Slashdot Mirror
Rasterman Says Desktop Linux is Dead
anguished writes "The future of Linux, its best hopes for blowing past everything else on an x86 machine, once was located in a little Austrailai website, with a window manager called Enlightenment, which we all hoped to be good enough to build and configure. In an interview with Linux and Main, the recently silent Rasterman talks about GNOME, KDE, E, and his view that the future of Linux requires new playing fields."
I posted this in another thread, but it got buried, so here goes...
For you and me, KDE and GNOME, along with any of the good standard distros makes GNU/Linux a great, pretty-easy-to-use choice.
But that's not good enough.
What I'd like to put together is Linux for Technophobes. The machine that Joe Schmoe, who has never used a computer, can walk in to Wal-mart, take home his new box, and be able to use it for email, web browsing, and word processing with zero assistance from anyone else.
He should open the box and find a simple (a la iMac) one-page sheet that shows him how to connect the mouse and keyboard.
A simple wizard sets up the net connection with him.
I'm picturing a very simple interface for the Basic mode. One big button that says Email and has a picture of a mailbox. Another for the web browser. Maybe a couple more apps, but not many.
And, if you click on the Advanced mode button in the corner, you get switched to KDE or GNOME.
Let me know what you think, and maybe we can put something like this together.
I am concerned about any program, any piece of hardware, any treaty, any law that treats me as a consumer, not a citizen
What, again?
:)
How many times has Linux died this year? I've lost count
"Backups are for wimps. Real men upload their data to an FTP site and have everyone else mirror it." -- Linus Torvalds
Of course the desktop is dead.
If we want a desktop that works,that will compete, there are two things that have to happen.
We need a single distribution. That's right. We need totally focused efforts.
We need a single desktop. No more of this "I can choose 10 window managers." I'm not saying take away the choice, but we need to pick one system and say "THIS IS IT" and the community can code for THAT.
Until we have focused, unified efforts towards bringing out a rock solid desktop, it won't happen. There is too much choice for the consumer.
However, I find his defeatist attitude annoying. I think the reason for it is simple: he seems to be a pure technologist, and therefore upon observing that the technically superior OS loses on the desktop, he gives up hope, embracing the idea that making the coolest, whiz-bangest WM for the ultra-31337 geeks is the best course of action (and while at it, take pot shots at the KDE and GNOME dudes).
What we need is more people who know how to market Linux to software companies so that the damned applications will get developed. This is not a technical problem, it's a business problem: there are too few desktop Linux users, thus a relatively small business imperative for software companies to incur the overhead of porting applications. Furthermore, the fear of free clones of your application and the culture of imitation in the Free Software world scare companies aware from producing commercial products for Linux (note that I think this fear is unfounded: a sufficiently complex, powerful application takes an awful lot of effort to clone. Your work should stand on its own quality).
The reality is that we need to find more ways to entice companies to develop commercial, closed source software for Linux if we want it to succeed on the desktop, for the masses. Don't say it's already there, we all know it's not. And we need to remember that the solutions to business problems are usually not found by technical means.
At the start of each new school year, Microsoft hits our campus hard. They hang big banners, set up booths in the student center, and get the managers to make the on-campus computer store employees wear Microsoft t-shirts.
The BYU Unix Users Group gives its own response. This year, we're going to have a booth in the student center too. We're inviting students to bring their machines, and a group of volunteers will install Linux on their machines on the spot, for free.
We're making up flyers that read, ``Thrusday and Friday only! Get a FREE COPY of OpenOffice Suite version 1.0 (must have student ID or employee ID). Save HUNDREDS of dollars on your computer software this year!''
We're not just going to be pushing Linux, but Free Software in general. For those who are queasy about jumping full-force into Linux, we will offer to install Mozilla and OpenOffice on their Windows partitions, so they have some familiar ground to refer to when they boot into Linux.
The biggest debate in the group at the moment is which distributions to recommend to the newbies who bring their computers to the booth. I argue that since we're installing it for them, those who live on-campus and are on the university's network should use Debian because of the ease of maintenance. Others claim that Mandrake/RedHat/SuSE are more user friendly in general, and so they should be advocated instead.
In any case, we're doing what we can to let starving students know that they don't have to shell out hundreds of dollars to feed an addiction to proprietary software, when perfectly usable and functional Open Source alternatives exist for them. KDE+Mozilla+OpenOffice+Evolution is a powerful combination that makes Linux very much a viable desktop operating system.
Plus, anyone who switches over has the best support team around: the campus Unix Users Group! A perusal of our mailing list shows that we don't sleep at night until your problem is solved. :-)
An unjust law is no law at all. - St. Augustine
A lot of people run Linux. A lot MORE have "tried" it, and then say to themselves "then what"?
