Linux *Won't* Fail on the Desktop?

HanzoSan sent in a story claiming that Linux will Succeed on the desktop, and not just the server market where it already has had much success. I think that the latest version of KDE has demonstrated that it can compete, but with the increasing dependance on file formats that have no support on Linux, it's going to be awfully difficult. That said, Linux has been my desktop for many moons, and I don't plan on changing it (Maybe If Apple released TiBook's with 3 mouse buttons I'd at least have an option ;)

Universal File Formats

[JohnHegarty]

Until some universal file formats are agress by all the compaines out there , then it will no take over. But when your document can be opened in an os , on any word processer... well that will be the end of ms won't it.....

Mindshare

[crumbz]

It is important to see Linux succeed as a viable alternative to Microsoft. Especially here in the U.S. With the global trend of open source software picking up steam (German govt., China), the U.S. cannot afford to rely on Microsoft as the provider for all desktop and server OS as well as major applications. Look at the security problems we are seeing currently and multiply it by an order of magnitude with each major OS upgrade.

It really comes down to a balance of money, intellectual property rights and giving users the tools they want. Let's hope that the U.S. doesn't squander it's lead in this area because of a lack of options.

Hmmmmm...

[PowerTroll 5000]

I support disseminating Linux as freely as AOL does its CDs.

Perhaps that might be a good idea. The big advantage of free software is the fact that this could be done. You can't beat the price. However, people do not have the same awareness of Linux as they do AOL.

How about an ad-campaign a la IBM Infrastructure commercials [slashdot.org] to explain Linux in plain English? Without awareness, few would be likely to pick up the CD.

It's more than just KDE

[Krusher55]

Succeeding on the desktop is more than just KDE or even nice applications. It requires substantial hardware support. People don't want to be severly limited in the type of printers, scanners, video cards, sound cards, etc. they use and they will expect them to work exactly the same as under their current system. The cost saving of using Linux is wiped out if you have to spend more to get a supported printer or if you have to spend an extra half hour figuring out how to change the resolution of your screen. Linux is still a little ways from that point so widespread desktop use is unlikely to happen any time soon.

Re:Well, what's the DESKTOP killer app?

[fallacy]

"But other than that, What is being offered? "

The opportunity to not be tied to endless EULAs, support contracts, pricey upgrades. To create an environment how *you* want it, not how someone else thinks you want it.

I'd rather be incompatible between versions *for free* thank you.

It's not about what's best

[mrblah]

What everyone keeps forgetting is that people don't care about making the best choice in operating systems. They're only concerned with making a good choice. If MS is good enough for the vast majority of users, then the new user is going to decide that it's good enough for them. MS is what "everybody else uses" ... and that's a strong message.

Build a tool ...

[linuxdoctor]

I especially like this suggestion which the author suggests as a paradigm shift: "Let's completely modularize each tool function (such as layout, fonts, kerning, textures, linking, math and tables) and make each a separate interactive GUI tool. Like an erector set, applications could be constructed for specific needs. And like hammers, saws, wrenches and screwdrivers in the physical realm, such tools are easier to utilize than large factories (or contemporary application programs)."

This is the classic call to arms of Unix, way back when. "Build a tool that does one job, and does that job well." And then make the tools work together. Unix was originally built for programmers, but there is no reason to believe that "ordinary users" cannot benefit from that philosophy as well.

I say, go back to first principles, and we all win. It worked for hardware in the 1980's with the advent of RISC technology. Software too has become too bloated.

A major problem with the article

[phaze3000]

To paraphrase open-source advocate Richard Stallman

I don't think RMS is going to like that one.. :)

Seriously though, I think there's one major issue which the article writer has forgotten: fear.

Many (most) IT directors think that 'No-one ever got fired for choosing Microsoft'. If they go with Linux and it's a failure, it could well be their neck on the line; if they choose a Microsoft option and it's a failure, well everyone already knew Microsoft were crap, but what choice did we have?

The only way this can be combated is with slow erosion of the Microsoft market - it used to be that "no-one ever got fired for choosing IBM", so it's certainally possible to topple the Microsoft monopoly - it just isn't going to happen overnight.

Re:If TiBooks had 3 mouse buttons?!?!

[pressman]

Ok, well, what's the difference between using a multiple button selection device on a laptop and using modifier keys to do the same thing? You're still essentially pressing keys to mimic a multi-button mouse. My left pinky finger has become quite adept at modifying mouse commands on my iBook. I don't miss my one button or multi-button mouse at all while using my laptop.

Re:Well, what's the DESKTOP killer app?

[R2.0]

That's why one of the settlement conditions offered by the 9 holdout states is so important: MS would be compelled to auction off licenses to develop Office for other OS's.

If IBM, Sun, maybe Corel or Redhat were to bid on it, MS couldn't complain that it got a raw deal (although they will anyway), and Office will be ported to Linux.

As a bonus, it will be pried open so that maybe MS will have some incentive to fix it. I'd switch OS's just to get the pagination to work!

Re:Well, what's the DESKTOP killer app?

[silicon_synapse]

But most people just don't care. They're perfectly happy to be running Office 97 on Windows 98 with a picture of their grandkid as the wallpaper. Sure it crashes once in a while, but that's normal (in their point of view), right? If they switch OSs, they have to relearn the most fundamental ideas about how they use their system. Why should they bother?

---
Extra! Extra! Read all about it [slashdot.org]! Slashdot editors censor dissenters [slashdot.org].

One word...

[FortKnox]

...Marketing

I know flames will fly, and not a lot of people believe in it, but that's what MS has a big advantage in. People watch TV. People see MS ads. People might occasionally see an apple ad. People only see IBM's Linux Server ad (and the common person has no clue what its about).

Also, maybe having some local demo's in malls. Just to let people play with it, like they do in bestbuy, etc...
See what its like so you don't need to be afraid...
If someone made a good commerical ad and had demos in public places that showed how pretty it is, how inexpensive it is (people will need to buy it for the support), and how there aren't licenses and most everything is free, then you'd have a "general layman interest."

