Windows and Linux User Interfaces

Anonymous Coward writes "Greg Raiz, Boston based interface designer and former Microsftie takes a look at Linux and outlines key shortcomings and strengths of an OS that could take on a giant."

Comments

Maybe true, but not necessarily desirable

[Tet]

To gain momentum Linux needs a central installation architecture that all applications must use to properly install and run. The OS should ensure that applications are installed before they can be executed.

Perhaps there's some truth to this. If Linux is to gain more widespread adoption, then maybe that would help. If so, then I personally hope Linux remains a niche OS. What he doesn't seem to grasp is that some of us would rather remain true to the Unix ideals and philosophy than to chase mass market popularity. I want to just be able to extract an archive and run a binary contained within. I don't want to have to inform the OS that I've done so, and have to "install" the software. I want to be able to compile an app and run it from my home directory. Why should I have to register it with the OS in order to do so?

Re:Maybe true, but not necessarily desirable

[AKAImBatman]

No, I don't think it's really about the fragmentation. It's about the mindset. You can always teach someone to use a modern Linux distro, as many anecdotes have recounted. As long as you're there to answer questions, the users can make it through the difficult parts. The problem is that this isn't 1995 anymore. Back then GUIs were a new thing for the IBM PC, and everyone was willing to seek out a local computer geek to help them make it work. Now that users are comfortable in Windows, however, why should they switch to a completely different system that requires them to relearn how to do everything?

Apple has answered that question. "If you move to the Mac, your applications are simple to install, your files are well laid out, the computer self-manages itself, the user interface is less confusing, you can quickly search for files, organize them in new way, nearly all maintenence is automatic, the system is free of spyware and viruses, AND you can still use Microsoft Office!."

The Linux community's answer has been, "Hey, we don't have viruses either! Erm... except for those one or two. But someone released an anti-virus that spread itself to eliminate the first virus! Oh, and did I mention that it's free! And you can have a home server!"

Unfortunately, the Linux answer has not been very appealing to the market.

This link contains a great analogy of the situation.

Re:Maybe true, but not necessarily desirable

[ScuzzMonkey]

Don't put absurd requirements for Linux that the latest MS Windows can't accomplish..

But that was exactly his point--Linux NEEDS to be able to accomplish things Windows can't accomplish, dramatic and useful things, to overcome the barriers to adoption. I think the risk-free install is a bit pie-in-the-sky, but his point is well-taken... there is an opportunity to do some big, dramatic things to make it easy to adopt Linux on the desktop, because there is no corporate imperative in the way. If you saddle yourself to only to features that the latest Windows can accomplish, you're discarding one of the main advantages in the fight.

Re:Maybe true, but not necessarily desirable

[Jesus_666]

Note that you don't register your apps with OS X, which I like. Most apps should not need to access system configuration files so there's no need for them to mess with the system in any way. So why tell the system you're there? There's only a few apps that would profit from this (the ones that actually extend the system, like kernel hacks), everywhere else it's just useless data.

Instead of giving Linux the same user-hostile installation routine as Windows (Apps scatter their stuff all over the system, which of course makes an uninstallation routine necessary) I'd like them to adopt a shell-friendly version of Apple's Application Bundles. I'm thinking of something like this:

~ $ tar -xjf Gonkulator.app.tar.bz2
~ $ which gonkulator
which: no gonkulator in (/usr/local/bin:/usr/bin:/bin:/opt/bin:/home/user/ Apps)
~ $ mv Gonkulator.app ~/Apps
~ $ which gonkulator
/home/user/Apps/Gonkulator.app/Contents/Linux/gonk ulator
(The App Bundle handler recognizes folders with the .app extension and scans them for executables - probably using an extended Info.plist)

Linux is FINE

[Work Account]

I have been reading about Linux being "ready for the desktop" for like 8 years now.

I for one am sick of it.

I use Linux every day on the desktop.

Yes, at first it was a bit confusing, but over the years it has matured ten-fold.

My parents use it, my grandmother runs Fedora, and I convert others on a daily basis.

ENOUGH already with this GUI/desktop debate. It is over and done and we have done it.

I beg to differ

[Janitha]

Linux should stop copying Microsoft feature for feature and embrace the differences and features that advanced users love.

I would disagree. What about enlightenment, fluxbox, openbox?

This article doesn't really make much sense overall.

Re:I beg to differ

[AnonymousBystander]

Linux should stop copying Microsoft feature for feature and embrace the differences and features that advanced users love.

that would contradict

As a general rule most people do not enjoy switching, upgrading or installing anything new

SUN dropped GNOME as Desktop

[Anonymous Coward]

Sun seem to be dropping GNOME as their Desktop

http://www.gnome.org/~gman/blog/02112005

Read more on above link!!!!