;-)
Linux just doesn't have any good, free software, and that's what's needed to run a desktop.
At my last company, when I complained about Office attachments on the email and intraweb (against agreed-upon policy), the IT guy just gives me an Office CD and winks. When I state I run Linux at home, I get the "it's not my fault is it" (with the look of "you know, if it hurts when you slam the door on your head don't do it" look).
Linux will not even BEGIN to be appealing until people can "take their work home" (Office warez CD). As cool as CodeWeavers Crossover is - I've used it - it isn't "free" with the OS.
That's not a slam - I encourage commercial software on Linux, but the office-worker-at-home and the AOL user -- the majority of Windows users -- just want everything for free. They don't believe in Free Software or the GPL, and they don't believe installing MS Project on every computer is really stealing.
Eleet coder wanna-bees is another group -- slightly more technical than Mom -- that Linux won't win over. These people download the ISO's as soon as their released, burn em, but only try every 3rd release and then on a spare computer. Since Linux won't run his pirated games (or at least not full speed), Linux sucks. Besides, you can't run MS Visual Basic on Linux, which is an industry standard. Everyone knows you gotta program Linux in Assembly, or sometimes C.
For Linux to become more appealing to the masses, it doesn't need a lot of polish -- it's "good enough" right now. What's needed is for Microsoft needs to get tougher on licensing, which they won't do UNTIL they are SURE they have locked out the threats (by extending the Internet, apparently)
Apple has you beat by a mile. It's been dying twice a year since the mid-80's.
You like your Macintosh better than me, don't you Dave? Dave? Can you hear me Dave?
Desktop Linux is far from dead. It's NOT dead.
Just that it's not heading in the right direction.
Lots of things have been said about the ease of use thingy, but that's just scratching the surface.
What's important, looking at the larger picture, is that Linux is filled with programmers wearing beany caps.
Translation : Linux programs are wonderful, but it's just NOT the world needs.
Look at Windows. Lots of clumpsy and over bloated programs, but at least, they do what the world wants, and buys !
We have put too much emphasis on SOURCE CODE, because we wear beany caps - that is, we are the people who almost always CHANGE THE PROGRAM BEHAVIOR OURSELVES, that's why we demand the source code to the program.
But the world outside of us is that people do NOT want or need or know how to change the program's behavior, all they want is that the program does what they want - whatever they want.
That's why we have NORTON UTILITIES for Windows, and there's none of Linux.
That's why we have so much MUSIC, MP3, STREAMING, VIDEO, MULTIMEDIA utilities for Windows
On the other hand, what do we have here ?
KDE, GNOME, ENLIGHTENMENT, yeah, big deal !
The users need MORE THAN WINDOWING ENVIRONMENTS, they need UTILITIES that do stuffs for them !
That's what we fall short on.
That's what we need to double and tripple our efforts on.
Not that we do not have the knowhow to do it, nor that we don't have the programmer-aid to do it.
We have Kylix from Borland (FREE !) and how many of us are using Kylix to develop USEFUL UTILITIES for the users ?
Do something about this problem and we will see the Desktop Linux comes alive.
Muchas Gracias, Señor Edward Snowden !
What, again?
Exactly what I thought. People are so busy planning grand futures for Linux, and so disappointed when the software evolution fails to take us there, that they forget to enjoy the present.
Linux will have a future. Just take my word for it. The journey, however, is more important than the destination.
"I have opinions of my own, strong opinions, but I don't always agree with them." -- George H. W. Bush
Desktop Linux (and BSD, excepting MaxOS X) is really only appropriate at large installations where the environment is completely controlled and administered by professionals. While it's fine for a power user to install on their home computer, it really isn't appropriate for mom and pop. For that matter, neither is Windows. This means that desktop Linux is most likely to be found supporting scientific applications, Software development houses, Health care support, corporate desktops, data entry and call centers, and cash registers. It may become a viable home desktop system in the third world, should countries like China, Korea, Peru, etc decide to invest the money necessary to create localized infrastructure to support a wide scale Linux deployment for it's citizens similar to the old teletext systems used in Europe.