That "general layman interest" is a catalyst Linux needs. Its powerful. That's when people "try" things. Isn't that all we're asking for? Just "try" it??

From the article

[Selanit]

>In a nutshell, the Linux community must develop both
>a quality GUI system for configuring hardware and a
>standardized system for installing and removing
>software. Developers must be persuaded to provide
>Linux drivers, especially for "Winmodems," and to
>port their software products to Linux.

Agreed on the need for a GUI "system properties" type hardware configurator. KDE's hardware configuration leaves something to be desired. (Specifically, it doesn't offer much in the way of actual configuration options. If you want to do any non-trivial fiddling with your hardware, you might as well go straight to a console, 'cause you're going to need it anyway.)

As for installing and removing software, it would be good to have a more-or-less universal software management system. The two current contenders are RPM and Debian's apt-get, of course. Both have advantages and disadvantages -- for example, it's more common to find fresh builds of programs in .rpm format; but apt-get handles dependencies more gracefully. Perhaps what we need is a synthesis of the two, which would use the .rpm file format and apt-get's syntax. Instead of having a centralized package depot like apt, or many randomly distributed files like rpm, you strike a balance: maintain a server that lists current URLs for packages, which would be hosted on the project's page instead of centrally. Typing "rpm-get install Snicklefritz1.3" would check the central database for current URLs of the RPM and its dependency BruberMIPS0.9.5, download them from two different sites and install them. (Note: the "spell" system in Sorcery GNU/Linux [wox.org] works kind of like this, only it downloads source and auto-compiles instead of downloading pre-built packages.)

In addition to persuading companies to release Linux drivers for their hardware, we also need to convince them to open-source the drivers. I seem to recall ATI already did this. There is even less reason than usual to make your driver proprietary; after all, the driver is useless without the hardware to match. People would still have to buy the product in order to get use out of the driver, and in the meantime students could study the driver code to learn about low-level hardware interaction. And stuff. (nVidia, are you listening?)

Re:Universal File Formats

[Lumpy]

This is true if the IS and IT people out there stay lazy. Me? I configured EVERY new machine that comes in to make Word not to save as a DOC file. but as RTF. this didn't implode the whole business causing chaos and burning HR people with sales people flinging themselves out the windows as many Microsoft lovers here would like people to think. Noone noticed. RTF flies around fine.. and I now have people asking clients to send them a rtf file of that document.

This is how chaing to a universal format starts and spreads.. Non lazy IT admin makes a change... now if only another 50 IT admins do this... DOC would be a rarity within months.

Where is Linux's "Digital Hub" Strategy?

[desslok]

One of the killer things about my OS X Powerbook is how I can plug my digital camera or FireWire cd burner and it not only has a driver, it already knows what to do with the device. With my camera, it automatically asks me if I want to transfer all the pictures off the camera.

I've never bothered hooking it up to my Linux box. I'm sure nothing would happen.

Where Apple goes, often most of the industry follows. Jobs' "Digital Hub" strategy is dead on once you've seen it in action. It makes a computer really useful for the home user.

For the business, I am increasingly in doubt. Microsoft file formats are so common it's futile to try to use Linux in the office. If the free office suites do the job, fine. But I think the only place Linux will succeed are in custom installations (like the Burlington Coat Factory point of sales units) or where cost is essential (like the city using Linux for offices in the Florida Keys).

For the business user on the go, Linux won't make it unless there is a desktop with the kind of commercial development behind it like Apple's or Microsoft's. The level of integration and consistency of interface needed is far, far away in the Linux world.

Linux on the Desktop

[dTd]

I don't think this is really an issue. Linux has been my only OS at work and at home for 2 years now. Ever hear of plain text? Very cross platform compatible it is. Openoffice opens all the MS format docs I've ever had to open. I've not felt pressured in the least to install any other OS at all besides for curiosities sake. My wife uses linux, my mom uses linux, and she's just turned 60, my kids use linux. You just don't need anything else and if you think you do, you just haven't looked at all the apps that come in any new linux distribution. This desktop argument is stale, old news, and a waste of time. Linux has already taken over the desktop on the 6 machines I have access too, and quite a few more of my friends machines.

Re:Well, what's the DESKTOP killer app?

[Salsaman]

"Get the various "features" of broken web-browsers supported in Linux browsers."

Mozilla has a 'quirks' mode which does just that. Unless you declare strict html in your dtd, Moz will default to this mode.

Love Linux but Installing W2000

[Anonymous Coward]

I just finished a couple of weeks trying to be primarily a Linux client on our company network (I was using Mandrake 8.1 -- a Great release). I am now installing windoz 2000. Why? Not that I don't like linux, I've put it on multiple laptops, I have a server I run at home and I like to do my perl/cgi development on linux. No, it is because of these reasons:

1. No NOTES client. We use LN for e-mail and many DBs. Tried VMWARE desktop 3.0 -- too slow, frequent lock-ups (which require the blue checking HD deal -- and take time). Also didn't like the smaller screen (tried the full screen mode and this locked up both the VM and linux twice). Tried hitting LN through a browser (works, but doesn't have a fraction of the features and ease on the client).

2. Limited support for Netware. Only way to map to a network drive was to use the console to do ipx_configure and ncpmount -- it works and I can put it in my start-up script, but not easy for the average user ...

3. Never did get the network printer working. Tried HardDrake MANY times with MANY settings and never once had anything exit the printer. Even worse -- no messages at all about where my test pages might have gone (even an obscure queue not found message might have helped in my trouble shooting).

4. Getting sound working was a trial. After buying a new sound card and disabling the MB on-board sound, I still needed to purchase the OSS driver to get it working, but don't play around with the controls or you will have what sounds like a 78 piled high with dust. And volume is all over the map Xmms needs my volume WAY UP, but the Mandrake boot song WAY LOW - forget and you are blasted out.

5. Not being root all the time is the mantra - and yet everything I tried to do seemed to want me that way. SU all day long. No SU editor - I guess you need to evoke a graphical based editor from a console where you've logging as SU. Maybe I'm an idiot ... Next time I'd just ignore the warnings and install everything as root.