The Challenger: OS X versus Linux

[ZachPruckowski]

This probably won't be a popular comment, but I think Mac OS will eventually be bigger on the desktop than Linux.

1) Easier Support - your computer breaks, you know who to go to
2) Less of a learning curve.
3) Less confusing in terms of options (there are a lot of types and kinds of Linux, or so it seems).
4) Media acceptance. Macs are more well known than Linux, which isn't Linux's fault, it's just the fact that OS X has Apple behind it.
5) Application Support - Things are ported to Mac quicker than to Linux usually. Apple also stands to get more software compatibility when they go to Intel computers.

::Braces for "-1 Flamebait"::

The answer you wanted

[Work Account]

SUPPORT.

As a Linux developer you and I often deal with companies that will not publish open specifications regarding their hardware.

As such, it is necessary to "break the law" and reverse engineer these devices in order to create decent Linux drivers that interface between the device and the application/user level software on the GNU/Linux kernel and operating system tools.

Some say that if Linux slowly gains market share of say 20-30% that manufacturers will stop making Windows-specific devices.

Another benefit would be support: all of us Slashdot/Linux guys would instantly become experts on people's home PCs if everyone ran Linux.

As it stands now, most of my friends have trouble figuring out how to .ZIP up files on Windows or do other simple things such as send an email, etc.

It is funny/amazing to watch them because they write code like protein-folding applications in Linux but on a Windows computer all they can manage to figure out is how to start Solitaire :)

Re:The answer you wanted

[AKAImBatman]

Welcome to the Chicken and Egg problem. Hardware manufacturers and Software producers are not going to support Linux until it's popular. Yet Linux won't become popular until the Hardware and Software vendors support it. As a result, you need to make the product more attractive to users in another way.

Apple has the right idea. By pushing the technology far ahead of the competition, they convince users to accept some of the shortcomings in exchange for a large number of features unavailable on other platforms. This increases the Mac user base, forcing the Hardware and Software vendors to support them.

I know in my own personal case, I could run regularly Linux right now if I wanted to. The problem is that my kids have a large number of educational titles and other kids' software (no, they don't operate correctly under WINE) that they must reboot to use. If I leave the system in Linux, I hear no end of complaining from my wife who has to reboot the next day.

Now I (and a few others) have suggested methods by which the Linux Desktop could pull ahead of Windows, and possibly even the Mac. My own suggestions would be distro specific and would not harm any existing distros. Yet the community resists such changes strongly, stating that "Linux is perfect the way it is". Many automatically assume things I didn't say, based on their past experience.

So in the meantime, I and my small team will attempt to implement these ideas whenever we have sufficient time. Unfortunately, it is very difficult to reconcile our schedules, and without more volunteers I fear that the project will not get done in time to make a difference. Which is really too bad, as I feel that it could make Linux a significant competitior in the Desktop market.

Ever heard of Apt? Emerge?

[quakeroatz]

As a corollary to the first rule, users don't like installing applications. Part of the fragmentation problem for Linux is that the fragmentation forces a problem for software installation. Users are forced to untar, un-gzip, copy, configure and sometimes compile in order to properly install software. To gain momentum Linux needs a central installation architecture that all applications must use to properly install and run. The OS should ensure that applications are installed before they can be executed.

Wow, I couldn't have described apt or emerge any better. Isn't it common that those who review Linux OS's vs. Windows almost always head to the biggest vendor (Redhat) which is exactly the wrong idea: directly motivated by Microsoft's position on the closed source market? Biggest is best is necessarily a universal philosophy. Also, there are rpm managers in Redhat that do the same thing as apt, I think you can even use apt on Redhat without too much trouble.

Sure one might say, "How would the avg. Windows user know to apt-get install ?"
I would answer, "They could figure that out long before they understood how to dl and compile source code, and would certainly require less user knowledge and decisions than going apt-get install , which rarely asks for user input"

I see a ton of skilled Windows IT folk that are scared away from Linux because they try to compile everything. Apparently they haven't heard, and/or common linux knowledge doesn't include important tips that would make Avg Joe Windows user's first Linux experience much more enjoyable.

Re:I do NOT think Linux needs an install architect

[uofitorn]

It's implicitly assumed he meant it was a niche OS among the Windows/Home PC crowd.

While I agree with your comments for the most part, to say that the majority of all new servers are slated to run GNU/Linux is typical of the naive slashdot groupthink and is not remotely true. However, I'm in the middle of installing FreeBSD 5.4 on a Sun Blade 100 machine so I'm doing my part!

wireless support on linux is horrible

[taxman_10m]

After a few more days I'm probably going to wipe my linux partition unless I can get my Linksys wireless PCI card working with WPA encryption. Who knew it would take an act of God to configure correctly?