To proclaim that desktop linux is dead is foolish though. I've seen some very large scale desktop Linux deployments Boston area genomics companies, universities, and software houses. These are often commercial Unix to Linux migrations, so I'm not arguing that it's hitting the Windows desktop market hard. But if you know your stuff there's definitely work to be had in this market. As long as I'm paid well for this stuff, I'd hardly call it dead! --M
I think the main issue that's preventing most people from switching is that it isn't worth it. Linux, on the desktop, is not that much better than Windows XP on the desktop. Its not noticibly more stable, its not noticibly faster, but there are noticible downsides (application support and ease-of-use) to using it. I've been running Linux on a desktop machine for years now, and have recently settled in pretty well with KDE 3.0 and Gentoo. I use it not because it really gains me any technical merit I don't get in Windows XP, but because I hate Microsoft, the windows-style command line interface, and that blasted tooltip that keeps popping up in the corner of my screen in XP. Still, whenever I boot back into XP (to run Photoshop or the occasional game) I have to admit that Linux really isn't technically superior anymore, at least not in ways that a desktop user would notice. XP is reasonably fast, reasonably stable, and reasonably easy to use. For those less rabid then me, then, its an easy choice. They can endure the pain of switching to Linux, for a dubious set of benifets, or they can stay with Windows. This has been the situation forever. Why did MacOS never manage to take back its market share from Windows? Its been superior (from an average desktop user's point of view) for a very long time. Simply because people didn't percieve enough benifet from doing it. Windows was *good enough* compared to what MacOS was at the time. Now, if the timing had been different, had a Linux 2.4/KDE 3.0-style desktop been available around the introduction of Windows 95, would Linux have taken off? Hell ya. People would have seen a significant benifet in moving to Linux. Thus, if Linux ever wants to beat Microsoft on the desktop, it can't settle for being a "better Windows." It has to be *more*. Not just different, but a generation ahead technically. Now, this is what Microsoft does best. When they're not designing stuff like Palladium, MS engineers come up with genuinely cool stuff. A lot of it may be ripped of from other sources, and the first implementations may be less than perfect, but overall, they keep advancing the desktop. If Linux wants to be the next Windows, it has to beat Microsoft at its own game. It has to think up the next generation of user inteface and implement it before Microsoft can.
A deep unwavering belief is a sure sign you're missing something...
I am serious about putting TechnophobeLinux together. Please reply if you are interested in helping.
I hope you change the name for Christ's sake. Hell, even "AnalProbeLinux" would sound friendlier than that!
Programmer man-hours are a limited resource, whether you work for Microsoft or Linux. Linux has a larger talent pool, but that effort is divided into dozens of differnet desktop enviroments, all of which, IMHO, are inferior to windows (and I haven't seen whole lot of improvement here, either). Konq is a terrible way of browsing the file system, not to mention slow. You can't even copy/cut/paste reliably between applications. So forget about coding for them. Linux is wonderful for programming and remote access, speed and reliability, but when it comes to the UI, it stinks.
To make laws that man cannot, and will not obey, serves to bring all law into contempt.
--E.C. Stanton
It's quit tempting to look on Windows as the Nazis, or the Mongol Horde -- a force that must be crushed if civilization is to be saved. If this is your working analogy, then there is only total victory or inglorious death.
However, much as we might like it, the world is not populated by dragons and operating systems are not the tools of St. George.
Linux is not dead. Not now, nor is that a likely event any time in the near future. It's equally unlikely that Linux will soon drive Windows into the sea.
Windows will continue to be dominant on the consumer desktop for the immediate future. Windows has the applications, the games, and the thousands of developers grinding out the product. Could they do better work on Linux? Possibly, but it's not going to happen. Not with a relatively tiny marketplace further divided by flavors of installation and interface.
Linux will continue to drive servers and as the desktop of enthusiasts. It's a niche operating system, now, and likely forever.
For those that gnash their teeth over the evil empire, fear not! All empires crumble with time. But when something comes to push back the dark forces of Mordor, it will almost certainly NOT be Linux. It will be something clean and new, something that has a Vision (upper case "V") of computer interaction that goes past the creaky, cranky interfaces we have now and gives us a new way to relate to our machines. When it happens, Windows will go into the C/PM bin before Bill Gates can debug his digital living room.
And Linux will still be there, clanking along, doing it's job.
There is some space between death and triumph. Kind of like Switzerland.
Linux is doing fine thank you.
Oh sure, it is a bit slow selling on desktops but that will change as more and more consumers find out that Microsoft can more than double the cost of every PC you need.
The Microsoft office suite is $400 or so a seat. And, they are getting nasty about blocking the install on home, laptop and second or third systems by the same person. For $76, StarOffice suggests 5 personal installs. And, if $76 is too stiff, use OpenOffice.
Once the white box boys figure out that they can deliver all PCs with a free copy of OpenOffice and simply charge $15 or so to have it preinstalled, the casual market for the Microsoft Suite could dry up completely. And, the same may be true with large organizations such as corporations, governments, etc. Why spend $300-600 more per PC when you can go with linux, OpenOffice or StarOffice and double the number of new machines you buy?