Enough rant ... Linux is neat ... Linux is stable ...

And yet it is still too complex for my average windows type user. So even though I feel like a turn-coat -- I am back to W2000 (actually NT4 has been a pretty good and bullet-proof OS for me over the last couple years) and a dual boot Drake 8.1 for development.

MacOS X beat ya to the punch...

[toupsie]

In less than one year since its release, Apple has smashed the years of KDE and GNOME on the UNIX desktop frontier with MacOS X. Frankly, KDE/GNOME look like redheaded stepchildren compared to MacOS X -- looks too much like Windows (YUCK!). Darwin + Quartz + Aqua is such a beautiful combination. Practically anything you can run on Linux (outside of network IP specific apps) can be compiled in MacOS X. Plus MacOS X has the "killer" productivity application, Micro$oft Office and the ultimate graphics app, Photoshop (GIMP is no where close--sorry). Your X Windows applications will run on MacOS X for the most part. About the only valid argument Linux users have against MacOS X is the cost of hardware, but that is just a short term cost.

Don't get me wrong. I don't hate Linux. Its a daily part of my life...as a server OS that maximizes my old i386 hardware. I won't be using MacOS X Server either. The PPC hardware is too nice to stuff in a closet. It begs to be used by human hands.

I think its time for hard core Linux zealots to really examine what a beauty MacOS X is. Pop over to CompUSA or an Apple Store, shove the crowd in front of the new iMac to the side, click on the Terminal icon and see what a pure UNIX experience is really like. After that, I think your fear of Steve Jobs and his magical black turtleneck will go away.

Note to CmdrTaco: If I hear another mouse button joke and Mac from you, I am going to hand Ms. Fent an original iMac hockey puck mouse so she can beat you into submission. The PowerBook G4 has USB, take some of that dowry and buy one.

It's a long shot.

[SlashChick]

I work with (and for!) a lot of people who use and like Windows. I was also part of a test group at our company that switched from Outlook / MS Office to Netscape 6.2 email / StarOffice, and I have several juicy tidbits regarding this that fit in well with this article.

I just got finished setting up three computers with Windows XP Home from Dell. Computers really are a commodity now -- the Dells were gorgeous, easy to open, and functioned perfectly for a cost of $588 each (shipped!) Google "Dell refurbished" for other good deals. But I digress.

I set my mother and my dad's secretary up with the new computers (two at the office and one at my parents' house.) Keep in mind that Windows XP is about as far from Windows 98 (which is what they had) as you can get while still being Windows, and Office XP is somewhat different from Office 2000.

With two clicks I had set up a system whereby they could connect the secretary's 56k modem (my parents live/work in the middle of nowhere) to the Internet and have everyone else's computer connect through hers. I then set up remote disconnect -- where it shows the icon in your system tray and you can connect and disconnect the modem from any computer in the office. Windows XP comes with a nifty disk that you can put into any Windows computer (besides Windows 2000) and set up the connection sharing.

With another few clicks I had set up the Files and Settings Transfer Wizard, which uses Ethernet or a serial cable to connect to the other computer and download settings (fonts, favorites, etc.) I even backed up other programs and had them transferred automatically.

When my mom plugs in a digital camera, a wizard pops up and shows her all the pictures on the camera. She can then copy them to a disk or to the hard drive. She can print 4x6s, 3x5s, or wallet prints from the OS. Burning files to a CD is as easy as selecting the files, dragging them to the CD drive, and clicking Write To CD. Yes, folks, Windows XP may have a whole host of Big Brother issues (most of which I turned off upon installation), but it sure is easier to use. The whole experience reminded me of the Macintosh.

Compare this with installing Linux. Even setting up Linux to see NTFS drives is a pain, let alone transferrring files and settings (since that is most likely what you are going to want to do upon installation.) I've used Mandrake pretty extensively, and even it has some weird problems (like asking which version of XFree86 you want to use, and not automatically detecting the monitor and setting a reasonable resolution.) It took me hours to figure out how to get Mandrake to change to a lower resolution (Ctrl + -). There is little documentation. And this is on Mandrake 8.1.

There is just a lot of stuff on Linux that is poorly documented and/or buggy, and that carries over to the Windows versions of open-source software in a lot of cases. Netscape 6.2 (which I am using on a daily basis) is easily one of the worst email clients I have ever used. It won't let you switch on-the-fly between text and HTML mode. Attachments randomly refuse to open. At least it's stable, which is more than I can say for any version of Staroffice (5.2 or 6.0 beta.) Save a file as Excel format? Crash. Open a large file? Crash. Apply special formatting? Crash. I'm running Windows 2000, so no, this has nothing to do with Microsoft. A favorite quote of a co-worker also on this project is "Yeah, I use StarOffice to open documents, but if I want to get any real work done, I just use Excel."

It's not there, and after seeing Windows XP (which, BTW, has no activation bull if you buy it preinstalled), I'm not convinced that it ever will be. I will happily use Linux on the server, but I consider Windows an excellent client OS.

See my post history / journal if you want more info.

Fragmentation is a bad thing

[pongo000]

Knowing the /. conventional wisdom on this subject, I'll probably get branded as a troll. But here's my take: So long as there is the degree of fragmentation in the Linux world as currently exists, a mass migration to Linux via the desktop is highly unlikely. Why? Because people don't want to have to chose between two desktop managers, between ten different word processors, between x different flavors of y.

This is not an indictment against freedom to choose! But it's been my observation that most people (especially the tech-unsavvy) don't want to have to choose if at all possible. They want one desktop, one word processor, one of y.

For Linux to break the M$ stronghold, distros will need to provide two things: (1) A "simple" install which provides the typical user with the minimum (ideal: zero) number of installation options, and (2) an "expert" install option for those of us who want to tweak our systems to the nth degree and not use an install process aimed at the LCD of the population.