Fedora Core 4's network configuration gui is worthless. Ndiswrapper hung the machine. And it took me hours and hours to find that I needed some WPA supplicant something.

Re:wireless support on linux is horrible

[WhiteWolf666]

Before you wipe your Linux install completely, try SuSE 10.0 (http://opensuse.org/

Ndiswrapper, and WPA_supplicant, IIRC are part of the base install. You'll have to use ndiswrapper commandline, but the instructions are in the documentation (avaliable on your Kicker as the help icon). Then use YaST2 to configure your network WPA password.

SuSE does the best job regarding wireless configuration of all the linux distributions I've tried. It's not perfect; the Kwireless applet they use to a bit clunky for switching networks. OS X has it soundly beaten in that regards. But it works, WPA included.

Microsoft Stole Gorm!

[diablo-d3]

According to this insider's blog, Microsoft has stole Gorm, among other open source software.

Re:To me, this issue always disturbs me

[gg3po]

I call BS. I'm a graphic designer. I work on a Mac (PS, Illustrator, InDesign) at work, but a Linux box at home. I've done a lot of freelance work on Linux from home lately. I typeset and did all prepress for a 700 page book over the summer using Scribus, Gimp, and Inkscape. I keep hearing about how Linux fonts suck, but they always look great to me. *Much* better than the non-anti-aliased fonts that are the default on Windows. In fact, the only way I can make fonts look half-way decent in Winbl0w$ is to turn on their crappy ClearType® feature, but I've noticed that even that screws up the colors at the edges of the characters. If you have some specific examples of font issues, please link to some screenshots, until then I'm of the opinion that the "fonts suck on Linux" argument is a complete troll.

One Thing He Left Out...

[eno2001]

There are many areas where GNU/Linux distros (and the other *nix-like OSes) go far beyond anything that is possible in Windows or on the MacOS in terms of features. Many of these "beyond" areas would be VERY attractive to the average user. Unfortunately many of them are obscured behind the CLI or very complicated concepts. A few examples:

1. The xmdx extension for X window system (X.org) which would allow multiple machines to act as one shared screen over the network. Combined with the proper simple user interface and an xmdx aware pager, A user could execute their web browser on Machine A and go surfing. They could then drag-and-drop the browser to Machine B's desktop and keep on going down there. If this was further combined with an xmdx aware sound server, A music player could be made to follow it's user from machine to machine without ever stopping.

2. Virtualization might seem like a concept that would be useless to grandma, but you're not thinking straight if you believe that. If a GNU/Linux distro were set up to to run on top of a Xen paravirtualization environment in a transparent way and across multiple machines, imagine the user friendliness... To grandma, it looks like a desktop that is always where she left it and it never stops. She can shut her machine down and the Xen domain would migrate to the central home computer/data store.

3. Clustering. Again, a lot of people would think it's a dumb idea for "Joe Average" to have a cluster. But is it REALLY a dumb idea? I say no. Why should people be forced to throw away old computer systems once the latest version of Windows won't install? Why can't they just have an automatic cluster solution with a very transparent UI that provides them with MORE power than they would ever get from a single Windows box?

Just in general, the key should be to take very advanced concepts that don't even exist in the Windows world and make them available to the end-user in a very simple, transparent way. This is all possible with Linux. But most Linux folks don't think this way and therein lies the problem.

right on the point

[kobold2]

Progress has been quick to match features with Redmond but this type of progress will only allow Linux to play catch-up, never to lead. In order to break away Linux has to do the things that Microsoft hasn't done or perhaps will never do to differentiate and become a practical desktop alternative. Linux is verbatim free to do whatever it can. :-)

Why this elitist attitude...?

[cr0sh]

so basically the strategy here is to take the shittiest developers from the windows platform and get them to write garbage code on linux?

This kind of attitude I am certain holds back many people who would be adopters and great supporters of Linux. I am certain there are many VB and other 'doze programmers who would be readily willing to help develop the pieces necessary to create a "VB-like" development system under Linux, but I bet they are put off by this attitude.

I have used and developed on Linux since 1997 - not a veteran by any means, but I have been here a long time. Guess what: I am also a VB programmer (VB6 was the last version I used at a former job - I don't use it at home anymore). I don't write "shitty" code. My code is clean, well structured, and well commented. Regardless of what language I use (and I know a slew of them), I always try to make my code shine first (maintainable), then optimize it where it is needed last. If such optimization requires me to "roll up my sleeves" and get down-n-dirty with some C/C++ and/or assembler, so be it (although this need hasn't been true for a while - last time I had to do some C coding was when I created a scanline triangle rasterizer for custom 3D engine I was coding in VB - ahem).