Money is money.
And, right now money favors linux hands down.
Plus, that does not take into account the progress that Xandros, Lindows and others are making to expand the number of viable desktop systems under the linux banner.
The absence of QuickBooks, TurboTax and a few other key applications is a problem right now. GNUCash is fine. And, other software does substitute for much of what people think they need Microsoft for. But, it takes time for that information to filter out. But, it will filter out. Those who sell PCs (not the big OEMs) will be taking the lead packaging complete systems including software for a whole lot less than the Microsoft burden. Then customers can decide if the extra money is really worth it. It is not if you can make the choice.
And, if you write custom applications anyway, Java or Delphi/Kylix is right there to give you the same powerful GUI based RAD development systems you expect on Microsoft stuff.
The more machines you need the bigger the price benefit helps linux.
And, if you think that consumer PC buyers really want to pay twice the price for a system just because it has some Microsoft software on it that they rarely use, you are crasy. The typical consumer simply is unaware of what they can buy and use. That will change.
NexuSys - Linux support by the best
I couldn't agree more with the parent poster. It's not "Linux on the Desktop" that's dead, but the DESKTOP itself that's dead (or dying).
Normal people don't want to use computers, in general. They want to do tasks that they consider worthwhile. They want to communicate with others asynchronously. CURRENTLY, this is done through email, and CURRENTLY it requires a computer. Who says email NEEDS to require a computer? What if your email could be read to you automatically when you walked into your apartment? Most people would see this as a usability improvement over:
1. Sit down
2. Turn computer on
3. Wait
4. Double-click
5. Wait while phone dials
6. Click
7. Click
8. Scroll
9. Click
10. Click
11. Stand up
People don't want to use computers. They want to get things done. They want to create letters and presentations. Currently this requires a computer , a printer, and a lot of typing. Does it have to be this way? No! A lot of research has gone into voice recognition and computer vision. In the future we'll just describe a document or presentation in basic terms, using a natural interface like voice or gestures, and a device will spit out what was requested.
I predict computing's next "killer app" will be something that allows people to get rid of their computers.
As a power user on OS X I don't feel constricted by this. I still run X and various Unix tools thanks to fink and I find the UI to be straightforward and easy to use. In other words, the simplicity helps me get on with stuff rather than wasting hours reading through FAQs or HOWTOs just trying to figure how to share a folder or whatnot.
The same cannot be said for a Linux desktop. I'm constantly wasting my time trying to find some stupid option in the zillion control panels KDE/GNOME puts up for me, or swearing at the stupid help system that doesn't integrate distro help with KDE/GNOME help with manpage help etc., or scratching my head trying to figure out to get my scanner to be recognized, or grinding my teeth because the distro fills its multiple menus of apps with cryptic apps with names starting with g or k.
It doesn't have to be that way. Unless Linux becomes usable for everyone, not just experts it will never get on the deskop. Besides, the more users there are, the more jobs there are for admins and developers to meet demand. I would have thought it's in everyone's interest to see it succeed.
Linux is strongest as a server.
It's easier to enter that market and to build a reputation. That part of Linux is working very well for the community. With all the news about various companies using Linux for processing vastly significant amounts of data for vastly significant purposes, in some aspects, Linux is leaving all others in the dust.
It's Linux's reputation that will eventually bring it to the desktop, however. It's not the eye-candy of elightenment. It's not all the cool object-oriented inner-workings of GNOME. The reputation of Linux's reliability, availability and affordability that will eventually pull it onto desktops of home and corporate users.
First and foremost, if a more agressive push to the desktop is to happen any time soon, is to more completely and accurately emulate the Windows look and feel. It doesn't matter that it's "inferior." The "inferior" argument hasn't held since day-1. It needs to be familiar to the people who want to use it. If they expect "Network Neighborhood" then give'm Network Neighborhood.
It is not yet time to strengthen the weaknesses at the expense of existing stengths. Linux has a lot of strong points that are not being put to full use.
The demand for the desktop will come in time but there should be no major push for it. If there were to be a huge push for it, it would mean a radical series of changes such as a more well-defined "LSB" and strict adherance to it. We would need to come up with a "Linux Standard Desktop" definition that GNOME and KDE and any other players should target themselves to. Graphics and multimedia standards will have to be rigidly defined and adhered to.
These changes would have to happen very quickly and abruptly. It would cause a great deal of stress and confusion across the board. I say let it happen gradually and take the pressure off the desktop developers. There is no rush... not yet anyway. (Maybe after Win2k is pulled from the shelves.)