Distro vendors themselves will need to agree on what a "simple" install is comprised of...and use the same components. Otherwise, we're back to square one on the fragmentation issue. Developers can make this process easier by putting aside their petty disagreements and pooling their energies to make production-quality software a reality, rather than the endless stream of beta-version software that never seems to quite make the jump to release-quality.

Device drivers...

[mjh]

Developers must be persuaded to provide Linux drivers, especially for "Winmodems," and to port their software products to Linux.

I think it's a bit more complicated than that. Developers don't have a way of providing a universal device driver that will work under any release of a kernel. Heck, a device driver for 2.4.10 won't easily work in 2.4.17! Exactly how is a device manufacturer going to release a driver (either open source or binary) that an end user can *easily* install? As it is right now, device manufacturers who support Linux have to do so with little added expense. Mostly because most of the people using Linux are technically adept enough to get their devices drivers working. But if Linux gets more popular on the desktop, the cost to device manufacturers of supporting Linux is going to dramatically increase as end users aren't able to install their device driver by themselves. I think this is going to be a limiting factor on Linux's popularity.

Until a device manufacturer can easily install their device driver in to just about any running linux kernel, I don't see them jumping on board to provide linux drivers. Until that happens, I don't see linux making much headway on the desktop.

I don't like this. I run debian on every computer I own. I'd really like to see Linux become popular on the desktop, but I think it has to overcome many hurdles. One of which is easily allowing device manufacturers to install their drivers.

$.02.

Re:The Key to Linux Success...

[clontzman]

StarOffice/OpenOffice: they need to iron out the last few bugs and market it, for crying out loud!

I don't disagree with you, but the problem is... who spends money to market a free consumer/business product that they've already spent a ton of moolah developing? I'm sure it's hard enough for companies to justify spending money to continue developing products they're just going to give away (or sell for little more than the cost of media and docs), but to have to drop money on *advertising* them as well starts seeming like throwing good money after... other money.

At the same time, remember the IBM OS/2 Fiesta Bowl sponsorship? Different situation, but not totally. What a disaster of a promotion, although history seems to show that IBM couldn't market a cure for death.

Re:Linux on the desktop...

[aardvarkjoe]

Well, there's an idea. If we got everyone bitching about linux usability on slashdot to actually try using it, maybe they would stop.

Re:Whose desktop are we talking about?

[Lumpy]

so you didnt give him a real distro then? I'm confused as to why did you give your dad a non newbie linux box?

my mom.. I email her a file... she copies it to her KDE desktop and double clicks on it. it asks for the administrator password and it is done.

I walked her through that once, and now she does it on her own...

What kind of alpha-ware are you making your dad install that isn't available as a rpm or easy to install binary package? I've eve seen a couple of apps now bail on rpm and use the Loki installer now...

please get him off of slackware, an advanced linux distro and give a newbie the braindead distro... redhat. it works great and is easy.

Re:Grrr! Stupid mouse remarks!

[sir99]

Read it again.
TiBook. TiBook. TiBook.
That would be a laptop, which implies a built-in mouse. He wants a laptop with a built-in three button mouse, instead of buying an external mouse and dragging it around with him.

Re:Ugh, more garbage?

[Lumpy]

actually the apps are there. they just need fine tuning.

the number one important thing that linux needs is a decent installer. Loki gave us one in their final death throwes.. it's awesome. and the like of KDE,gnome and EVERY app should drop what they are doing and start adapting it to their app.

To hell with making some minor bugfixes this week, get an installer on your app that even a lobotimized monkey can use.... that's the loki installer.

I've said this before...

[Second_Derivative]

The problem with drivers is not to do with arrogance of hardware manufacturers. Dare I say it's exactly the opposite. The problem with drivers as opposed to programs under Linux is that programs have something that drivers don't: a reliable ABI. Granted one usually needs to static-link against glibc, however if you import libc symbols and libX11 symbols your ELF binary's pretty much guaranteed to work everywhere.

Enter the kernel. Here there's no such thing. Just try running a 2.2 module under a 2.4 kernel, for instance. People tend to upgrade their kernel just like they upgrade any other part of their computer, and the interface is a real moving target.

A hardware manufacturer has a tough time when he wants to distribute a binary-only driver. They'll need to recompile and twiddle their driver every time a kernel release comes out, which is becoming more and more frequent these days. Binary drivers also seem to be viciously resented [debianplanet.org] by the open-source community; if you refuse to release your hardware specs you're an information fascist. It's so bad that we've now got this kernel license mess, where some modules refuse to let you link against them if you're not GPL. Read the f***ing GPL: there's an exception against GPL 'infecting' other code from within core services like the os kernel. Even Microsoft, evil as they are, doesnt mind what license your software running on their system is, so long as you don't touch theirs.

A popular approach these days seems to provide a GPL'ed 'resource manager' module which exports a stable ABI for accessing that device. The proprietary bits are then packaged up into a library or XFree86v4 module (which by the way is a much more stable proposition, with binaries being not only stable but crossplatform on x86. Linux is just as low-level and really ought to take a page out of their book). However this is no better than Microsoft DRM: it is code written for political, not technical reasons and it really does not need to be there.

Face it, Linux the kernel will succeed not when the scheduler runs like greased mercury, not when they finally decide which damn VM to use, but when they (a) provide a stable kernel services interface and (b) give hardware vendors a choice about how they can let you use their hardware (what if they've got good intentions but have licensed technology from someone who made them sign an NDA?). Isn't there a bit in the GNU manifesto about Free Software being there to promote choice for the user?

Re:Linux On The Desktop Is Viable Today

[hacker]

So I don't know that it's not ready, except for thoses who don't understand or are against change. I agree that it is not quite where windows is at, after all these years, but don't throw it away either. Many offices could readily change and have the tools they need using Linux, and gain the stability and speed we come to love.

The single common denominator I've seen so far is that all Windows users switching to Linux, expect Linux to _BE_ Windows. They want to right-click on the desktop and get "Properties", and they want a "Start->Run" paradigm. They try to "de-configure" the Linux machine to live and breathe like their previous Windows environment, instead of learning why Linux _EXCELS_ past Windows, and exceeds where Windows fails, they just want Windows.. on Linux.

