makes better products when they are under threat.

Wins for everyone!

Audio

[Bambi Dee]

It took me forever just to figure out that most of the commonly mentioned problems weren't my problems and that an arcane .asoundrc setup comprising dmix, dsnoop and rate conversion plugins ("slaves"?) was the answer to getting playback and capture working - and I'm still puzzled with regard to the roughly 10 devices I find listed in the Audacity preferences, none of which really work, or as to why upstanding applications like Kaffeine or Amarok work better when told to use OSS (emulation) instead of ALSA proper...

Finally I decided to buy a modern, super-supported "real" soundcard - and while it played sound just fine, I couldn't get it to record anything at all. Back to store.

I don't want to blame the developers; they probably don't get a lot of support from the manufacturers. I'm no Linux expert (obviously) so if my time weren't worthless and if I didn't basically enjoy Linux I might've given up; the amount of terminology to sort through before you know what anything's actually good for is staggering (LADSPA JACK ALSA OSS Portaudio GST ESD aRts...) It's working now, but I hope I won't have to do this again on another computer any time soon.

(Sound in Doom 3 horribly broken but who cares)

corporate applications

[wasabii]

A competitor to Exchange. A good email client that works with it flawelessly and has all the abilities of Outlook.

CRM software integrated into that.

A single-authenication package that is ready to go. Like AD. This means making all current software use Kerberos flawelessly and fixing all the outstanding issues with LDAP lookups (caching, offline support, etc).

But how does one go from "Python" to "Trac"?

[tepples]

Any case, the single sentence `aptitude install trac` or the double click on synaptic installed a properly running Python environment, Trac itself and all its Python dependencies. How can it be much less difficult than this? How does one make the connection from "this app is written in Python" to "I need a package whose name is Trac"?

Remember: you don't use "Linux"; you use "Fedora 7" (or whatever). So, if you lack the knowledge *stick* with whatever is packaged for your version of your distribution, just like you only install on Vista properly packaged for Vista apps. Trouble is that not enough developers have the time to package their apps for every distribution under the sun. This puts us back to the "this OS has too few apps; I'm sticking with Windows XP because I can more easily find apps that are packaged for it" complaint.

Re:But how does one go from "Python" to "Trac"?

[SanityInAnarchy]

How does one make the connection from "this app is written in Python" to "I need a package whose name is Trac"?

Because the name of the app is Trac. That was just an example -- the old, open source BitTorrent is written in Python, so you would follow the same instructions, only for BitTorrent instead of Trac.

The connection is "I need an app named Foo", so you then go to Synaptic (or Adept, or aptitude, or apt-get) and find Foo, and say "Install Foo". You don't even have to know if the app is written in Python or not -- Python will be installed for you if it's needed.

Compare this to trying to install a Python app on Windows. The only reason things are as easy as they are on Windows is that they will wrap both the Python app and Python itself into one installer, and each app will have its own version of Python, wasting disk space and RAM.

Trouble is that not enough developers have the time to package their apps for every distribution under the sun. This puts us back to the "this OS has too few apps; I'm sticking with Windows XP because I can more easily find apps that are packaged for it" complaint.

It actually doesn't take that much work.

You will only really have to make three packages, at most: A source tarball, a Debian package, and an RPM.

The source tarball means that anyone who knows about make and has time to waste can install it on any distro, or build their own package. It's also close to no work for you, since you probably already have everything you need in a directory somewhere -- you just have to tar it up.

The Debian package will probably work on both Debian and Ubuntu, and also oddballs like Linspire. Bonus points if you want to host your own repository, like certain projects do -- it probably isn't much more difficult. It might also be included in one of the mainstream repositories, but even with no repository, end-users know how to download and double-click on stuff, and it does work to download a .deb and double-click on it in a GUI.

All of the above is pretty much the same for the RPM -- it'll work on RedHat Enterprise Linux, CentOS, Fedora, and even oddballs like Mandriva, it may be included in a mainstream repository, but worst case, people can download and double-click on it.

That's really it. Distros are similar enough these days that if you provide these three forms, you've covered probably 99% of Linux users. The other 1% will be people on really strange distros like Slackware and Gentoo, and if they're using a distro like that, chances are they know how to use your source tarball -- in the case of Gentoo, chances are they made a "package" before you even knew that Gentoo existed.

The pink elephant is (warning: flamebait)

[gefahrmaus]

GUI coherence and ease of configuration are the list toppers.

Most alpha geeks already get it, and know about the pink elephant in the room: Mac OS X is the fabled the "BSD Desktop Platform" for those reasons alone.

<flamebait>
Until there are "Linux Human Interface Guidelines" (LHIG) handed down from a working group, wielding Nazi-like powers across all distros (be they powers of rating, or powers of recommendation)*, Linux will be where it is today: running in Parallels on Macs, performing tirelessly in the basements of corporate datacenters, or hosting websites run by ISVs and ISPs (and even that is being threatened by IIS in the current trends at Netcraft).

Next year, or the year after that, in perpetuity; none of them the "year of the Linux deskop." Each year will deliver more powerful, feature-rich, iterations of Mac OS X sporting a visually coherent interface that "just works."

* Rating: 0-5 stars for adherence to the "LHIG"; or "Recommended for general use, by your mother." Shame the creators of the software to follow the guidelines and end the renegade engineering mindset that "an engineer can design an interface." They can't. Look at how Java engineers fucked up the view (in MVC) in JSF by appropriating DOM and HTML attributes for their component backend.
</flamebait>

Linux is ready

[Skapare]

Linux is ready. The problem is AMD (formerly ATI) doesn't want to deal with this. They (AMD) could choose two different paths to Linux success (either opening the documentation to let Linux developers create the drivers, or create the drivers themselves). Instead, they choose to go with an OS in which an older version is now considered an upgrade. What can you do? Don't buy AMD/ATI components that don't have Linux support or open documentation.

APM & ACPI

[msmithma]

My main desire is more robust ACPI/APM support. My servers motherboard just got CPU frequency scaling support in the latest FC6 kernel update and it's buggy at best (it's over a year old). Furthermore it still isn't recognized by ACPI without forcing. Power savings are a big selling point to data centers (The current main Linux market) and to laptop users (the target consumer if it wants to move into common use). How amazing would it be to say my laptop can run open office for 6 hours on under linux but the same machine can only run MS office for 4?

Why should choosing an UI be the relevant issue?

[Shark]

I think the key isn't so much to pick an UI... The real power would be to make the UI not matter. For example, the NetworkManager part of Gnome. It is far from perfect, but these guys got one thing definitely right. They put work on something that is potentially a godsend for GUI users. Yet they took the care to separate the UI side from the working side. So if someone wants NetworkManager in KDE, all they need is a KDE front-end to it.

I *think* gstreamer might have had a similar idea, but I haven't looked into that much yet.

Basically, my take is that the trend shouldn't be to enforce a GUI standard. It should be to re-establish the culture of separation between the app and its GUI. My message to devs: If you really think your app will be useful, try to make it useful to people regardless of their GUI preference. Heck, even people who use CLI should be able to interface with it.

A GUI should always be considered a 'plugin'. Nothing wrong with writing a default one, but I'd definitely give bonus points to someone who makes sure their app isn't tied to the GUI environment outside of their interface plugin.

Linux is about choice, as a user, I want that choice offered to me. As a dev, I would want to give that choice. Forking isn't all bad, but why fork a whole app just because of UI issues?

Heck, defining this standard properly, I bet you could get both KDE and Gnome to agree on it. Wouldn't be easy for sure, but definitely possible. If both KDE and Gnome agreed that for your app to be compliant to their standards, it has to also be compliant to this app dissociated from UI standard. Everybody wins. KDE apps are one front-end away from seamlessly integrating into Gnome or whatever other environment also complies to this new standard.

Frequent optional updates on the stable distro

[Richard_J_N]

At the moment, most of the people who run a particular piece of software are using a stable distribution, and therefore are usually running something 3-9 months old. This has the following problems:

  - The majority of potential bug-reporters aren't running the latest release. Bug reports are less useful.
  - When you do report a bug, and it gets fixed upstream, you still don't get to enjoy the fix until the next distro cycle.
  - The release-test-fix-use-release cycle is much longer than it would otherwise be, and a lot of talent is wasted.
  - Most users have to live with most bugs for 6 months.

What I'd like to see is every package being built at least weekly from CVS by the distro. Then, the user can optionally install certain specific apps at will from this "testing" repository, without having to update his entire install to unstable. Now, the cycle runs:

    User sees bug; user installs latest binary, and quickly verifies bug still present; user reports bug; developer fixes bug; user gets latest build. (all in less than a week, rather than 6 months).

I know it's possible to do all this by hand, but why not automate it? Most users will happily spend 10 minutes on a bug-report, but not 2 hours. [Ubuntu's "prevu" does help a bit here, but it's still not automatic enough.]

How bout marketing?

[Enrique1218]

Linux is very mature on the programming side. Linux has the basic set of programs available to make a minimally useful computer. Now resources should be allocate to educating the consumer about it. Moreover, a strong push is need to woo OEM into partnership to get Linux pre-install on computers. Furthermore, overtures should also be made to get developers on board as well. Windows is not the be all and end all of desktop computing. Linux should be bankable in some niches. Gain a foothold and start building a base from there.

Why don't apps support file permissions?

[sowth]

The worst thing is permissions editing in Nautilus. ...

This is a big problem with many apps and toolkits I've seen. How many file save dialogs have you seen which allow you to set permissions and the group? I can't recall seeing any. It's like they were made for a system which doesn't have file permissions. This needs to change.

Re:Its been in development for almost 20 years

[marcosdumay]

Well, it is fast (main complain about microkernels), stable and able to run on some computers now. It seem to not have any problem but lack of coders (and no, I'm not helping).

It is also the most viable general use microkernel implementation out there. Microkernels may have no imediate gain, but long term they seem to be much more easier to expand.

Of course, with later Linux developments, we may have a moot point since they are moving most things to the user space already... Linux too can become the first general use microkernel.

Directory Service

[Uhhhhh]

One piece (collection) of software Linux really isn't competing with is Microsoft's Active Directory. I, unfortunately, do Windows IT work and while at times it is incredibly frustrating I have to admit that Active Directory is decent and that it really is way ahead of its Open Source counterparts in terms of ease of use, consistent UI, and software integration. Also, things like Group Policy that allow you to make setting changes or software installations on all the computers you are responsible for really simplifies things.

Sure, you can do these things with open-source software by, say, using OpenLDAP, sending commands via SSH and setting up your own Repositories but it certainly isn't as easy or intuitive as Active Directory.

If Linux had a free directory service that easily supported Windows, Linux. and OSX I think that would be a major step forward for Linux/Open Source in small-medium sized business and cross-platform support would be a bridge to seeing Linux on the business desktop.

Re:Directory Service

[Ash-Fox]

One piece (collection) of software Linux really isn't competing with is Microsoft's Active Directory.

I actually found Novell's eDirectory to be quite comparable actually. It even has more neat little features such as integrating with the identity manager, which unifies every login you need, despite requirements for different passwords and so on. Not requiring people to know the passwords and secured cryptographically. Allowing you to share webbrowser passwords between the webbrowser of your choice at that time.

setting up your own Repositories

I'll admit that setting up a WSuS server is easier.

Re:Directory Service

[Uhhhhh]

I came across FreeIPA today, actually. Seems promising...

A few little buglets

[Richard_J_N]

A few things which need to be fixed:

1)Kcontrol should be scriptable. There's no easy way to apply KDE settings from the CLI. (unlike gconf). An easy way would be to implement DCOP for every item in Kcontrol. Motivation: it often takes me longer to set up KDE as I like/recommend it than to do the rest of the install; copying ~/.kde across is not appropriate for a machine which will be used by someone else.

2)KDE and spaces in filenames should be (optionally) transparently converted to/from underscores. Spaces look much prettier on the desktop, but are a real nuisance at the shell. [For example, xmms does this with playlists]

3)KDE - traskbar - drag to re-arrange the applications. (gnome now does this; a patch was submitted years ago, but has never been merged).

4)Ubuntu is great, but it would be wonderful to have an easy way to turn on the "geek" options. For example:
    * The boot screen - would be much better if it were verbose, in the same way as gentoo or knoppix. These are both outstanding - and visually attractive. Ubuntu (with splash=verbose, or splash=none) however remains monochrome, will not accept vga=794, and is actually too verbose (so much trivial information, no easy way to see what services are starting).
  * Switching back to emacs/bash keybindings. (Ctrl-A = go to start of line, not "select all", etc)
  * Ubuntu doesn't ship all of PHP (eg php5-gmp is not compiled).