In the mean time, keep "Linux" in the public's eye and make them want it more and more by focusing on it's existing and growing strengths. Showing the public a weak, buggy and kludgey desktop will only sour public opinion regardless of how much work and pride it represents the developers. The "first impression" will stick regardless of what changes happen after the fact.
Linux on the desktop is not ready for prime-time. Let's not put it out there until it's ready. For now, let it remain the domain of the "L337" and let the public have Windows + Samba.
Having C:\Windows and C:\Program Files is okay on a windows box; they're just points of no-entry, aka advanced stuff you never need to look at. Instead you have "My Documents" to put documents into, and "My Pictures" for pictures. As you get more advanced you could even install a new program. It goes, logically into Program Files, and you get a link automaticly in the start menu.
Now lets look at the linux version. There's /home/joeuser, which has nothing. You could add, say, documents, pictures to it. So now we have those two nice folders. Now Joe is feeling brave and starts learning about his computer. He finds in his home dir: .bash_history, .kde3, .mcop, .mozilla, .qt, .bashrc, .DCOPserver_localname_localname_0, .ICEauthority, .kderc, .mcoprc and .xftcache.
Okay, well, so be it - just ignore all of that. Now Joe wants to install NewCoolApp. He starts the installer that was written up for TechnophobeLinux, which kindly asks him to provide the administrator password for the installation. Said and done, and the installer spews files all over his disk. They go into /bin, /usr/bin, /usr/local/whatever, /usr/share/whatever, a bunch of man directories, some in /etc, and maybe some in /opt/whatever as well.
Honestly, how many people here have actually read the guidelines for filesystem layout? I know which stuff goes in /bin as opposed to /usr/bin (which is also mostly different on different distributions btw...), but Mr. User is most likely to have one partition for everything on his simple desktop system, and none of it matters. Say what you want about the stupidity of putting apps in C:\Program Files\Vendor\ProgramName but at least it's fairly obvious that the "program files" end up under "Program Files" (duh) and possibly C:\Win(NT|dows)\System, which kinda makes sense since they're system files.
Joe is going to have a lot of questions rather quickly. For instance, why isn't there user stuff in /usr? Who is /usr/share shared with? What's optional with /opt and why isn't the rest optional? And why is my home directory full of config files if config files go into /etc? And why are there at least two */bin dirs (containing not only binaries but other runnable files btw)?
Say what you want about the Windows registry, but at least it's not laying around in plain view in Joe's home directory. And separating /bin and /usr/bin makes perfect sense on a server handled by a skilled person who could actually do something if /usr would be unavailable anyway - Joe certainly wouldn't be able to poke around the system using /bin and /sbin tools to set things right.
If you're truly going for an easy-to-use idiot-friendly linux, you're going to have to take some tough decitions. Toss the old layout out the window, pick something like /apps, /config, /system, /documentation, or whatever - and spend a long time compiling stuff from scratch to make it work. I once had plans to do this but never reached anything usable (see LFS for a good beginning). You will probably be flamed until you glow red from people saying you're fragmenting the standard and what-not, but sorry guys, the current layout is for server-techs, not for Joe.
(Sorry about the rant)
If you feel like actually doing something like that, feel free to contact me.
Tomorrow will be cancelled due to lack of interest
Nop, you need the subdirs. I think the "problem" has no solution. Windows is organized arround APPLICATIONS. Unix is organized arround TOOLS (little specific programs). Now we have a mix of tools and applications (application beign openffice, xine, etc.).
/progs meta filesystem. Like /proc this /progs could map all files installed by a package.
/prog/apps/openoffice/ would list all files instaleed by openoffice regardless of if they where mixed with other files (say you used /usr/local as prefix and not /opt/openoffice).
/progs/apps/openoffice to uninstall a program! Dependencies could be mapped to that dir also!
Linux will always have this duality. If you know Unix, you will learn to chain the tools and do productive things, as well as learn the applications you need. Joe Average doesn't need this and will NOT learn them. He only cares about the applications: separate programs that are standalone where you only need to learn to use the app.
Joe uses the PC for certain fixed problems solved with apps and sacrifices some flexibility in favour of user-friendlyness.
Power users will learn the difficult way of doing things. They will be able to do complex stuff the standalone application can't (chaining tools, scripting, batchs).
The inherent complexity of the file system could only be solved by mappings. If you are Joe Average, you could be presented a different file structure. But all the tools will still have to be somewhere regardless you seen them or not.
A needed step would be database of where when what and beign able to present a list of every file an application installed. Something like an explicit package management: say a
Example:
I'd love (and probably many others will) to use rm -rf
Using a registry could be a problem, unless filesystem maped.
unfinished: (adj.)