People who are too lazy to learn a new environment, are not going to be users you want helping to contribute to the advance of Linux in general.

Linux requires work. Linux requires time. People need to understand there is no "Linux, Inc." that manages this. It advances at the speed of.. well, nothing. Whenever something needs to get done, it gets done... or doesn't.

Migrating users also need to understand that Linux _IS NOT FREE_. It costs money, lots of money in fact. Time, bandwidth, servers, payrolls, salaries, equipment. Just because something doesn't work, or "sucks", does not mean that it will get fixed. I see literally _THOUSANDS_ of people complaining about Linux problems. When I ask them if they have reported the issue, they say "No, I'll just wait until it's fixed". _THIS_ is the real problem with the "professional" quality of Linux. We have talented programmers, documenters, packagers. We just don't have talented users that provide _USEFUL_ feedback so we can improve the software we write every day.

Linux is ready for the desktop, and has been for years. Are migrating desktop users willing to learn how to use Linux on the desktop? Not yet.

Re:Well, what's the DESKTOP killer app?

[walt-sjc]

I don't think it's realistic to target grandparents at this point. Business suffering under the constant threat of the BSA, hundreds of thousands of dollars a years in license costs, viruses / worms that shut down their networks IS a good target market. Government is another. Do YOU really want a large chunk of YOUR tax dollars going to line Bill G's pockets?

If we took only 1% of the money that governments / business spend on MS licenses and used that to fund open source software, we could offer a viable replacement to MS in 1 year.

Of course there are reasons.

[FallLine]

All those tools need to agree on a very exact format specs or else the user cannot use the tools together.

All those tools need to operate in roughly the same way or else the user will have to learn a lot more.

All those tool developers need to remember that 99% of their market is for that subset of features that the dedicated application needs. Thus, adding much more will bloat it with complexity and size that simply does not reward the user.

All those tool developers all need to setup their applications with the majority of the users tasks in mind so they don't force the users to do more work than is necessary.

All those tool developers should provide a certain amount of interoperability besides just file formating and such. e.g., How does the user perform an "undo" after one tool has been applied?

All those tools need to agree to collaborate on support problems rather than pointing fingers at other tools.

The point is that creating seperate tools in this fashion is simply not appropriate for most applications. The organization and development costs for this "tool" methodology to make it appropriate for the end users totally exceeds the costs to produce a superior application under the "traditional" unified application framework. The analogy that I'd make, in response to the "tool box" analogy, is what tool do most users, that actually use tools, actually carry with them? A leatherman (and maybe a limited toolbox at home). The toolbox is too bulky and ackward in most situations where a leatherman (or like tool) is totally appropriate.

What you are doing is laughing at the Swiss Army knife that is MS and kin that tries to be everything to all people and assuming that the toolbox is the best solution because the swiss army knife is almost useless. Well it's not impossible to devise a better unified tool than both for most users. Its name is the Leatherman ;) While there is still plenty of room for the toolbox, its use is largely confined to professionals and enthusiasts that require a high degree of specialization.

Oh, man...

[SlashChick]

This article actually brought a smile to my face...

"Quite a few distributions of the Linux desktop are close to becoming products that can successfully compete against Microsoft Windows."

Translation: It's still not there yet.

"Each system can be installed without harming Windows."

Indeed, that's the first step. The second step is to automatically transfer / map "My Documents", "Favorites", "Fonts", etc. I haven't yet seen a distribution that will willingly copy over files from Windows, but Windows XP will willingly copy files and settings from any other Windows computer via Ethernet. Linux needs this to have a successful dual-boot audience, and it would be nice for system upgrades as well.

"With closed-source systems, users are stuck with programs and upgrades they cannot change."

Who says? I regularly contribute my feedback and bugs to everyone from Microsoft to MySQL to Trillian. I pay for the products, and I send in every bug report / feature request I find. In most cases, I don't want to program it myself anyway. If many people request a feature, it will be there. And often the programmers come up with a more intuitive way to impement it than I would have. I'm okay with this, and so are the majority of users.

"The Microsoft approach limits a user to available software. With Linux, a user can grow."

This makes no sense. There are development tools aplenty for both Windows and Linux. If your company uses Windows, chances are high that someone, somewhere, has an MSDN subscription and has the suite of Microsoft's visual development tools that they would be willing to let you borrow. Of course, you can also use third-party development tools (some of which are free) for both OSes.

This article should never have made it to ZDNet. Sometimes I wonder whether ZDNet scans article submissions for "Linux" and just posts those, knowing it will generate heated debate. Ths article is really flamebait -- it says nothing new, and it makes both sides come up in arms. Too bad. *sigh*

Re:Universal File Formats

[digidave]

I asked a writer to send me a word doc as rtf and even included instruction. She was going to be paid $250 for the story, but wouldn't bother saving it as rtf, so we didn't run her story and she didn't get paid.

I think the real problem is that most people are way too lazy to learn even a slight variance in what they do, let alone a huge one like change their OS:

luser: Where's Word?
bofh: KOffice.. right there on the toolbar.
luser: But it doesn't say 'Word'
bofh: It's the same thing.
luser: I like Word.
* bofh renames link to say "Microsoft Word" *
luser: Thanks!

Re:Build a tool ...

[ethereal]

Not that I'm a big Microsoft defender, but I'll point out that this is exactly how Microsoft Office works - a bunch of automation-capable modules with some glue code in between. So it's possible to write your own programs that embed Word documents, etc., the same way that you would embed the IE web browser component. When you say "ordinary users should be able to to this", think of management types building apps in Visual Basic, and consider whether that's really what you want or not :)

If this is really the direction that Linux apps need to go in, then it may turn out that Mono or dotGNU will turn out to have been some of the better things to happen to Linux. Not because of the language interoperation (which I'm not convinced will really work right), but because of the standardization on common objects and interfaces that will be required.

Gah, I'm defending Microsoft Office and .NET (but not Microsoft itself, at least) in the same post. Somebody shoot me now, please.