5)Fonts. I'm in the sizeable minority here, but I *hate* antialiasing - I much prefer sharply defined fonts to blurry ones. However, this requires properly hinted fonts, which means the MS corefonts, particularly Tahoma. Can we please have some free (not just gratis) fonts that are well-hinted for non-antialiasing?

6)Firefox is a dreadful memory hog! My new machine is a 64-bit system - the major advantage is to allow Fx to address > 4GB of swap.

[I've already reported bugs on these, but they aren't getting any attention.]

Beggars can't be choosers

[tepples]

Your points #1 and #6 are hardware support issues. You chose your hardware False. A donor chose my hardware. I live in a world where individuals get birthday presents, individuals get Christmas presents, and non-profit organizations get tax-day presents.

If you want to run some other OS, it's up to you to make sure that you buy hardware that lets you do that. Are donors smart enough to choose something other than Dell, Gateway, or HP?

Re:Beggars can't be choosers

[Chandon Seldon]

False. A donor chose my hardware.

In that case, your donor may have given you Windows-only hardware. The fact that there's no one actually screwed up doesn't mean that you can run Linux on it, unfortunately. Linux, like any other OS, still only runs well on supported hardware.

Re:Beggars can't be choosers

[tepples]

your donor may have given you Windows-only hardware. So how does someone who relies on donated hardware run GNU/Linux? At some point, buying a Windows license becomes cheaper than buying hardware supported by Linux.

Re:Beggars can't be choosers

[Chandon Seldon]

So how does someone who relies on donated hardware run GNU/Linux? At some point, buying a Windows license becomes cheaper than buying hardware supported by Linux.

That's true. And if minimizing your immediate cash expenditures while using all of your donated hardware is your goal, then that's the point when you'll have to buy a Windows license. My guess is that using GNU/Linux has other advantages that you'd want to figure in to your decision making.

Luckily, most hardware happens to be supported. For many "random" computers, the only hardware component that is unsupported is a Broadcom wireless card. In that case, you have a third option: Plug in an ethernet cable. That's probably cheaper than either a wireless card or a Windows license.

Hell, it's even possible that futzing with ndiswrapper is a good cost / benefit tradeoff for you, especially if you have a bunch of similar machines that the same ndiswrapper setup will work on. I'm just really, really tired of hearing about how "GNU/Linux isn't ready for the desktop" because someone wasted six hours trying to follow poor forum directions about how to get ndiswrapper to work when even that doesn't support their specific chipset.

Oh, to have mod points!

[mad.frog]

Wish I could mod you up, but instead I'll comment: yeah!

Give me a stable API (and ABI) that I can rely on. I don't care which implementation is underneath, but having multiple APIs for windowing systems is really a pointless waste of time at this point.

DirectX, baby

[pilotlicense]

Full DirectX support. Yeah, yeah, I know. The whole Microsoft thing. But it'd be cool to have.

Re:DirectX, baby

[Ash-Fox]

Full DirectX support. Yeah, yeah, I know. The whole Microsoft thing. But it'd be cool to have.

I would say a good chunk is already supported.

Re:DirectX, baby

[pilotlicense]

That is not a good chunk, in my opinion. And, I said *full* support.

Re:DirectX, baby

[Ash-Fox]

That is not a good chunk, in my opinion.

It's enough to run games like World of Warcraft, Oblivion, Half life 2, S.T.A.L.K.E.R.

The implementation of the "Direct3D 8/9 infrastructure" may not be complete, but it's pretty close to it. The majority of other DLLs (beyond dsound.dll) aren't as important as game installers tend to come with a DirectX installer that provides those DLLs (which don't require reimplementation since they would work fine under Wine).

After all has been said....

[SerpentMage]

I looked through the comments and sliced and diced the data... And I think the most essential point was being missed.

How would you refocus Linux? Refocus Linux to what people would pay for. I am not kidding in this. An individual told me that Open Source is nice and great, but what is really nice and great is developing something that you can make money at. Having people part with their money means something and thus until your software is able to do that it is not worth anything.

So I think my answer would be to your question, "Refocus one of the BSD's to something that people would be willing to pay for." Who knows what that might be. I would not even refocus Linux, but would focus on PCBSD, or one of the other BSD's. Less hassle and pains from a crowd perspective. The Linux crowd has reached a critical mass of mess where I fear nothing is going to get going further...

Think of it as follows. If you truly had the money focus on a BSD and talk to companies and say, "Hey here is a new market and you can keep your drivers binary. And guess what I have X $$$ to focus on marketing this software."

Multimedia authoring.

[Anonymous Coward]

Personally, I'd focus on Multimedia authoring, it's always been Linux's weak point.

First, audio composition and editing, as well as a Virtual DJ studio. The backend (ALSA + Jack) is brilliant, but the lack of synths, mixers and whatnot of the caliber of Absynth and Traktor (from Native Instruments) and Live (from Ableton) makes ALSA/Jack fairly useless. 0 latency is kinda useless when the closest thing to a good softsynth is a as much of a pain in the ass as Ardour and Rosegarden are to use in comparison. A Linux analog to VST would be brilliant as well, and not only for the plugins, but mostly for, lets say, my current setup: Ableton Live at the core, with Reason plugged into is via VST as a drum machine, and absynth plugged in as the synth.

Second, raster graphics. As much as people keep repeating that Gimp is as complete as Photoshop, it isn't. CMYK support, 24 and 32-bit colour support, support for excessively large files, better tablet support, are all very much needed in a bitmap editor. I'd certainly throw much resources in Mr. kanzelsberger's direction, for work on Pixel Editor, which is a truly brilliant application.

Third, Vector graphics. There really hasn't been a truly top-tier vector drawing application for Linux since Corel Draw 9 was briefly ported. Inkscape is neat, but it's far, far more useful as a GUI frontend to SVG editing than a full fledged Vector drawing application. Xara Xtreme has promise and potential, the Windows version of Xara was ever even quite close to being on par with Illustrator, Freehand, or Corel Draw, but it has promise and it's probably the best bet right now.

Fourth, 3D modeling. I won't argue that Blender isn't ridiculously powerful, it is, there's no denying that. But it is also the single least intuitive program I've had the mispleasure of using. I've always referred to it as "a programmer's modeler" having "an artist's modeler" to go along with it would be great. (and don't even mention Maya, first the price tag is/was obscenely ridiculous, second Autodesk, since acquiring Maya, has been gutting it for use in 3D Studio Max. Maya is barely a pale shadow of it's former glory).

These may not seem important to many, and the tools currently available on the platform may be enough for quite a few. But as a DJ, musician, visual artist and graphic designer, well, these things are pretty well the reason I even use a computer.

Re:Multimedia authoring.

[Ash-Fox]

Personally, I'd focus on Multimedia authoring, it's always been Linux's weak point.

Agreed.

First, audio composition and editing, as well as a Virtual DJ studio. The backend (ALSA + Jack) is brilliant, but the lack of synths, mixers and whatnot of the caliber of Absynth and Traktor (from Native Instruments) and Live (from Ableton) makes ALSA/Jack fairly useless.

Indeed, the Linux desktop is also very heavily in decent video editing software (if you ignore the fact you can run some decent software under Wine).

raster graphics. As much as people keep repeating that Gimp is as complete as Photoshop, it isn't. CMYK support, 24 and 32-bit colour support, support for excessively large files, better tablet support, are all very much needed in a bitmap editor. I'd certainly throw much resources in Mr. kanzelsberger's direction, for work on Pixel Editor, which is a truly brilliant application.

Does Krita appease you?

Vector graphics. There really hasn't been a truly top-tier vector drawing application for Linux since Corel Draw 9 was briefly ported. Inkscape is neat, but it's far, far more useful as a GUI frontend to SVG editing than a full fledged Vector drawing application. Xara Xtreme has promise and potential, the Windows version of Xara was ever even quite close to being on par with Illustrator, Freehand, or Corel Draw, but it has promise and it's probably the best bet right now.

From the same software suite as above, there is Karbon, I'm also aware that OpenOffice.org and StarOffice have vector art capabilities, but I haven't managed to look at that yet. There is also sodipodi.

XML

[MidoriKid]

In fact I'd spend a lot of the money on getting everyone (or as many projects as I could) to agree to a configuration file format that could easily be interpreted by an application. A one-size-fits-all library could be written to get the settings from file into memory and back again, then it would just be a matter of organising that data into a front-end that's meaningful for the user.

Sounds like XML and the myriad of libraries that support it. The problem with XML (ignoring the buzz inspired bias against it) is for any given application, a custom configuration file format would be much less verbose. As far as "one-size-fits-all" goes, I can't think of any .conf I've edited that couldn't be expressed in XML.

CAS/CAD systems...

[jopsen]

The only things I use Windows for is MathCad and AutoCad (both required by my school)... Now I guess AutoCad is too big, and probably impossible... But Linux already has CAS systems like Maxima, even commercial ones... But a free (as in freedom) CAS system with a GUI, that would run on both Linux and Windows... Would without doubt replace MathCad, both on my school and my computer, very fast. Now I know we've got different frontends to Maxima, putting LaTeX ontop of it, but none of them are usable for non-geeks, like my mates.

Re:3D - buy Nvidia

[Anonymous Coward]

3D support is 100% with all NVIDIA cards.

Patents

[tepples]

There isn't a single PDF reader besides Adobe Reader that supports subpixel rendering which makes the font rendering hurt my eyes. Are you willing to have your installer pause until 2019 while waiting for Microsoft's patents on LCD subpixel rendering to expire?

Maybe it's the installer?

[SanityInAnarchy]

I've always been able to resize Windows partitions with no problems, but I use commandline tools like ntfsresize. Maybe the installer uses something else?

I do know that it does have to defragment a bit -- it does move files out of the way in order to shrink the partition.

Re:Maybe it's the installer?

[number11]

I've always been able to resize Windows partitions with no problems, but I use commandline tools like ntfsresize. Maybe the installer uses something else?

QTParted. Which doesn't seem to defrag.

What does GIMP lack?

[tepples]

Bring GIMP up to the point where it would compete with PS. What, specifically, do you find GIMPshop to lack compared to Photoshop Elements?

Work on polish. Why, when you can make more money working on english?

Re:What does GIMP lack?

[alfredo]

In PS you can fade a filter or add effects. As yet I haven't found a way to do that in GIMP.

I use GIMPshop. It's good for much of what I do.

Polish women are beautiful

Re:What does GIMP lack?

[tepples]

In PS you can fade a filter Have you tried duplicating a layer, filtering the copy on top, and turning down the filtered copy's opacity?

Re:What does GIMP lack?

[alfredo]

Hmmm. I will give that a try. Thanks.

audio and video

[stites]

Right now audio and video on desktop computers is a war of competing proprietary standards. Microsoft, Flash, and a host of other proprietary companies compete based on their own proprietary standards. The game companies are forced to invent their own unique standards because none of the other standards run fast enough to guarentee good game performance. DRM has reared its ugly head in an effort to be the antithesis of open standards.

I think that we should put a concerted effort into creating good open standards for all audio and video applications. Then we should build the infrastructure, shared libraries, programming aids, etc. necessary to support audio and video applications. Then we should write good application programs for the entire range of audio and video applications. And throughout the development cycle we should emphasize performance so that games can run using open standards instead of having to roll their own.

This amount of structured development is too much for the typical open source evolution approach. I doubt that a comprehensive system will evolve through the actions of open source developers acting independently. If somebody has the money to finance a large scale open source development effort then the audio and video areas are a good place to spend the money.

-----------------------
Steve Stites

Re:audio and video

[Salsaman]

I quite agree. In fact I have been working in this area for the last 5 years, trying to get projects like LiVES, videojack, Weed/Livido and OMC off the ground. However, there is no funding for this to be found anywhere - I have applied to literally dozens of groups, including the FSF, google and other places, but nobody seems interested in free standards in multimedia.

Anyway, you can check out some of these projects:

http://lives.sourceforge.net/
http://www.piksel.org/videojack
http://livido.dyne.org/codedoc/
http://lives.cvs.sourceforge.net/lives/lives/weed- docs/
http://lives.cvs.sourceforge.net/lives/lives/OMC/l ives-OMC.txt

.lxe files

[CrazyJim1]

Similar to .exe files, they would allow interoperability between Linux flavors. Currently people need to install or compile to get stuff to run between Linux systems if not using Java.