All of this isn't to say there aren't crappy VB coders - but there are just as many crappy C, C++, Perl, Python, etc coders as there are crappy VB coders. In the end, the language is just a syntax to tell the computer what to do. How you code and structure your program according to the syntax of that language is up to the individual programmer. If he is bad, the code will be bad, if he is good, the code will also reflect this. Personally, some of the crappiest code I have seen has tended to commercial game development houses (although I do give props to ID for having good code that is fairly readable, with comments where absolutely needed, and clear layout of algorithms where you can easily read the code and figure it out without too much headache). This could be due to a number of reasons (too many cooks and such), but then again, I have seen excellent code created by other commercial software companies as well. There are excellent VB coders out there - all one has to do is read a few back issues of Visual Basic Developers Journal to see that (the code they publish is *excellent* code, in terms of structure and readability).

Painting all VB developers as being "shitty" programmers does nothing but disservice to the abilities of those VB programmers who develop great code and software. It also does nothing to help their perception of the Linux development community. These are the programmers that the Linux community needs if it is ever to make serious inroads to the desktops in businesses. Many businesses worldwide utilize tons of in-house (and otherwise) custom developed VB applications, running on their Windows desktops, that aren't typically represented at all by similar software available for Linux. If the desktop is ever really wanted to be "owned" by Linux, the community of developers on Linux needs to see past the arrogant and elitist attitude it has towards VB developers, and instead extend a helping hand toward getting them onboard helping to develop an easy-to-use, RAD tool for the Linux desktop (it is my personal opinion that we are already 99% of the way there - notwithstanding the various VB-like RAD tools that do exist for Linux, I believe a modern approach could utilize Qt or Gtk coupled with Python under KDevelop or a similar IDE to create a very nice and extensible RAD tool that would attract many VB developers and others to developing software for Linux desktops).

Windows-to-Linux conversion tool?

[JonToycrafter]

The "get people to switch from Windows" debate is raging again here - it reminded me that I wanted to know if there was a Linux-to-Windows conversion tool to ease the process. If not, let me say what I'm thinking.

I'd like a Windows executable that will scan my system, identify settings (TCP/IP settings, SMTP settings, dialup/VPN, background desktop image, you name it), and burn a CD of my settings. Then, I want a Linux executable that will read those settings, and set me up in Linux as close as possible to Windows.

This tool should ideally also work Windows-to-Windows for moving to a new computer. Ideally it should have a plugin architecture so folks can write add-ons. The XMMS folks can write a plugin to suck in my Winamp settings and so on. Done correctly, this tool could even analyze my installed programs and suggest what programs I'm going to need. "I see you have Yahoo Messenger - you'll want to get Yahoo Messenger for Linux or Gaim. Once it's installed, I'll pre-populate your settings."

Throw it all on a live CD and you have a great way to convince folks that switching to Linux is easy.

I'm not a developer, but I'm a Windows power user (the key demographic, yes?) who'd be happy to be on a team of folks interested in this.

Linux is Ready for Mac OS X Users to Switch...

[grouchofan]

My sister is a graphic artist in the DC area. Her company has had various restructurings and layoffs in recent years, dumping more work on her.

This got me thinking about how much money a typical graphic shop spends on Mac hardware and software. I wondered if a Mac shop could switch to Linux (PPC or x86), save money, and still be able to do everything they could before with a comparable ease of use. I spent a while analyzing OS X 10.4 Tiger and Linux in a series of articles on my blog.

What I came up with is that Linux itself is certainly easy enough for a Mac user to pick up, and can be customized to look and act enough like OS X that while there would be a learning curve, it wouldn't be a steep one.

There are lots of open source and free packages out there to do the kinds of things Mac designers do, and most of them compare favorably to their commercial counterparts on the Mac.

The only real drawbacks I saw were (as noted here in the article and commentary) that software installation is a touch easier on the Mac and that the Linux applications aren't (yet) capable of reading Macintosh files.

This led me to the conclusion that a "new" designer with no existing library of graphics in proprietary formats (e.g., PageMaker, FreeHand, Illustrator) could pretty easily use Linux and OSS.

A more experienced designer with lots of graphics in proprietary formats could still do it, but would have a heck of a time getting those graphics moved to Linux. In exchange for that effort, they'd save a ton of cash on hardware (since Linux runs on cheaper x86 iron) and software (since most of the Linux stuff is free).

If you would like to read the (even more) long-winded version, see my blog at http://mikesalsbury.com/mambo/content/view/243/