Linux on the Desktop...

[Fizzlewhiff]

The leading distros are improving in leaps and bounds for making Linux easy to install. There's still a few things that I think are lacking after the install but we are starting to see improvements in this area thanks to Ximian's Red Carpet and Red Hat's Up2Date.

I'd personally like to see a facility to make it easy to install something you might have missed during the original install. For example, Joe user installs Linux and when its all over and done with he wishes he could connect to a file share on his Windows box. He remembers seeing something about Windows connectivity during his install but doesn't know how to get back to that dialog or what the package was even called. His choice is to either reinstall or go to a newsgroup and ask for help, which leads me to my main point.

I think the linux community needs to lighten up when it comes to "newbies". Linux users should think of themselves as evangelists and when a new user asks a question not be so quick to flame them for not reading the HOWTO before coming to them with such a trivial question. If you go to your local church and ask an elder or a member of the clergy a question about somthing that has you confused do they jump up your ass for not reading the bible first for the answers? No, they are happy to see that you are interested and they try hard to help you. Why can't we be the same when someone approaches us with a question about Linux, no matter how trivial it may be? I'm not saying we should be there to answer all their questions, but in the process of answering their first questions we might want to show them where to find the answers so the next time they can help themselves. We just need to be more tactful when educating new users.

The distributions are doing a good job, the developers are doing a good job, now it is time for the users to do a good job. If Linux is to succeed on the desktop it is up to the users to give it a good image.

Re:Linux on the desktop...

[moongha]

As a person who has installed Linux on a Mac, I find the above post ludicrous. Infact, why has it been modded so highly when the majority of people have no idea what the actual experience of installing Linux on Mac is like?

Comparing the 'useability' of Linux to the MacOS is laughable. It really is. Why do you think people pay a premium for Mac hardware? For the performance? It's for the UI, and the UI alone.

So I don't think I'm being particularly unkind to Linux to suggest that it has a long, long way to go to be comfortable at all to the average Mac (or Windows for that matter) user.

Remember, one of the first stages in solving a problem is accepting that it exists. Denying it to everyone you meet isn't going to make it go away.

Re:Universal File Formats - one solution

[walt-sjc]

This is what I did. I setup a Windows box that acts as a document converter. Incomming emails are scanned for .doc's by procmail which sends off the word doc to the windows machine. A VB script takes that file, opens it in word, and causes word to save it as RTF. The rtf is sent back to procmail which adds it back as a second attachement. So now each email has both the original .doc file, and an rtf version (you want to keep the doc file for various reasons (sometimes you lose info.)

I also setup drop-box directories for employees to put old word docs and a vb script generates an RTF version.

You can do the same with other "common" proprietary file formats. We also have a few windows boxes setup that can be accessed via VNC to run various legacy / proprietary apps (I thought about writting a proxy that finds the next "free" machine automatically."

While this doesn't totally eliminate windows, it cuts it way down. The document converter alone eliminates 95% of the reason to use Windows.

For people with a larger need for Windows, VMWare can be useful.

When Micro$oft is no longer FREE, Linux prevails..

[ausoleil]

As it stands now, it is incredibly easy to copy and install any Microsoft application on the desktop, with or without licensing. Products Keys are traded and generated without compunction, and are bandied about in web-sites, on the Usenet, between friends, stolen from work and so forth and so on. Further, the BSA estimates that 1 on 4 OS installations on new machines are pirated. So it is easy to copy and install any Microsoft application for free, up to and including Windows XP. Quite many home users have no software other than the included OS and apps installed by the vendor that are legal. Think: how many folks go and buy full-blown Office licenses for their homes? I'll wager my Linux server that it is far less than 50%.

That's what is competing against Linux on the desktop: freely available no cost Microsoft OS's and applications.

That makes the killer app for Linux desktop success as simple as pie: real licensing from Microsoft that requires real product activation. When it happens, Linux is suddenly very viable as a competitor -- people will be REQUIRED for their versions of Windows XP-Pro-Gold-2004 and Microsoft Word 2004, etc. et al.

Why Linux will never succeed on the desktop

[Anonymous Coward]

On very simple reason: consistency. You MUST HAVE CONSISTENCY in order to make your interface "granny-safe". The average user is not a software developer with way too much time on his hands to fool around with and learn every single different weird way each app does Thing X, or twiddle endlessly with window manager file settings.

Open source and community development is a wonderful thing; I'm all for it. But it has one terrible downside: getting developers to all work (for free no less) in the same direction is like herding cats.

As a result, Linux is pandemonium. A world of choices and options and little settings you can hack in. That's great. But for interface design, it is ONE OF THE WORST THINGS YOU COULD POSSIBLY HAVE.

Maybe the Linux community has their helicopter beanies screwed on to tight. I dunno. But it is a fundamental law of interface design that it should be CONSISTENT. And not just every app using the same widgets: interfaces are about application interoperatability.

Here's where KDE and GNOME just fall flat on their faces. Even basic operations like cut and paste have inconsistent procedures and inconsistent results across multiple applications. And important ones, too! (take Emacs vs. Netscape, say). Drag and drop is nearly nonexistent, due to every library having its own drag-and-drop API, or none at all. There is no consistent interface for select-and-yank other than raw ASCII text (heavens). There is no consistent launching procedure. There is no consistent application installation procedure. There is no consistent location for files.

And as for widgets, now we have some real fun. I count no less than eight major widget sets in common use in X, each of which has its own unique operation rules. Font management is poor, bizarre, and inconsistent across different apps. Nearly every library has its own totally different way of handling internationalization.

And then we get to the jillions of "options" thanks to long-time X users mired in ancient ways of doing things. Apps must still offer focus-follows-mouse, despite YEARS of studies showing to be a hideous interface procedure. The #1 interoperability feature is still select-and-yank, a hack worked into early versions of X back when all it was doing is supporting multiple terminal windows. There's only one selection buffer, so every app has to come up with a way of maintaining selection even when you select something else in another window -- and they all do it differently. Hundreds of window managers and themes make support and management a nightmare. Demand by All Those Developer Users has prevented X from dumping lots of its old stupidities (like its graphics model) in favor of, you know, models that aren't twenty years old.