Synaptic?

[tepples]

I want to play a fun game that works. StepMania? Frozen Bubble? Nexuiz? The games that come with GNOME or KDE? Wii console through xawtv?

With window$ you just double-click on an exe. With Linux you download a tarball, go to the command line, do a tar zxvf thefile, cd to the directory, type ./configure and then wait. Why install from source as an end user? Why not just use your distribution's package manager to install the game from a binary package?

So maybe we just need a standard library? Possibly a meta-library that programs can reference at install and automatically install the missing libraries? They have one. It's called APT.

Oh and while I'm dreaming here, let's just scrap the keyboard/mouse and go with full-blown voice recognition so we don't even have to type. I'd like you to try reading aloud every Slashdot comment that you have posted within the last few weeks, especially in a noisy room. Voices wear out quickly.

Re:Synaptic?

[WhiteHorse-The Origi]

APT is a package manager which uses repositories. It's great and I use it, but it requires someone to create a repository with a package for my platform. I think APT is a good start but in the future it would be great if it could grab any software for which the source is available, resolve all dependencies, compile, and install it. Maybe even ask me if I want to make a desktop icon or menu shortcut.

One example game I tried recently is called Cube or Sauerbraten. It wasn't in the repo so I downloaded it. The canned binaries don't run on my amd64 so I had to compile it from source. I did use Synaptic to search for and install the dev libraries needed but it was the usual: try to compile, wait for errors, figure out which library is needed, then try again. I know this is getting less common and things have improved, but I do like having the latest version of software when the developers release it, rather than when it gets added to the repo a few months later. A perfect example is Wine. It's constantly updated and new releases are always coming out. Each release adds significant support for things I like to do. They probably have their own repo I could add but it won't have an amd64 package. So in turn I manually override the repo version, download and compile the latest. I am learning SVN and it's helping to keep things current but it's still a pain to track down all the latest releases when my computer could automagically do it for me(similar to apt but one step higher).

Admin Interface

[tom581]

I'd like to see a standard established for the system administrative interface. Of the two major commercial distributions for Linux being RedHat and SuSE, I'm not fond of the RedHat approach of a bunch of disparate tools called "system-config-splat". On the other hand, SuSE has been very consistent with their use of YaST across the versions, and provide a text mode interface for those situations where you have to network into a system to perform management. The RedHat "system-config-splat" tools require a GUI, and tunneling GUI over an ssh connection is slow. YaST is rich with plugins for managing virtually all aspects of a Linux system. RedHat and "system-config-splat" is lacking in that area. By making Linux easier to manage, then it will be adopted by more inexperienced shops. This will win over the Windows admins. And face it, when it comes time for OS selection, the IT Team has a lot of say over what gets installed on the systems.

Re:Admin Interface

[Ash-Fox]

I'd like to see a standard established for the system administrative interface.

bash!

I hope you're joking.

[SanityInAnarchy]

You give an example of a text file, that's easy to understand, and is four lines long. You then suggest literally five times as much XML, just to describe what's a valid value in the text file?

Alright, first of all, "connectionsMax" doesn't need to be validated for anything other than being a number. The actual implementation may have a limitation like you suggest, but I don't really see a reason for it to, except maybe theoretical OS limitations like maximum number of threads running at a time. It seems you agree, because you suggest 0 as a min -- in config files, setting a max of 0 usually means "unlimited", because presumably this is a config file for something which needs at least one connection.

Limiting SSL to boolean, I can see. I can even see doing validation -- certain things, like visudo, already do validation on save, and other things, like Postfix, do validation right before a restart, so that Postfix will keep running with the old configuration if the new one is invalid. Most things can deal with invalid lines in their config files anyway.

I don't necessarily disagree with the features you want, the XML just seems superfluous to me. And your "export/import" works fine for those who like vi/emacs/nano/whatever -- just check out visudo, vipw, vigr, etc for an example of how it's done. But it doesn't work so well for a graphical editor, where people could have more than one file open at once, not want to close it just to save, etc.

Also, you do underestimate text files. There's plenty of room in there to throw a lot of comments (including your "short" and "long" examples), they're as "user friendly" as anything else if you can find them -- or maybe I just don't know any users who find it easier to click a checkbox than to type "true" -- and they are actually much easier to navigate (using a search) than a GUI representation of them would be, unless you did a lot more work on that GUI than your XML example.

In fact, it's only recently that GUI configurations became searchable, and it's still not very good, compared to what I can do with a text file.

The downside is, you don't always know which text file to edit, or how to get to it. Also, text files are slow, for anything more than simple configuration. But that's about it.

A good interface is both simple and deep.

[David Gerard]

We're talking about broad focus here. I think I can put it in a sentence:

Good interfaces are friendly to both technophobes and geeks.

Ubuntu is a winner because it treats newbie confusion as a reportable bug that needs fixing. That's its big win over Debian, who (love 'em as we do) treat newbie confusion as a cue to supply pointers to manuals.

But that's just a start. The way to get there is to always make your interfaces technophobe-friendly but with depths to reward the geek.

Examples: MacOS X. LiveJournal. Firefox.

Every part of the system needs to be considered (or reconsidered) in terms of being obvious as well as deeply right.

Software integration

[o517375]

Linux would benefit greatly from the integration of existing software. What I am thinking primarily of is SMTP. IMAP, groupware, and mail clients. These need to be tightly integrated into TWO applications similar to Outlook and Exchange. I realize there are applications out there that appear to be this, but they are really kludges of existing applications. Don't misunderstand; I like Postfix and Exim, along with the rest of the apps, and we use them now on the gateways. But I would merge all of the code and add a ton of features. One of the first features would be wireless handheld bidirectional integration. Since I'm not a developer, I can't say how this could be done. But is at the top of my wish list. We use Exchange right now, only because of the wireless handheld Blackberry integration and easy contact and calendar. So I am waiting for this to happen

There needs to be integration of a database and a RAD front end similar to Access. Access is the only reason we continue to use Microsoft Office. We could easily switch to Open Office except that our business relies heavily on Access. We could easily switch given the opportunity, because the backend to our Access application is Mysql. So I am waiting for this to happen also.

Wish lists can are rants in disguise. But you asked.

Re:Software integration

[maxwell demon]

Linux would benefit greatly from the integration of existing software. What I am thinking primarily of is SMTP. IMAP, groupware, and mail clients. These need to be tightly integrated into TWO applications similar to Outlook and Exchange. I realize there are applications out there that appear to be this, but they are really kludges of existing applications. Don't misunderstand; I like Postfix and Exim, along with the rest of the apps, and we use them now on the gateways. But I would merge all of the code and add a ton of features. One of the first features would be wireless handheld bidirectional integration. Since I'm not a developer, I can't say how this could be done. But is at the top of my wish list. We use Exchange right now, only because of the wireless handheld Blackberry integration and easy contact and calendar. So I am waiting for this to happen

Actually IMHO the solution here would be to remember the Unix tradition: Have your programs do one thing, and do that right. Ok, an email program does one thing, right? Well, no. Let's ignore that current Email programs try to be everything even remotely connected to communication (why the hell do I have to configure an Email account when all I want is to have a calendar, because I'm completely satisfied with the other email client I already use?). Even simple email today consists of several separate services: There's the communication with the server, there's local storage and management of mail, there's displaying mail, and there's things like spam filtering mail etc. Now what if I like the Evolution GUI. but consider the Thunderbird spam filter superior? (Disclaimer: That example is simply made up and should not be construed as statement about the quality of those applications) Well, I obviously have the choice if I prefer the better GUI, or the better spam filter. But why? Those parts are largely independent from each other. There should be a standardized spam filter API, and then I could mix and match any spam filter implementation with any GUI implementation.

Also, think of the spell checker. Firefox has a spell checker. OpenOffice has another spell checker. Then there's ispell. And I'm sure you'll find other applications which come with their own spell checker each. Every application has its own dictionaries. All will find different spelling errors, and if I teach a certain spelling to one app, none of the others will know it. Ideally, there would be a simple, well-developed spell checking API, which every program from Firefox to ispell would use. There could certainly be several implementations of that API, and I'd be able to select the one I like best, knowing that every application I'll use will then get that same spell checking. If I teach the spell checker a new word in OpenOffice, it will be available in Firefox as well, and if I spell-check in XEmacs, then again it will know the new word.

Note that having several programs doesn't mean no integration. For the user it doesn't matter at all if the calendar window opened comes from the email program or from a separate calendar program communicating with the email program through a standardized calendar API. From the user's view, he tells his email viewer to open a calendar window, and that calendar window appears. However, with modularity, if the user decides to use another email program tomorrow, he still can use the same calendar. He'll not even have to import the data, because it's, after all, the same application which actually manages the calendar data. Of course, he can also switch calendars if he likes another one better, without the need to learn another email interface, because he doesn't need to change the email program just to use another calendar.

Re:Software integration

[o517375]

I think your point is good in theory. First you mention Unix tradition. How attractive is the Unix tradition for desktop. IMHO, not good. Unix desktop has very low market share and very limited feature set though it's been around since before Windows. Second, software integration as you describe only works well in a perfect world. In a complicated reality, when you throw things like SSL integration for example, kludging together all of these things becomes a system administrator's nightmare, even with great package management. Thirdly, developing the parts without a clear vision of the whole, to me, runs clearly counter to software development. The developer should start out defining the objective. If that objective involves many complications (such as for an email client) because people want many calendar functions integrated with email functions and contact functions, well what do we say? You can't have that because it's not good practice? Or do we say, "Hell yes, I'm working on it." The real problem with end user GUI open source code projects, such as an Exchange-Outlook replacement, is that the objective is so complicated resulting in lots of "boring" code that no one wants to really write for free. Even Zimbra which I hear is very good is mostly a kludge of existing software and it's expensive for the licensed version.

Re:Software integration

[maxwell demon]

If that objective involves many complications (such as for an email client) because people want many calendar functions integrated with email functions and contact functions, well what do we say? You can't have that because it's not good practice? Or do we say, "Hell yes, I'm working on it."

The user wants integration at the interface level. He doesn't care the least bit how the code providing this interface is organized. Well, except when the monolithic code structure bites him, because he has e.g. to maintain his private spellings in several applications, or he can't use his calendar without an email account.

Think about it: Normal operations already involve a lot of applications. If you open an application under Linux, what you see is actually determined by at least three processes: The application process itself, the window manager process (e.g. Metacity), and X. Most people probably are not even aware of that fact, because things fit together quite seamless.

What the user cares about is that he can open his calendar from within the email application, and that data is shared between those applications where meaningful. For that, you don't need to have both in the same process. Also note that having them separate doesn't mean you cannot develop them together (I guess the first web server and the first web browser were developed together, despite the fact that they are different applications which even usually run on different computers).

The real problem with end user GUI open source code projects, such as an Exchange-Outlook replacement, is that the objective is so complicated resulting in lots of "boring" code that no one wants to really write for free.

Well, another problem should be that one would have to know Exchange-Outlook well enough to know what the users actually expect. Something which is especially unlikely for hardcore OSS people, i.e. those who'd be the most likely to write such a thing.

Re:Software integration

[o517375]

Well, another problem should be that one would have to know Exchange-Outlook well enough to know what the users actually expect. Something which is especially unlikely for hardcore OSS people, i.e. those who'd be the most likely to write such a thing. I would think that an Exchange/Outlook replacement would be of far more interest to an Open Source programmer than KDE when Gnome already existed, and Open Office. My God! Who wants to spend time cloning Microsoft Office! I break out in hives whenever I have to even use it.

Re:Software integration

[maxwell demon]

Of course Open Office wasn't just written as OSS from scratch, but it came from opening up Star Office. According to Wikipedia, the development of Star Writer already started in 1984. I know for sure that I got Star Writer preinstalled on the PC I bought in 1995. Given its long history, I'd say it's not really a clone of MS office (I'm not sure if MS already had MS Word as early as 1984, but it certainly wasn't dominant at that time).

And BTW, KDE came first; Gnome was started because at that time there were licensing issues with KDE (Qt did only come with a GPL-incompatible license back then). I'm sure Gnome wouldn't have been started if the KDE licensing issues had not been there.

Reducing the CHAOS that is Linux

[enmane]

Sorry but it's true...

I've been using linux for 7 yrs now and after 5 yrs of hardcore use and then one yr of moderate use, I'm now finding myself also using Windows more often and here's why...

Programs!

I've had this theory that the porting of FOSS programs from Linux/Unix to windows will hurt Linux and those on the opposite side of the fence said "but wait, if we create the programs for Windows then that's good; when they switch to linux they will be able to use the same apps and feel at home." It's believable but it isn't factual. The truth is that linux changes too quickly and there are too many changes within distros.

The beauty of Windows - I need program A so I just go to the developer's site and download it - thanks to FOSS software. Try that with Linux - it can't be done; there are too many distros and versions out there for an rpm, deb, etc to get built even within a single distro.

Before you go out and tell me about apt, synaptic, etc - I'm aware. Each and every one of these tools relies on a developer to create the package and add it to the repo - this is another obstacle. Remember, in Windows - I go directly to the developer and take the package and install. In linux, I'm waiting and hoping that a developer will hear my plea to add the package to the repo before I can install it. Sure, I could build it myself BUT that's provided my distro doesn't do anything funky with file locations, versions, etc.

That's the number 1 reason I find myself in the Windows world these days. I'm tired of waiting to have them built and added and I'm tired of the warning that if I do it myself that I could "break" something. Just fire up WinXP and download/install to XP - now that my favorite Linux apps are available on XP ;-)

1) SOOO - my vote goes to the linux distros unify how their distros work so that developers can build packages.

other than that,
2) decent filesystem - looking in /mnt for partitions isn't intuitive; having to go "here" for the executable and "there" for the config file is pretty lame too. Oh, you want a config file - go check in /etc/, you wan't a networking config file, go to /etc/sysconfig/netwo.... - It's not THAT bad but it's not nearly as easy as it could be.
3) get ACPI/APM to just work and stop forking around! There are so many tools out there and each with their own special niche - it's maddening. What does my distro do? Add them all and hope the user figures them out! Just maddening.
4) updating the look of programs - the old unix X-window styling is still around on many programs. XFig for 1 but it isn't alone; great program but being ignored due to the ugly interface. It would be nice if X took care of updating the interface.

In short, here's how I'd define linux progress "2 steps forward, 1 step back."

Unless we're talking about OO speed then it's "1 step forward, 2 steps back." Funny how the "individual" program Swriter takes longer to load than "Staroffice 5.2" which was this integrated desktop environment and deemed "bloatware" by the OO developers. They blew their horn and claimed that they were going to strip it down and speed it up - LMFAO - 6 yrs later and SO 5.2 wipes the floor with OO, in terms of speed. I'm praying that someone comes along and rewrites OO such that it performs only twice as slow as MS Word - as opposed to the 5-7x slowdown currently enjoyed.

Re:Reducing the CHAOS that is Linux

[Ash-Fox]

I've been using linux for 7 yrs now and after 5 yrs of hardcore use and then one yr of moderate use, I'm now finding myself also using Windows more often and here's why...

I've been pretty much using most modern OSes everyday for the past ten years.

Before you go out and tell me about apt, synaptic, etc - I'm aware. Each and every one of these tools relies on a developer to create the package and add it to the repo - this is another obstacle. Remember, in Windows - I go directly to the developer and take the package and install. In linux, I'm waiting and hoping that a developer will hear my plea to add the package to the repo before I can install it. Sure, I could build it myself BUT that's provided my distro doesn't do anything funky with file locations, versions, etc.

Linux Standard Base solved many of these issues years ago.

1) SOOO - my vote goes to the linux distros unify how their distros work so that developers can build packages.

LSB has already done a lot to help this, LSB type packages have worked quite well for me.

2) decent filesystem - looking in /mnt for partitions isn't intuitive; having to go "here" for the executable and "there" for the config file is pretty lame too. Oh, you want a config file - go check in /etc/, you wan't a networking config file, go to /etc/sysconfig/netwo.... - It's not THAT bad but it's not nearly as easy as it could be.

I've used Linux distributions that used self contained application directories (like C:\Program Files\App name\) for everything. This isn't a Linux issue, Linux can do this. This is a distribution/POSIX userland specific issue.

3) get ACPI/APM to just work and stop forking around! There are so many tools out there and each with their own special niche - it's maddening. What does my distro do? Add them all and hope the user figures them out! Just maddening.

Sounds like Mandriva Linux with it's insane amount of 3rd party kernel module support. Majority of Linux distributions don't go to that extent.

4) updating the look of programs - the old unix X-window styling is still around on many programs. XFig for 1 but it isn't alone; great program but being ignored due to the ugly interface. It would be nice if X took care of updating the interface.

Feel free to contribute that suggestion to the maintainers.

I'm praying that someone comes along and rewrites OO such that it performs only twice as slow as MS Word - as opposed to the 5-7x slowdown currently enjoyed.

I've noticed StarOffice 8 (Based on OOo) opens up pretty instantly. Looks like you have your wish.

Re:Reducing the CHAOS that is Linux

[enmane]

[quote]I've noticed StarOffice 8 (Based on OOo) opens up pretty instantly. Looks like you have your wish.[/quote]

with or without the preloader? Feel free to try 5.2 and let me know what you think. I'm not sure which distros are using LSB at all. Feel free to educate me as to the percentage of Linux distributions that are using LSB compared to the ones that are not.

Re:Reducing the CHAOS that is Linux

[Ash-Fox]

with or without the preloader?

Never seen the preloader under Linux. Don't see anything like it in my process list either.

Feel free to try 5.2 and let me know what you think.

I can't seem to find any trials of it on Google.

I'm not sure which distros are using LSB at all.

LSB is generally 'supported' by most desktop Linux distributions such as Debian, Ubuntu, Mandriva, Redhat, Fedora, SuSE and so on. The only major distribution I can think of that does not support LSB is Gentoo -- Although that shouldn't be too surprising.

The smaller distributions that deal with just a single specific problem like a mini-router-firewall, livecd desktop system and so on, do not likely support LSB.

Please note, just because a distribution supports LSB, does not mean that it's packages are LSB (generally packages that come with the distributions are distribution specific for obvious reasons).

The biggest problem with LSB however is mainly very few developers use it. Many OSS projects tend to rather make only the source available or some very simple .tar.gz file that contains some compiled binaries. Even bigger OSS projects like Mozilla and OpenOffice.org choose to ignore it.

That said, one of the most insane? LSB things I've seen was, StarOffice 8. It is in LSB, but due to the fact Sun decided that it wasn't enough to make sure the suite would install, they included parts of a LSB environment with the packages. Wrapping it all in a java based installer which comes with it's own JRE (really overkill).

Re:Reducing the CHAOS that is Linux

[enmane]

Then I'd like to use your system ;-). OO and SO definitely don't open THAT quickly for me and I've used just about every version. Here are some places to try out 5.2 and note that it hasn't been supported since Sun picked it up so it's dated but the speed is still there...

I googled "download staroffice 5.2" and got the following hits,
http://www.trustmeher.net/freeware/staroffice-dn01 .htm
http://download.reitoria.ufsc.br/staroffice/ingles /

Anyhow, I hope that a project like LSB gains acceptance - I use PCLinuxOS - a great distro. #2 on Distrowatch after Ubuntu and with a fraction of the support. It doesn't look like LSB is doing that much to me though - great idea but as you've already stated, the implementation isn't working out all that well, just because everyone has their own ideas on how to do things "better."

One word: PowerTOP

[Anonymous Coward]

I would massively analyze all of the code in applications and the kernel with PowerTOP. PowerTOP is a tool made by Intel that monitors how much energy each application uses over time. Performance per watt is the name of the game. Linux developers need to use this tool to profile all of the applications and modules to shrink the Linux energy footprint. This will give Linux and the associated suite of applications tremendous appeal for laptops, embedded devices, servers, and the ever shrinking desktop computer.

It's amazing some of the things the PowerTOP team have found so far. Things like the blinking cursor of a GVIM sending an interrupt that brings the CPU out of sleep. Even the most innocuous applications can contribute to significant energy waste. We are talking about over 50% improvements in the energy efficiency of a computer with a few simple changes in the software.

Imagine if all of the applications in the default distribution were fixed to avoid unnecessary power drains. First of all, the performance would be better. Second of all, the energy savings would be so huge over Windows it would be insane to use anything but Linux on a laptop or server. With any hope distributions will start using PowerTOP to evaluate whether applications are ready to get merged into the distribution so users can feel safe they can install applications and not sacrifice battery life or energy efficiency in the data center.

Re:Its been in development for almost 20 years

[Anonymous Coward]

Saying there is no immediate gain of a microkernel oversees one crucial point: security!

Due to a set of well defined (small) interfaces the kernel is much more secure than a lump of processes all running in the same kernel space. Thats where Linus is wrong IMHO.

How many resources do we get?

[SanityInAnarchy]

I'm coming a little late to this discussion, and I'm afraid I'll be lost in all the "too many distros" and "my wifi card doesn't work" spam. But I do have a few ambitious ideas that no one else seems to have mentioned...

First, stop writing stuff in C.

Recently, I copied all of my important stuff onto another hard drive and then back, and lost most of it, due to a buffer overflow in cpio. I should have had backups, but then again, if I'd done backups the way I archived stuff there, I could have run into the same problem, where my backups would all be corrupt, and I wouldn't know it until I went to check.

So, my first project would be to figure out what language to use, and develop it to where it's production-ready, and supports enough low-level stuff that we can write most of the kernel in it. If no such language exists, I'd write a new one. So many languages are close -- lately I like Python, Perl6, and JavaScript -- but none of them are anywhere near being a complete win over C, even for applications. The best-looking ones are also entirely too weird -- Haskell, Erlang, and so on.

With the right language/environment, I wouldn't even be afraid of running everything in ring 0, except for legacy code. I realize why this is usually considered a bad idea, and I have a solution, but I'm trying to keep this post short...

This would involve completely rewriting the kernel eventually, but it wouldn't have to be done all at once.

Filesystems are another area of interest. I would rather get the "kernel rewrite" above at least started before tackling this one, though. Right now, the choices are: Write your FS in the kernel for speed, or write it in FUSE for flexibility. I'd rather just write it wherever (since kernel/userspace would have no meaning in the above "kernel"), and have it perform well and be able to do everything I want.

There are some good starts here -- both KDE and Gnome have a VFS layer in userspace. Apps use URLs to access everything -- files are in the "file:/" hierarchy, but there are others. For example, in KDE, I can type: "fish://user@host/some/where", and it will use SSH to access those files.

But the problem with these approaches is that they aren't supported by legacy apps. Sure, you can wrap them in FUSE, but then you take a large performance hit.

The other problem is that there's the GNOME VFS and the KDE VFS, and as far as I know, they don't cooperate. That's my rationale for doing it at the "kernel" layer -- every app would be forced to use one API, or have no access to the filesystem. You could still have competing implementations, but at least you don't have to worry about supporting one or the other in your app anymore.

The filesystem API itself is problematic, too. Right now, there's no reasonable way to do well-performing transactions, without practically implementing a filesystem on top of the filesystem. This is what both MySQL and PostgreSQL do, by the way -- they just throw everything together into one file, and implement their own journaling. You may have noticed that there are tons of filesystems for Linux, and they all suck -- this is one reason why. I don't believe write barriers are enough.

I know about ZFS. For some reason, whenever I mention these obvious problems with the POSIX filesystem API, people point to ZFS as if it magically solves all filesystem issues.

All ZFS demonstrates is that the Linux kernel's filesystem API is deficient, that the layers we've chosen simply will not work. Redundancy, multi-device support, multiple machines, bad block relocation, cryptography, compresson, indexing, and relationships are all things I think a "filesystem" should have to deal with. Right now, here are the solutions I know about:

- RAID will do redundancy locally. DRBD will do a very simple RAID-1 over several machines. They obviously don't share code, or DRBD would be able to do RAID-5, for example. And they operate at the block level -- good, because you can run any filesystem on them, but bad

Re: Os X vs. Linux

[Eli Gottlieb]

Someday Apple's lockin may very well go as far as M$'s. But people will still buy Apple machines on that day because, unlike Windoze, OSX actually works extremely well on its target market of machines and supports the features people want.

History shows that people will support a regime that takes away all their freedom but gives them what they otherwise want.

Disclaimer: I run OSX with a virtualized Linux machine for development.

this is what i would do

[sakura the mc]

i would focus all that fanaticism on freebsd development

Re:start over from the gound up

[Anonymous Coward]

You missed the part about "stinking" feces. Oh man, I just took a HUGE Linux! I feel 5 lbs lighter!

1 dist to rule them all.