I find it ASTONISHINGLY TELLING that Linux users trying to fix the interface consistency nightmare THEMSELVES cannot work together. Hence we have no less than two major, incompatable (!) camps each dedicated to interface improvements: KDE and GNOME. That alone tells me X will never succeed on the desktop.

Say what you want, the reason MacOS and WinXP have such good interfaces compared to Linux is simple: they have hegemons who dictate that This Is How Things Will Be. Since everyone follows those rules, their interfaces are consistent, directed, backwards- and forwards-compatible, and modern. Without a bully, the thousands of voices all go their own ways, and you wind up with an interface by committee.

Linux itself succeeded because it has a bully: Linus. And that's a good thing. X windows will fail exactly because there is no one to fill Linus's shoes on the desktop.

Re:Linux on the desktop...

[Cheetahfeathers]

I recently bought a shiny new G4 with OS X. For many things it is very good, and I love the menu bar at top of the screen thing. But there are many ways in which the UI of my Linux and Solaris boxen with Window Maker is massively better for me.

The corners of the Mac are ignored, save for some lame option about screen savers. The buttons of the tile bar on the Mac OS X are close togother, rather that on opposite sides, making that part difficult. The cut and paste on the Mac is rather difficult, being a combination of mouse and keyboard, rather than pure mouse use.

The few times so far I booted into OS 9, it's been pretty bad (for reasons other than the UI), so I haven't learned much about that. :P

Compared to the Windows box that I use for games, though, the UI on the Mac is wonderfully advanced.

Re:Whose desktop are we talking about?

[moongha]

Sure, I agree, computers are not easy.

However, some systems (eg. MacOS X) do a much better job of making things easy (on top of a FreeBSD-a-like) than Linux does. Why shouldn't Linux aspire to this?

The problem is that its greatest strength (its open nature) is also its greatest weakness. It has no unifying direction where it comes to UI, like it does at the core OS level.

The look and feel of OS X was devised by a small group of people, and then implemented by a number of professional programmers. Unfortunately, on Linux, everyone wants to be a creative genius when it comes to UI, and there is no unified look and feel.

lack of standardised UI = poor ease of use.

Re:Windows market disappears

[Time Doctor]

Just because it seems far-fetched now, doesn't mean Windows emulation is a good idea. That's like saying the afterlife is so cool, so lets all kill ourselves, or something.

Re:Whose desktop are we talking about?

[Geeky]

I find it easier to talk people through command line stuff than GUI stuff. Far easier to say "Type this... OK, what does the screen say now?" than "click on that. Has a dialog box appeared? OK what does it say? Click OK. Now what does the screen look like?"

Talking people through GUI use when you can't see their screen is much harder than talking them through CLI commands. Especially if you can't remember the exact layout of every dialog that they might encounter.

Re:Universal File Formats

[radish]

- You might someday get a customer, consultant, partner, etc. that needs interoperability

We have 25000 desktops. Not very practical to change them all "in case one day an individual client needs support". If that happens (and in our business it's extremely unlikely) we can sort something out just for them.

- You will protect yourself from macro viruses

Outlook - now that gives us virus headaches. Office - not really. Virus scanners at the perimeter do a pretty good job, likewise document quarantine. I'd like to see something done about outlook, but if we tried to replace it with a plain & simple mail client there would be uproar from the users and the whole of IT would probably be fired! They're hooked on the groupware stuff.

- If Microsoft takes office in a direction you don't like, you won't be tied to it.

If that happens, we can change then. Right now I don't see a good reason to.

Don't get me wrong, I'm no MS fanboy, and if we had 25000 blank machines waiting for software there would be a strong argument for installing open stuff. But the cost of conversion is massive, and immediate benefits minimal. Please note I'm just a lowly developer, I don't get to make these decisions :)

Poorly Written Article

[sterno]

The title of the article and the contents seem to be unrelated. The title says, "Linux will prevail," then goes on to says:

"Unfortunately, many computer users are unaware of the extent to which they are "jerked around" by companies that sell a license restricting the freedom to use their software."

Well, it would seem that Linux doesn't stand much of a chance if people continue to remain unaware of that issue. He's not suggesting here that they are becoming more aware or ways that they can be made more aware, he's simply stating a set back to the cause. So hardly a proof that Linux will prevail. Next he says:

"Recent announcements by Sun Microsystems, regarding its expanded support for the open-source community and its decision to provide its own Linux distribution, are welcome news. "

Excellent! So all of the Solaris desktop users may move to Linux. I'm sure we welcome all 3 of them to our happy community. Next he goes on to list hurdles that Linux needs to overcome but doesn't provide any evidence that they ARE being overcome which is somewhat important if he's trying to proove his title. So then he moves on to say:

"No one would buy a car with a welded-shut hood, yet we continue to buy software that way. The Microsoft approach limits a user to available software. With Linux, a user can grow. If a tool is missing or awkward, someone can get under the hood and fix the problem. "

The funny thing is that increasingly, especially amongst the more expensive cars, it is becoming impossible to do any real work on them yourself. Sure, you can change the oil and other fluids but beyond that many cars are impossible for the average person to do work on. Finally, he says:

"

Two paths are before us. One leads to increasing proprietary control, protectionist measures and legal threats, while the other leads to open source, freedom and accelerated innovation. I, of course, choose the latter because it is "win-win." Vital innovation, new markets and vastly improved customer service win the vote readily over the purveyors of proprietary hoarding. "

To summarize, he seems to be concluding that Linux will prevail despite some hurdles because it would be really good if it did and really bad if it didn't. Wow, that's all the evidence I needed, kudos to linux, your victory is well in hand!