[Anonymous Coward]

this is something that I have been saying for a long time. But as long as there are 20+ different versions of Linux it will never get more that a fragment of the market share.

Upper level management are not interested into why we use this or that and "what do you mean it does not run as easily on our infrastructure".

Oracle/DB2/SAP or what ever future applications, it has to run and it has to run on what I bought.

Re:1 dist to rule them all.

[Ash-Fox]

this is something that I have been saying for a long time. But as long as there are 20+ different versions of Linux it will never get more that a fragment of the market share.

I can apply this to windows too:

this is something that I have been saying for a long time. But as long as there are 20+ different versions of Windows (Vista versions, XP versions, 2003 versions etc)

Upper level management are not interested into why we use this or that and "what do you mean it does not run as easily on our infrastructure".

I have seen many examples of where Windows does not run as easily and as well on Windows infrastructures as Linux servers and Linux workstations in certain circumstances (such as poor I/O in Windows making the throughput rather horrendous compared to Linux running on the same hardware, effecting the use of high traffic services such as fileservers - including webservers, database software etc) and vice versa.

Oracle/DB2/SAP or what ever future applications, it has to run and it has to run on what I bought.

Would be nice if XP SP2 ran on my 'corporate grade' laptop which was "designed for Windows XP" (according to the sticker) -- graphic card drivers do not work, sound card drivers do not work, internal wireless does not work (they did under SP1, but there have been no future driver updates).

I am sorry, I really don't see anything that holds water in your arguments.

what is the point of the question?

[Anonymous Coward]

Do you want to know what to improve about Linux?

Or do you want to know how to compete with M$?

I've been using computers most of my life. I would say I'm an advanced user though not an expert. I have no college-taught knowledge but I would say I'm very comfortable with them and I can build and troubleshoot my own computers.

Yet, Linux is daunting and I've yet to try to install it on any system. (Although I was tempted to a short while back and almost did)

Since I haven't even tried to install Linux, perhaps what is most needed is an improvement in the "advertising" and education about Linux. What is Linux? How can it help me? Just how hard is it to install? This stuff should be common knowledge to anyone computer literate, not just the few who want to boldly go...

The answer to the second question is simple. If you want to compete with Windows, Linux needs games. Not just a couple, not just a handful. It needs a quality gaming platform that's extremely easy to build games for. If you want an example, check out XNA, Torque and the others.

If you say "well it already has that" then it goes back to my first answer. Education. (not schools, but public awareness)

As a pretty hard-core gamer the only thing I know about linux and games is that it's good for servers but there's a very tiny handful of games for it, less than mac. I never hear about great, easy-to-use game creation systems.

Am I wrong? Again, consumer/public awareness, education.

(hah.. it's up to you guys to figure out how ... but if linux had exactly the same games as windows, I'd switch asap)

And a few things more....

[Anonymous Coward]

How about restoring the Shred program in Konqueror and in every application as an option on
the 'edit' menu. Even only doing this for 'root' would be a good start. As it is now, linux is a security risk for every user. We have read all the specious clap trap and lies about 'trouble with erasing magnetic media' that were used to justify exposing all linux users to spyware and phishers. Until and unless linux makes this change, all future releases of its kernel are worthless and dead on arrival as an install no matter what other 'improvements' it makes.
    Once shred is restored, how about including an iron low level disk editor capable of all that the old old norton utilities could do for DOS. A simple bulletproof way to share and participate in windows workgroups in a netware environment, not TCP/IP would be quite welcome. I say netware and not TCP/IP as the internet has become a cesspool of malware, some of it private and some of it international totalitarian government sponsored. Just check anyones router logs and see just how many '.cn's appear in the attempts at access. Another would be ways to play old windows games in virtual win98 or win2K windows would be nice...and to network them with other linux/virtualized win2k users over a home network would be fine. That would take away the only thing that windows is good for....games...and give it to us in our linux.
LInux 2.4 for now, as 2.6 has no shredder and is unsafe.

Amarok and gxine both use lib-xine

[eean]

Amarok uses lib-xine and gxine obviously does, so your example isn't valid.

The point is kind of. Mostly this a problem that distros caused and are the ones that have to solve it. Really having a couple of implementations isn't a problem. The whole MP3 issue was brought on by patent paranoia, everything was fine before.

#1 Object Oriented Shell

[jackspenn]

This obviously requires some major supporting backend work, but an object oriented shell would be awesome. Microsoft's Powershell is possible because of the years they put into the .NET framework and even still their is some room to grow, but a true object oriented shell for Linux would be awesome.

- Eric

Re:Its been in development for almost 20 years

[Eli Gottlieb]

Somehow I think that Darwin and OpenDarwin serve better as viable microkernel operating systems, since people actually use them and everything.

3D games

[nikanth]

My favourite openSuSE rocks and almost has all the features and UI is cool... But I would like to invest heavy resource to develop cool 3D games that would make the platformers/console owners jealous

4 points

[garphik]

-> Better support for hardware, its better from where I have seen, but it could get even better

-> User interface, The current user interfaces heavily imitate existing ones. There should be more originality in the user interface, some innovative designs of doing simple things.
Also things like GNOME / KDE do not give a sense of a complete package, there is little something missing here and there. But we are getting there ...
Konqueror is a very good 'explorer', still there is no such thing as the simple 'Windows explorer'

-> Better package managers, even more abstracted installers as simple as double-click install

-> More Stability, I think thats the only reason why people opt for Unix

All said and done, kudos to all the linux devs for putting Linux as a viable alternative in the OS soup.

Learn from history....Visicalc

[Anonymous Coward]

Visicalc was the program that made the PC worthwhile. If you make an accounting program better then quicken and peachtree for linux for FREE, you will get the business types you need, to convert to linux. GNU money ain't cutting it right now. Make it happen so those people can leave microsoft behind.

CD DVD Burning tools

[Anonymous Coward]

I would like the CD/DVD burning tools to be improved upon. Right now there are many options for Windows and most of them work very well which hasnt been the case in my experience with Linux.

Better hardware support..

[Gnodab]

Better laptop support. I've tried three different distributions on my new Pavillion Laptop, and not one has worked without massive tweaking and command line editing...not exactly a great incentive to switch.

Better wireless support...

actually you know what? Better hardware support all around. Make it as easy as Windows to set up. also a universal install file (like .exe on windows)

supplemental hard drives!

[gregconquest]

The biggest problem I've had with various linux installs involves extra internal hard drives. After having my own troubles and starting to fix my own problem, I decided to write a tutorial on how set up fstab to help others:
http://www.linuxquestions.org/questions/showthread .php?t=547302
After several iterations, I was finally getting close to the solution, and then, BAM! fstab does not set all the permissions for extra drives! And even fstab is not documented fully. The project languishes now . . .
Users need to be able to install linux and it ask them how it should deal with extra hard drives. This info needs to be written to an annotated fstab and the info should also set permissions -- with all of this reviewable and configurable via a GUI utility later.
And even then, when one of my "extra" hard drives is mounted incorrectly or has an error, linux stops the bot and I *have to* deal with it at the command prompt. it is tooooo much!
Greg Conquest

Re:Three things. BETTER GUI FOR SECURITY... apk

[Anonymous Coward]

Jesus christ, are you still flogging that shit? Nobody gives a fuck. A shitty network tool from a shitty website endorsed by two shitty tech magazines that nobody reads (out of the hundreds of such magazines nobody reads) is not the Word of CompSci God.
  Take some fucking zoloft for the obsessive compulsive disorder and stop posting. Nobody wants to hear your bullshit about how awesome you are and how everybody else sucks.

    - Sincerely go-fuck-yourself,
            Anonymous of Internet

Cinelerra-CV is trying, needs resources

[Velmont]

First thing I thought about seeing that headline was Cinelerra. The CV-team is trying the impossible; starting Cinelerra3. Everybody is skeptic because it's an enormous project; but with time and developers it could go.

I really want it to succeed; (I'm using Final Cut now, have to use a mac) so that would be one project in need of resources. :-)

How's about a poll?

[eclectist]

This seems like a good question for a poll. If you really want to know how ./ users would refocus linux development, give us 10-15 options, and let us vote.

DRIVERS

[Anonymous Coward]

Video drivers
Web cams detection & drivers
Ontegrated mainboards drivers
Linux certification for products, as mainboards, with penguin logo
GRUB: allow to boot from an iso image in the hdd or a memory stick
Compiz Fusion multilanguaje tutorial, with video for beginners
Portage, wget and similars, automaticaly manage errors, cron events for upgrading system daily and world (or all) weekly as easy as windows update for teh user.
Opengl hardware requirements for next generations video cards, Direct X is now the standard and what marks features, opengl games and apps should be the leader.
Cad-cam good app
Acounting-billing multilanguage, multilaws, it would be a enviroment, with plugins for each law for each country and some exporting tools with acounts correspondence. It would alow to format bils as a spreadsheet, and manage bills as a database, with modules for each kind of business, and where you would be a ble to add a product or service from the same billl, typing, and putting it in the database od products/services avalaible to bill.
Repairing, manteinance tools, as registry mechanic for windows.
Bios updates tools (they need DOS or windows usually) at least a DOS way of executing them

Automatic execution with wine or DOS (dosbox or freedos) executables, some way of making exe, com bat and msi files to execute with a double click using DOS and windows emulation automaticlly.

Standarization of Klik or any easy way of instaling new single applications, not to have to write several commands on console, scripts and other packaging can be standarized for a double click installation of apps, and them would change paths, enviroment, and install without that tools.

Incremental update upgrading, a patch system, where updating only changes the cahnged parts, not a full installation of an app each time you upgrade it.

A good compiler for play station 3, wich can use all the cell processors it has.

Ease of Use

[cjb110]

...it is the biggest weakness of Linux compared to Windows or OSX.

Obviously this isn't the easiest to 'fix', as a lot of it is subjective but even so the competitors have done it, so Linux has to as well.

Some of this is simple things:
CLI tools should have simple memorable names that relate to their function and not some obscure geek humor. Then their parameters should be standardized, e.g. is it /?, -? or --? to get a brief help? Pick one, then remove all the other variations from every single CLI tool that gets distributed with the main distros.

CLI tools should be marginalized, there should not be a CLI tool without a fully functioning GUI equivalent, and in fact the GUI should be more powerful. You can add features to GUI's and keep them usable, but with CLI you end up with all the entire ascii set as valid parameter choices!

Remove duplication, there should only be 1 'architecture' for things like networking, printing, sound etc...
This applies to the tools too, if there are two tools that do roughly the same, then either merge them or pick the best.

Hardware, as forcing the manufactures for drivers isn't 'achievable', you need a complete HCL (that's targeted at the basics, i.e. the default view only lists hardware you can go and buy today).

A single base desktop would be great, as it would focus attention onto usability and documentation. Obviously it can then be ported/copied into the various desktop flavors.

I think a lot of what I'm trying to say is pointing to this:
Linux needs a drive to a state where there is one way of doing 90% what the user would ever want to do.

and one the best ways to do that is to remove choice, sometimes remove it completely, but most of the time move it to a config file.

It should make SUGGESTIONS...

[EEBaum]

Linux and Windows are on the complete opposite ends of the Suggestion spectrum. Windows is a mother-in-law constantly yelling "do it that way! i can get it cheaper at Sam's. Someone's at the door." over your shoulder. Linux is 34 million unassembled apparati in one big pile and 34 million instruction manuals in another pile.

I know Linux is all about freedom and choice, but would it really go so much against its principles to make suggestions? Suggestions such as, "There are 38 FTP clients available on your system. Why not try this one?" or "There are 8 ways to install software. We recommend using yum."

The same story often happens... user wants to do something. user looks around. user finds program that supposedly does what he wants. user looks at manual. manual says what things do. user tries to use program. user becomes frustrated because program does not do what he wanted or expected. user is frustrated. user complains online. user is told to RTFM. user says FTFM

See, the user already tried reading TFM. Problem is, the settings out of the box don't do what he wants on the program, and the instructions, while technically correct, don't make any useful recommendations for what settings are commonly used and how someone might commonly tweak them. Further, there's another piece of software out there that would do the job much better. But how would he know? The software just came in one big pile, and he tried the first one that had a name vaguely related to what he needs to do.

I'm not saying it's a good idea to have a "Press 'Start' to Start" arrow pointing at a button labeled "Start," but a step or two in that direction, across the Linux world, could make a HUGE difference in usability by people who don't have a full-time-job of time to dedicate to making Linux work properly.

my list of things

[Anonymous Coward]

-A coherrent desktop environment with focus on doing a few things very well instead of everything under the sun but half-arsed. This does in fact go for every app under linux with a graphic UI.
-Better multimedia creation support.
-Less dependancies through out the system. I'm sure some functions could be moved to common libs so you don't need to install app x y z and so on to get the app you want running. Or make apps have all they need in their own folder, think mac os does that.
-One common package format. It is insane that as soon and app gets patched there has to be one package per dist release and platform. It is redundant beyond belief.
-Better support for games.

CAD

[msim007]

Personally I would like to see a decent CAD application (or even better .dwg compatible BIM app like Revit). If you use CAD software you are more or less stuck with windows & autodesk. It would be fantastic to get an open office type app for CAD work.

Re:The Focus on documentation

[Anonymous Coward]

Videos:
Utube has a few excellent videos that are so well done, a nervous user may just try it. However, many so far I see are just not there yet. Too often I just see a series of 'then type this.. oops forgot to type this' repeated again and again...which just confuses the busy, tired from work, nervous newbie who is coming from visual basic and needs to see some encouraging results fast. Make a video with just a little planning so you get it right the first time with a decent clear voice over explaining what you are doing. Do a series of videos on a few basic topics.

Then you will start to encourage all those timid ones who are starting out on Linux and are definitely not code warriors. Then they can feel brave enough to actually try it.

Documentation:
Get a documentation wiki (perhaps a developer wiki too). Then ask for a lurker of the group to document in the documentation wiki what is written by the devs in their mailing lists, when they help a user. The wiki docs become the basis for updating the docs. One wiki I found is two years old. The guy posted that it is now more than 90MB after spending about 20min every two days. It was recently used for updating the project docs. Looking at the recent changes it is All done by ONE guy.

You only need ONE volunteer. One guy said He notes the info down and emails the users who asked on the list if they need to update it with anything further. Another project , the guy occasionally posts a link to the mailing list after an interesting thread and the devs can check it if they wish. I am going to go and perhaps suggest something quite radical in that even If a few very vocal ones in the project hate the idea - try and do it anyway (you can always scrap it if it fails). why? I have seen one wiki done which was rejected a very vocal few. Two guys got together and did it anyway. After about a year, people now are quoting it and then those who just hated the idea are surprised because the wiki is more up to date than their docs. This motivated the docs team to get a hurry on, and the docs now update much faster. I just saw one of those opposed, quote from it and recommend it to a new user. . As long as the wiki dude doesnt make waves and is seen as a constant supporter - it may eventually gain acceptance. But the official docs will quickly start geting more updates.

Reward? Well For the documentation wiki, I recently asked one and he said "well, a free airplane ticket to a conference every now and then would be very nice, but I just enjoy supporting the project". Try asking among the lurkers. With a bit of encouragement someone may step up.

Outlook Calendaring Replacement

[egarland]

Linux needs to be able to dislodge Microsoft Office in order to become a viable option for large-scale replacement of Windows in the corporate environment. Open Office can handle most of the document creation work and Firefox/Thunderbird do a great job at handling web browsing and email.

What is needed now is something that can integrate with an Exchange server's calendaring and also integrate with a robust open-source calendering server system to replace Outlook's calendaring functionality with an open, standards based system. We need to embrace and extend Outlook.

Much ado about nothing

[Anonymous Coward]

Well, a lot of the issues discussed here have been resolved in a distribution called Freespire

Bring KDE and Gnome together

[pxpt]

Now that the two main GUI camps have had time to play and find out what works and what does not I think that it is time for KDE and Gnome unification. I know that this would be almost impossible politically but think of the advantages. For one, you would have a unified look and feel for your applications and everyone would target a single environment for development. I want to get around to writing applications for Linux but this front end issue is so pathetic that I am holding off until there is a single target to write for. I know what some of you are going to say - just pick either KDE or Gnome to write for and ignore the other. NO, I refuse to do so on principle - there should be one GUI! I wonder how many people feel the same way?

Both the KDE and Gnome camps should get together and work out what should be specified in the new GUI. After doing this they should then construct the GUI and release it as the new way of doing things. Personally I think that the way to go would be to leverage the CRL (yes I know its a microsoft thing but the Mono project is coming in leaps and bounds in this area). Even if this is impossible to organise at least let them thrash out a common look for their two environments in the interim.

The lack of a uniform look and feel is one of the greatest problems we face in general acceptance of the Linux desktop (along with easy installation of course) and so I think that throwing some money in this direction would help things enormously.

Oh, and don't get me started on the X Windows side of things...

A few ideas

[Anonymous Coward]

I don't know what of this was mentioned already - too much for me to read all of it - but here's what I'd do differently:

- I'd like to be able to sort the stuff I install by license. A completely free version which I can download for free, put my name in, make a change or two (if I like) and sell as my own without any GPL limit would for instance be good. Might also be handy for games that run on every pc, recovery tools, and similar applicatons where a standard base might be practical, but open source, gpl-relicensing, or any such thing might not be desired. The GPL is simply to education-oriented and business-unfriendly to be useful for every commercial application, and there appears to be enough completely free stuff available - there are some distributions that are completely free afaik, but they are usually too specialised. GPL 2 and 3 versions would also be neat, including free software, but excluding proprietary stuff. And then a more or less "commercial" version including proprietary drivers, non-open source parts, and the likes - probably the best for practical mass market desktop versions. All that from one vendor/distributor, at according rates (no charge for the unsupported free and gpl versions (should be possible to finance the effort with advertising on the download page, and some governments, universities, or companies might also be willing to contribute), any amount for the version including proprietary stuff (usually drivers), and certainly different kinds of support for according prices - regular updates for a low price or clicking on some advertisements, user support for usual fees for that kind of stuff, and even premium priced developer support for people who want to have help developing things (even if they don't want to put it into the gpl). That might also help finance the free versions, as the entry into support.

-I'd like to be able to create really specialised versions of Linux from one installer. The installers I got to know so far all basically do 1 to 3 different specialisations (desktop, workstation/programmer, and server) and the rest of the hard to understand options just add/remove the one or other program or redundancy. I need completely different installers and even Linux versions for embedded systems, all with their quirks which I don't know before having played with it for a week. I'd love to be able to create anything from an embedded system not able to do more than show the time and a stopwatch (think microwave oven) to a fully fledged power user system which also acts as a server for the home network. All that from one installer (the same which does the licensing issue above). All that without typing in anything but my username and password (if it's not an embedded system without authentication).

-I'd like to be more able to choose exactly the software I need - I don't have anything from installing 50 different text editors which all have advantages and disadvantages. All I need is a simple text editor, a good program editor which automatically formats all relevant programming languages (an extendable library with formatting might help), a simple wysiwyg text editor for a quick letter or reading readmes or the likes (preferably embedable in my email program), a simple html-editor, a fully fledged (office) text editor, and maybe a dtp program and a fully fledged html design program. All of them integrated with other programs - to insert graphics and other objects in the better text editors, or to insert their output into other documents. Similar is true for all other areas - and probably different for each person.

I realize that all this would be a lot of work - I suppose it's only possible if the work of the programmers is organised better - like wikis were people can change the source code of a given project, create builds for testing, add project build numbers to each version of the module, check for which of the build numbers which tests have been performed, a similar system to verify the compatibility with different versions of Linux, other catego

Not just visual coherency

[Anonymous Coward]

Application coherency. I hope to live and see the day where I can right-click the same folder on Nautilus, tree and right pane, seeing the same options. Printing is also a terrible issue GUI-wise.

yes

[Anonymous Coward]

Long ago, it was fun to be the cool kid with an OS that you honed into something unique.

Now I need stability, ease of use, software packages that don't suck. Network managment tools that allow top down enterprise support.

The community needs to make each application as useful as Gimp as compared to it's competition.

docking station

[hswerdfe]

I want to be able to connect my laptop to a frigging docking station.
I want to use my desktop keyboard mouse, printer, E-net, monitor, etc...

oh and I want an end to having to hack around with NDISWrapper, native support please...:)

Re:filesystem, init, xwindows, standard GUI conf t

[hswerdfe]

This might help...
http://www.gobolinux.org/
at least with your first issue.

3 things: Security, Security and uh,let me think..

[BDJones]

...Oh yeah: SECURITY!!! Make it the most secure platform available and over time it will become the dominant Business platform. Start by making the kernel impervious to root kits. I know, nothing is ever totally secure. But that is precisely the complacency we need to overcome. Why not totally secure? Not as an absolute reality, but as a goal we aggressively pursue. If the white hat community and the open source communities combined their efforts and pursued a totally secure kernel with the energy that the blackhat community pursued hacking the kernel, one suspects we could at least stay a step ahead of them. The problem is we tend to be reactive...always defending instead of preventing.

Home automation

[Anonymous Coward]

Linux stenghts could be used to link smart terminals in a smart home. could be used to create a completley secure virtual network between handhelds terminals cars almost anything. i dont know just thought.

Many suggestions

[derHerrLordKanzler]

My paroles to the FOSS developers:

1.) Get all those FOSS developers, who develop a "normal user program" like let's say for KDE or GNOME, to understand that they have to have to write a program with the needs of average users in mind. That means, every non geek or non authistic human has to use the software without any fear but with ease.

1a) Drop the RTFM attitude and learn how to design UI's that need no documentation. If they want non autistics to use Linux they have to accept that these "normal users" set the rules of the user interaction. Discover, that creating software that is usable without pain for everyone is hard, it needs smart developers, it needs geeks.

1b) Drop the "a foolproof system will only be used by fools" attitude. I know a bunch of VERY smart people who simply turn their back on linux. A comparison that might help is the following: I know nothing about electronics, circuits etc. When I want to play a Audio CD with my HiFi, I absolute don't care which microelectronic is working in the background, I have a play button, a track selection button an that's all I need. I don't care a shit about what's happening behind. Normal users don't care about the point-to-point-protocol and they shouldn't.

2.) Get away of the concept that more features will improve a software. Like explained in 1.) this is the geek, power user perspective, not the standard user perspective. The standard user only uses a very small fraction of let's say kmail. He/She doesn't need the rest, and hence he/she shouldn't see any disturbing buttons, menus, popup's, etc. Cut down the functionality to basic features and make those main functionality easy and obvious to use.

3.) Develop software that doesn't ask useless questions or shows useless information to standard user: "application Foo couldn't open directory because it couldn't access MIME Service Bar". It's of zero use to a normal user, never ever display such thing. Either the software can fix the situation itself or it should simply say that it's not working.

4.) When configuration is necessary ( for example access to the net ) only ask the user questions that he/she is capable to answer: "Should the ppoep module be loaded into kernel?" is a question that 100% of normal user won't be able to answer. "Enter the username and password for your internet connection" is something that the majority will be able to answer.

4a.) Ask the user a minimum of questions: When it's technically possible that an installation routine can find out that it is installing linux on let's say a Dell Inspiron 8000, then the routine should lookup a config list, find the Dell, crosscheck with the EEPROM's and automatically setup for this laptop.

5.) Stop stupid things like in Fedora 6: You can install xmms or some other player, but there won't be any sound of an mp3 file because the MP3 decoding module won't be installed. It's not even included in the distro for licensing reasons. SO WHAT'S THE POINT OF INSTALLING AN MP3 SOFTWARE?

6.) Stop acting like you're alone on this planet, creating the 7th audio layer for Linux won't help anyone. I installed grip on the Fedora 6 box of my father. I can rip, encode, but I can't test play any track to check what I wanna rip, grip plays but no sound comes out. THIS IS RIDICULOUS!!!

7.) Remove any trace of ASCII, delete it out of the history books, and hang the enventor(s) in some dark cellar. Remove the char type out of the GCC parser! Unicode is the most wide spread real alternative and should be used throughout. I don't ever wanna see some "?$" letters instead of a ä or Ü.

8.) Stop that stupid "I have to configure the keyboard for the console and a 2nd time for X".

9.) Remove X. X had many years time and didn't get anywhere on the desktop. The 17.000 extension to the protocol won't change that. NO DESKTOP USER NEEDS NETWORK TRANSPARACY. In the graphic sector create something like directX. Write perfect, super fast drivers for all video cards, support ALL hardware features.

Less betas, more finished products

[huckamania]

Linux desperately needs some sort of review process for projects, products, etc that is for the end users, not the developers.

There are way too many beta projects that are either stalled out or superfluous. Way too many application choices with almost no information to know whether the available apps are any good. I'd love to see something like downloads.com that concentrates on Linux, with user reviews and ratings, that isn't just fanbois or hopelessly out of date.

It happens so often on Slashdot where someone says, "gee, I wish there was something like this.", followed by replies that say "duh, ur stupid, use X.", "ura troll, Y has been out in beta for 6 months.". You can search for projects, and SF is fine for what it does, but it doesn't attempt to review or rate anything.

I'm not in favor of some licensing body, I just want to be able to see the choices and read a review.

OSS does not suck here

[Anonymous Coward]

http://www.openbsd.org/cgi-bin/man.cgi
http://openbsd.org/faq/index.html
http://www.freebsd.org/doc/en_US.ISO8859-1/books/h andbook/

Run, Forrest: RUN!!! apk

[Anonymous Coward]

"Jesus christ, are you still flogging that shit?" - by Anonymous Coward on Monday August 27, @02:42AM (#20368715)

LOL, yes, I am... & note one thing - I only do, when I see the usual F.U.D. b.s. from the crowd here that is usually along the lines of this:

"(Insert *NIX variant here) is more secure/securable than Windows"

To which I make this challenge, and to date (documented in this thread's exchange no less) nearly 40++ *NIX users run from it... rotflmao!

----

"Nobody gives a fuck." - by Anonymous Coward on Monday August 27, @02:42AM (#20368715)

Well, I really dislike stating the obvious, but... Apparently, on "giving a shit"? YOU do... I mean, why else would you reply so "ragingly", lol???

The funniest part is, I did NOT "throw this test into anyone's face" in THIS post, I only used it as an example of where folks who use *NIX said tools they have available are hard to use (per SanityInAnarchy telling me so as far as UBUNTU SeLinux).

(Frustrated on your end, @ your crappy score on this test giving you a 'wake-up call', dolt?)

Did you try it & saw how CRAPPY your security setup rates on your *NIX rig, vs. mine scoring in @ 85.185 currently????

I bet that is the case... & thus, your LACK of a score that exceeds mine, done on Windows Server 2003 SP #2 fully hotfix patched, vs. what you can get using a *NIX of YOUR choice.

(I guess it's easy to 'talk big', but when the chips are on the table, you all fold! Your lack of results, especially ones superior to the ones I showed tell everyone reading here that VERY thing...)

----

"A shitty network tool" - by Anonymous Coward on Monday August 27, @02:42AM (#20368715)

First of all, here's a fact for you, to improve your accuracy: It's not a "networking tool", it is a security test (that helps you to secure yourself no less on the OS of your choice it runs on, which is many).

----

"from a shitty website endorsed by two shitty tech magazines that nobody reads (out of the hundreds of such magazines nobody reads) is not the Word of CompSci God." - by Anonymous Coward on Monday August 27, @02:42AM (#20368715)

Secondly - 2 magazines, that despite their being "shitty" (per YOUR opinion only, big deal), often are cited here... and, only 1 is a magazine, dolt. SANS is a website, COMPUTERWORLD is a mag!

Tell you what - You find a BETTER & MORE COMPREHENSIVE TEST, that helps you secure your system no less (and is multiplatform) & I'll take that too, & yes, beat your score.

OK?

----

"Take some fucking zoloft for the obsessive compulsive disorder and stop posting." - by Anonymous Coward on Monday August 27, @02:42AM (#20368715)

Care to show me your PhD in psychology or psychiatry, since you see fit to dispense advice & meds?

AH, don't have those certifications?? I didn't THINK so... hell, I knew not. Too easy... big talk, no proof, as per usual (just like your lack of results on this test, lol).

Also, have you considered taking a dose of YOUR medication? I mean, look @ your "frothing + foaming @ the mouth 'eloquent' reply", lol... more like an "exercise in profanity".

----

"Nobody wants to hear your bullshit about how awesome you are and how everybody else sucks. - by Anonymous Coward on Monday August 27, @02:42AM (#20368715)

What b.s. is that? AND, where did I say "I am awesome" & "everyone else sucks"??

(Ah, you won't post proof of that either... because there isn't any for you to use here in this exchange, period! Typical b.s., from a defeated dolt, nothing I do not expect... very, VERY predictable on your part!)

After all: I merely put up valid evidence of a score from a test of security based on best practices that is noted by both SANS & COMPUTERWORLD...

(SANS = a respeced website regarding security

Re:Run, Forrest: RUN!!! apk

[Harik]

These clowns? Quality secure configuration there. That's called "information leak", and is one of the first things you're supposed to do when trying to secure a site. Hell, modern webservers have all that crap turned off by default.

Anyway, your posts are full of words and hyperbole, but I'll try your test if I can actually get it for linux. I'm not going to pay for it though, and it appears to be free for .gov use only?

Fix the sound

[Anonymous Coward]

Audio on Linux should be just as plug-n-play as video. It should be low-latency and high definition by default. The ongoing support for OSS and it's ilk is a mistake. Get rid of OSS. Get rid of ESD.

LSB Desktop is already a reality

[Burz]

http://www.linux-foundation.org/en/Specifications

The next version of LSB is supposed to define much more desktop functionality, and even an installation interface (hopefully this would include package names as well). I wouldn't say that LSB has to take everything (or even most things) from all major distros. The distros seem happy these days to adhere to LSB as long as it defines a core functionality that does not overreach or grow too fast.

But while distros are adhering to the LSB standard, application developers and users are oblivious to the standard. It is not being marketed to them, though it should be. Instead they keep talking of this imaginary "Linux" OS when they should be referring to "LSB".

The 6 mo. distro cycle is indeed destructive as you say. Even with Ubuntu which settles on an LTS version every couple years, the intermediary versions end up pulling the tech-savvy power users away from the non-techies still plodding away with LTS; this diminishes the former group's ability to assist the latter group. Either that, or the Joe Users end up constantly harrangued by the techies to keep upgrading their OS.

The hottest new doodads are very nice to have around... IF they are merely add-ons. In OS X I can replace my samba with the very latest version, but the fact that all new OS X 10.4 installs start off with the same basic version of samba is extremely important to platform stability. Eventually, Apple will incorporate the major new samba release into a future OS X upgrade.

The anarchy you speak of is a grave concern. The trick will be to convince a large enough group of people that LSB-compliant stuff is the defacto starting point... the thing you reach for and work on before you try anything else. We will also need to be tolerant of elite techies who want to do things their own way, as long as they do not advertise their wares as LSB. Think of it as 'standards etiquette': Everyone has freedom to do whatever except where it comes to false claims. People just need to be made aware of LSB and learn to ask for it in their OS and apps (i.e. when writing letters to Adobe) in order for that dynamic to take effect.

Fix this at a higher level........

[tempest69]

You can have dependency problems is you just use RPM's however there are tools such as "app-get" and "yum" that reduce this issue to virtually zero although you do have to be careful of the repos you pick since they have to work with each other. Were people run into issues is having to many enabled repos. To have all repos enabled and accepting all is just plain silly since you are going to have issues. What I do is to activate one repo at a time in a preferred order with the primary repo first. I never do an installation or an update without checking what I am getting and I will only answer "yes" when I am satisfied.

I agree that yum is great... for the dependency issues that it is dealing with it really does a fantastic job..

That being said yum is fixing issues that need to be fixed at at a much higher level... Most of the applications need to be ditching the dependencies.... The pieces and parts that are needed to get a program running are a nightmare..

If you need software that isn't covered in your repo , the installs get really ugly really fast, after a couple dozen oddball packages try and fight out a dependency war, you finally cave in and remove some programs, installing them on another box, (or a chroot jail).

Anyway, applications need to be well defined, where they aren't installing jack in the rest of the system.. Where the libraries they need, don't mingle with the rest of the system libraries. Where all the dependencies can be removed without harming any other app on the system..

What I'm proposing has space issues, no doubt, and some memory overhead. If we are to go forward in a meaningful way, we need to smash the complexity of the multi tentacled applications, down to the system calls and the display calls.. where the removal of the application is as simple as rm -rf ./apps/myCruftyInstall and then removing the shortcut from the desktop.

Storm

The clue's in the question

[gidds]

what application or kernel area would you attempt to improve[...]?

That's the problem, right there. As long as developers continue to view each application, each kernel or kernel area, each driver, each service, as its own little kingdom, then Linux will fall short of its potential. I believe the real gains will only occur when people consider the system as a whole.

Disclaimer time: I've never run Linux myself; what I know is from a few glimpses of friends' desktops, and reading an awful lot on this site and many others. But I hope I'm not entirely ignorant about it.

Many years ago, I ran an Atari STE, and later an Atari Falcon. It became a surprisingly powerful system: pre-emptive multitasking, standard web browsers and email clients, a task bar and start button, Unix shells, gcc, perl, etc. It did most things I wanted -- many of them very well indeed. However (which is why I bring up all this history) that came at a price: discovering, installing, configuring, and updating all the many apps and utilities which were needed, and getting them all working properly together, took a lot of my time and effort. Time not spent actually using it to do things!

When I switched to (the newly-released) Mac OS X, it was as if a great weight was lifted from my shoulders. Here was another powerful system that worked really well, but which didn't need an input of time and effort to make it so! The Mac's motto 'It Just Works' is pretty close to the truth: I spend very little time searching for apps, utilities, or drivers, persuading bits of the system to work properly, upgrading, recovering after crashes, or any of the other ways I wasted time on my Ataris or still do on my work PC. And that's not due to any single feature, application, or utility: it's in the way everything works together. The whole system has been designed to get out of my way and let me do my work with as little interruption as possible, right from the moment I first turned it on. And from all I've heard about Linux, it sounds like that's where it could benefit most. Is that a valid impression?

My bitch list

[MeditationSensation]

1. BitTorrent that works as well as Windows. I had a devil of a time messing with port forwarding and all that and still couldn't get more than a few KB/s speed. UPnP on Windows makes this a snap.

2. DVD viewing program with deinterlacing that's as good as PowerDVD's, for watching TV shows without all those horiztonal scan lines during motion.

3. Desktop Effects in Ubuntu locked up my dad's computer when I turned it on. He has a modern Dell system too. Pressing the shutdown button in Ubuntu locked up X for me one time. Just general polish stuff like that would be nice.

You can't play Halo on a GameCube

[tepples]

Seamless Inteoperability Open and share documents Please be more specific: What documents? Share in what way? I want to read and write docx and doc file properly. What problems have you had with the .doc filter in OpenOffice.org Writer?

Avialability Of Core Apps Photoshop (sorry, gimp doens't cut it yet) Please be more specific: [Why not?] See here Your search returned 2,160,000 documents. The "I'm Feeling Lucky" result was vague about what specific features GIMPshop is lacking. Which feature listed in the other 2,159,999 documents are you referring to? Or are you asking me to read each documents 2-10 in that list, collect every feature mentioned in each of those documents, and do some research to determine whether it is a feature in GIMPshop that is just done differently, a feature in Photoshop Elements that is not in GIMPshop, or a feature in Photoshop that Adobe left out of Photoshop Elements? If you are a professional who needs the full feature set of the full version of Photoshop, it might be worth buying a Macintosh computer to use as a Photoshop appliance.

Games Please be more specific: Doesn't a typical distribution of GNU/Linux come with several games? You're joking, right? (Sadly I think you're not)

Though you can't play Super Mario on a PlayStation, games for one platform do have counterparts on another. For example, if you like GameCube-exclusive platformers but you have a PS2, there exist substitutes: Crash, Jak, Ratchet and Clank, Sly Cooper, etc. Likewise, if you like Zelda, PS2 has two Dark Cloud games. Which genre for Windows did you find lacking on Linux?

But if you want a specific title for Xbox 360, buy an Xbox 360 console. Likewise, if you want a specific title for Windows, buy or build a Windows console. Even PCs running Windows 9x, Windows XP, and Windows Vista are almost as different as the PlayStation, PlayStation 2, and PlayStation 3 consoles.

What do the *users* want (hint - its not to be reading through USB driver help sites wondering why the printer doesn't work). I know what they don't want: Windows Genuine Advantage false alarms.

Finally, HARIK has some balls (others didn't)...

[Anonymous Coward]

"but I'll try your test if I can actually get it for linux." - by Harik (4023) on Monday August 27, @11:24PM (#20379941) FINALLY! A "penguin" with some balls... excellent!

Well, take a peek over there, because there are builds for it, for Linux... what kind of LINUX distro do you use?

I'll need a photo proof of your score. You can post it here, or email me it, & I have my old 84.735 score up online, but my 85.185 most recent high score I do not, but I can supply it with ease by email if needed, until I get it posted up online publicly.

----

"I'm not going to pay for it though, and it appears to be free for .gov use only?" - by Harik (4023) on Monday August 27, @11:24PM (#20379941) I didn't pay a cent for it, it was free!

----

"These clowns? Quality secure configuration there. That's called "information leak", and is one of the first things you're supposed to do when trying to secure a site. Hell, modern webservers have all that crap turned off by default." - by Harik (4023) on Monday August 27, @11:24PM (#20379941) I saw the index files & some backups of them, this is all (I am not a 'webmaster/web developer' by specialty here, so I really don't see your point on that being "bad", but I suppose you have some reason for stating that, be good to hear what it is, specifically, as to "information leak")...

Above all - & if you think that's some security violation? Let them in on it... it's the RIGHT THING TO DO, don't you agree?? I feel that way @ least, & I hope you do as well!

----

"Anyway, your posts are full of words and hyperbole" - by Harik (4023) on Monday August 27, @11:24PM (#20379941) Aren't most (as far as words)?

Plus, admittedly, my posts get "big"... mainly because I tend to quote others, as I have your points/replies, so that others reading can see the points I am responding to, & yes, it "bloats my posts" doing that, but hey - @ least it helps me not miss a point others are addressing of my own points.

APK

P.S.=> Good luck Harik, my current score is 85.185/100, & I am looking forward to seeing a score from someone with a LINUX OS based rig... but, I would especially like to see one from an SeLinux bearing distro (such as UBUNTU/KUBUNTU) or BSD variant (FreeBSD runs this test from what I understand, no hassles)... & thanks for @ least having the courage to give this a try on your end. That alone sets you above the rest here, even IF you can't beat my score noted above... apk

85.185/100 on CIS TOOL, photo of mark to exceed

[Anonymous Coward]

APK CIS TOOL SCORE 85.185/100 on Windows Server 2003 SP #2 fully hotfix patched (date 08/18/2007):

http://img.techpowerup.org/070828/APK_AToutLeMonde _85.185CISToolScorePhotoProof.jpg

There you go Harik: "the score to beat"...

Above all else - Good luck!

(& I look forward to discussing the points you find on your LINUX rig (I do strongly suggest using a distro like UBUNTU/KUBUNTU though, or a BSD variant (since it is always touted as the "most secure PC *NIX" & all that))).

I say that, on using SeLinux bearing distros, mainly because SeLinux kernel hook addons for Linux comes "baked in" from the start on UBUNTU/KUBUNTU from what I understand!

So, I think it is a GOOD idea to learn it for a LINUX person, if you do not know how to use it already, & how to strengthen it beyond default policies!

All, for "layered security"!

E.G.-> By additionally supplementing IPTables use (or even NetConfig dual homed/dual nic NAT firewalled setups, nice tool in Linux by the way I have always felt this one) via SeLinux's SOCKETS LEVEL control + its MAC control of things file/folder/disk vs. using only chown/chmod/chroot for filesystems/disk level userrights control for layered security.

I.E.-> We can discuss points on its scoring you agree with, & others you do not (because I have 2-4 I am fairly certain I pass muster on, yet was 'downgraded' on... we can go into specifics on yours too in that capacity too)...

APK

P.S.=> I look forward to it, & I hope you do too as well... we can ALL learn by it is why, & for better security (especially today, in this online world of virus/spyware/malware/trojans/worms etc)... apk

Re:85.185/100 on CIS TOOL, photo of mark to exceed

[Anonymous Coward]

You know, when you ask for "photo proof", the least you could do is supply that yourself. I assume that the reason you're asking for a photo rather than a screenshot is just in case people try to Photoshop it.

Of course, even photo proof doesn't necessarily mean a thing. That sort of thing is terribly easy to fake.

(I'm not saying you *did* fake it, you understand. I believe you. But I'm just bringing up some points you may want to consider.)

Re:Finally, HARIK has some balls (others didn't)..

[Wile E Pyote]

Hmmm, I'm very interested in seeing the results of this. Does this run on OS X as well? Oh, and hi APK. lol

Re:Finally, HARIK has some balls (others didn't)..

[Anonymous Coward]

Heh, Hello Wile E... & I too, am interested in his results!

And, not so much for the score, but for what he is scored down on, via his screenshot, & I am into discussing those portions especially, & even my own I was scored down on, with GOOD reasons!

(As I mention in my posts to he in this exchange, I have 2-4 areas the test 'scores me down on' that I am utterly convinced my system is 110% straight & solid on... I think HE may also find that type of thing on the LINUX version of CIS TOOL also).

Anyhow: Thanks for posting my score photo in the thread that shows Windows users how to get the score I did, or possibly, even better ones!

APK

P.S.=> No MacOS X model, only a .pdf guide so far (some Linux's go thru this too, as do some BSD variants (e.g.-> FreeBSD runs it, OpenBSD + MacOS X do not)).

On LINUX: I suspect though L.T. controls the kernel, some distros arrange file placements differently (like under the /etc tree, for files that keep state + security info. etc.), & thus, the test needs alterations for those versions/distros that do NOT run it from the LINUX family...

I mean, it HAS to be something along those lines, because most ALL Linux's are close in nature/the same especially @ the kernel/core level due to L.T. controlling it there, & BSD variants are all from the same base "codetree" for the most part as well, just like LINUX! If I am wrong here? We are going to find out about that, as well!

- in other words, I find it sort of odd that (especially since this test is java driven runtime interpreted code)... apk

Re:3D - buy Nvidia

[Anonymous Coward]

Thats why I use NVIDIA graphics card and Hp printer (Even if the drivers aren't Open, I will trust them on my PC not like some other

anti-trust company code --hint-hint) !

Three things

[pinky0x51]

1. Better hardware support especially for wlan and 3D graphic and with "better support" i don't mean proprietary drivers but free drivers or at least specs.

2. Free Java (IceTea) has to become ready to use and default on every GNU/Linux distribution

3. Free Flash (gnash, swfdec) have to improve and become default in all GNU/Linux distribution or even better part of the browsers.

Future planning

[gilesjuk]

Linus has admitted he doesn't look ahead far, this harms the quality of the software as future planning can reduce the need for rewrites and restricted implementations.

While Linus and others don't like microkernels, there needs to be a better way of adding additional drivers without recompiling.

Overlooked items

[Anonymous Coward]

Why is it very few people in the Linux community ever think about the businesses, artists and home users that need multimedia/graphics/audio/video software?

Sure, for audio and video, there's Ubuntu Studio, Studio 64 and Dreamlinux MMGL, and while all are much better than other Linux distros in terms of making video and audio creation and editing easier, so far nothing stands up to what you can do on a Mac or Windows.
Programs like CuBase, ProTools, Acid, etc. are amazing and have no real equivalent in Linux. (Yes, I have used everything you are thinking of naming that's available for Linux.) Audacity is probably the only one that has an easy user interface (even if ugly) and is reliable, the others crash quite a bit or are just plain horrific interfaces, but all are limited in terms of what they can do.

And when it comes to graphics, I'm sorry, the GIMP is nice, but it certainly isn't Photoshop. When polled, over and over again, most Windows users say they need Photoshop in order to make the switch to Linux. The only company making anything even close to comparable for Linux is Pixel, which is very cheap and works very well, but it's not open source.

Think about it: Record labels, recording studios, graphics designers, artists, bands and solo artists, advertising firms, publishers, mom and pop who want to edit home movies, students who want to make multimedia presentations etc. etc. etc. all could benefit from a Linux distro that has software that does these things. I think converts would come to Linux in massive numbers if this happened. To give credit, there are several individual programs for Linux that can each do one or two certain tasks nicely, but for Windows and OS X there are single programs that can do hundreds of tasks all in one program.

Every single time I read about what current Linux users think would make Linux better, or how to get more people to use it, I always see a bunch of programmers and security junkies talking about improved security, stability, hardware support...I mean c'mon, DUH, everyone wants those things, but they aren't going to lure the average user away from Windows or a Mac, they help, they've certainly lured many, but they aren't going to make the big catch.

Open Office is great, Firefox is wonderful, Evolution works quite nicely and KMail isn't bad either, but when it comes down to it, what are the two biggest things people go nuts over when it comes to computers and the internet? MUSIC AND VIDEOS! Seriously, think about it.

You want to know what will get more people using Linux? Go look at all the products on Adobe's website. Then go visit the Steinberg (CuBase etc.) website, then look at ProTools. Of course, none of those companies seems to have even the slightest interest in developing versions for Linux, and many are hoping they will, but what are the odds? Slim to none. We need equivalents for Linux of our own, open source, and that are easy to use, have strong performance, many features and look good too.

Does anyone here think that Adobe could get away with charging as much as they do for their products if they weren't selling? Same with ProTools and Cubase.

Riva FLV Encoder and Player are free, work great, and even use open source software to do the encoding! Riva just put a nice and easy to use interface on an open source command line tool! Why can't the linux crowd do that???

If you took all the video programs available in Ubuntu Studio and slammed them altogether in to one program with a great looking easy to use interface, you'd be golden. Same goes for all the audio programs. But then there is always the issue with sound in Linux, it's confusing, sketchy, and a mess. That needs work too.

Sorry to rant, but this just frustrates me to no end. I know dozens of people and dozens of businesses that would switch to Linux if we had better multimedia and graphics apps. And I hate having to keep a Windows machine going just so I can do these things myself.

How about ratings?

[justthinkit]

A lot of the posts in this thread are of the "more of this" variety. More docs, more standard, more innovation, more of what I want.

How about simply adding one or more ratings to each distro, with each intended to add some incentive to those who want a higher particular rating. This is a way of encouraging and crediting the huge and diverse number of Linux distros, while also rewarding those aimed at the average/typical user.

For example:

• a "Hardware detection and installation" rating. The more hardware recognized, the higher the rating. The more hardware properly configured, the higher the rating. The more hardware optimally configured, the higher the rating. This could be a three-part rating.
• a rating for "most adherence to usability standards". Define the standards that should be met, and then rate each distro on this. This would be useful for corporations looking to adopt Linux, or for old farts like myself who don't want to relearn/tolerate non-standard interfaces.
• a "leanest distro" rating. This would need some good minimum conditions -- e.g. OS must be able to surf the web, or OS must come with email, browser, and office suite. Or it could be even more detailed, rating each application on its own scale and then giving a cumulative rating.
• a "most complete" distro. I would think something like Debian would score well here.
• a "most GUI oriented". An OS would lose points for things that could be only done by editing INI files, for example.
• a "more forward looking" rating. Alternatives to Gnome or KDE, for example, would score higher on this scale.
• a portability rating. How much of the core OS and how many of the included applications can be run on other Linux distros. I have no idea if this is at all valuable, but it is all about rewarding things that have some utility and thus encouraging more distro developers to do good things with their distros.
• a gamer rating. Reward distros that come with the most games pre-installed. Or those with the best combination of minimal OS on one image with best add-on games-only image.
• Rating security-focused distros would be valuable. Have different criteria -- leanest, more secure, most complete, most updated, most documented.
• a best-for-newbies rating would be a natural. Again, criteria would be key.

Summing up: Linux users have the most choice, and too much choice. Ratings help users choose what is best for them. "I want a newbie-friendly, lean, gamer distro" might merge the ratings from those three categories to give best overall scores and best scores in each category. User could then choose to rank by sub-rating, etc.

We've reached the "nothing new under the sun" computing age. It will be all about packaging from now on. In the commercial world that translates to "best marketing wins". In the FOSS world it translates to "best fit wins".

Who has time to read 400 (or whatever it is) distro descriptions? Can any one of those 400 paragraphs possibly do justice to that distro? Will we ever stop arguing about which one is better?

If we objectively and usefully rate distros, we can move on to choosing and using them.

"IS THERE NO ONE ELSE?!?" (2x)... Achilles

[Anonymous Coward]

"Is there no one else? Is there NO ONE ELSE!?!" Achilles, Son of Peleus, potrayed by Brad Pitt in the film, 'TROY'... Says it all, better than I ever could!

APK

Changing Linux

[WindShadow]

Other than a higher priority on not breaking existing systems with kernel upgrades, I'm pretty happy with the way it works now. People have an option to delay upgrades if new features are not needed, but new security issues need to be addressed, and removal of drivers and kernel options prevents booting the new kernel for testing without breaking booting of the old kernel. Not in every case, obviously.