This is really a poorly written article and is little more than another puff piece about how Linux is the right choice, and windows is the wrong choice. It shows no new insights on the chances of Linux surviving and only points out the same issues to be dealt with that only about a billion other articles have pointed out.

it's already there

[Anonymous Coward]

depending on the distro, installing linux is often easier and less painless than on windows - here are some examples that would make a newbie throw his machine out with windows:

moving the ethernet card from the top bay to the bottom makes windows want to try and reinstall it! browsing to windows or windows\system trick doesn't work, it insists on the drivers yet again, and after a reboot, guess what?? no go, because there are now two similar adapters in the network control panel (this is 98SE, btw), each with differing tcp/ip properties. Easily fixed for me, hours of frustration for the avarage user.

When the "windows" magic breaks, there is no easy config file to alter, as is the case with linux, it's a matter of uninstalling, rebooting, reinstalling - fail, uninstall DEEPER, like tcp/ip itselt vs just the adapter, reinstall etc. This is a cycle of shit, excuse my english.

Or how about when a program installs and near the end of the install, asks you if you want to reboot to finish - windows crashes during the reboot and guess what??? the app is half installed and requires registry pecking to fix. If windows were a person, that person would have cancer yet be asked to run marathons - too undependable even for it's own operation.

Or here's another one - unplug an external mouse from a laptop with limited resources, reboot expecting to use the keyboard mouse. Windows rearranges irq's, and spits out messages about PCI bridges and the like - all stuff that would make the avarage secretary scream - and finds the keyboard mouse (did it forget it, lol???), but asks for the drivers. The solution here is simply to keep hitting next with nothing selected. But the masses would surely call their IT department on this, probably after trying for an hour to find the drivers...oh, and an normaly, this has to happen twice - assuming windows can manage to shut the machine down properly - upon reboot it does the same thing.

I guess the point is that in reality, i really don't think that windows is any closer to being as easy to operate as a tv or vcr than linux is, and, let me honestly say, if something does go fubar, it's much easier to have the user change a config file and change without rebooting, than the endless change, reboot, try again cycle, let alone mess with that God damn registry...

My 2 cents of truth from the frontlines of support,

Re:Well, what's the DESKTOP killer app?

[neuroticia]

Even WITH a reason you're not going to see people switching. I swear, my mother thought it was a violation of her warrantee to upgrade from Windows 95 to Windows 98. I had to remind her that she had a one year warrantee and that the computer was about 3 years old at that point. "Oh yeah." Says she. "But why do I need to change?"

I'm a geek, I like to think I'm a good geek, but I can't even bring myself to think about recommending Linux to someone who doesn't know what they're doing, or at least have someone who does living with them. It's hard enough to explain that when scandisk pops up they're supposed to leave it alone and let it do what it wants and just say "ok" to everything. But try explaining to them how to e2fsck /dev/hdc6. Forgettaboutit.

I've seen a quote floating around on here. Something along the lines of Unix being user-friendly but selective about who it makes friends with. I think it's going to be that way with Linux for a little more time.

-Sara

Installing OS == Rebuilding Transmission

[nakhla]

Think of your computer being like your car. Sure, there are some drivers who want to take it a step further. They might tweak their engines, swap out components, etc. all in order to make their cars the best they can be. However, most people are happy with their car the way it is. When you buy a car, everything is there for you. Most drivers don't want or need to know about the innerworkings of a car, let alone what to do when something fails.

We can apply this to OS's by comparing the OS to your transmission. I know ABSOLUTELY nothing about cars. That's why the transmission in my car is fine with me. If something goes wrong, I take it to the shop. If it completely dies on me, I buy a new car. I don't have the expertise - nor do I want to learn how - to rebuild my transmission. The average computer user doesn't want to worry about their OS. It's just supposed to be there and work. Installing a new OS is like rebuilding a transmission, and the average end user doesn't want to do it.

That is why Linux cannot succeed on the desktop until several events happen. 1) Linux must be installed at the OEM level. Computer have to come with it preinstalled. 2) The GUI has to be completely object oriented, easy to use, and easy to configure. News flash: XFree86 is NONE OF THE ABOVE. Look at how Apple took BSD to the masses. They didn't try and build an interface for X Windows. They built one ON TOP of BSD. That's what Linux must do. We can't rely on X Windows because it has too many shortcomings. 3) Applications 4) Unique features and enhancements not found on any other platform. As it stands right now, very little is innovative within the Linux community. Sure, the way things are done might be innovative. But, it all boils down to the Linux community trying to duplicate the things that Microsoft and Apple have already done. If Linux is truly to succeed, there needs to be some reason for users to switch.

Re:Universal File Formats - one solution

[Deven]

Have you made the code (procmail & VB script) available for this? It might help a LOT of other people...

Don't forget OSX

[pjt48108]

I was really pulling for Linux for a while, but then OSX came out, and now I think that Apple is leading the charge of *nix on the desktop.

Flame away. I can take it. But as eccentric as he is, Steve Jobs (and Apple) has once again seen the future and saved us from it.

Linux desktops sucks

[DapperDan]

It is sad to see so many people making excuses for Linux (note: any reference to Linux == desktop) and telling people what features they *really need* and what ones they don't. The unfortunate fact is that Windows (& Mac) are far superior to anything Linux has to offer.

Remember that most people do not care about the licensing issues other than free as in no money.

Now to get rid of the anti M$ feelings lets say Kde has all of Window's features etc, and Gnome has all of Linux's (remember were just talking desktop). Could any of you really step back and look at the two objectively and honestly say that Gnome was anywhere near Kde ? Even in the same ballpark? Are they even playing the same sport?

Linux has a loong way to go to be anything more than a niche desktop product. There are so many simple little things that it still can't do. For example, I was doing some work on a default Red Hat 7.2 install in KDE. I was browsing with Konqueror and went to copy a url that was in the text to paste it to another Konqueror window. Well it no workie. Now I know that if I run Mozilla it would have worked, but the problem is I shouldn't have to know that. It should *just work*. There are a lot of well-if-you-would-just-use-xprogram-or-run-yscript -you-could-get-it-to-work-without-too-many-problem s "solutions" to things that should *just work*.

If Office XP ran on Linux *today*, Linux would gain some serious ground but it would still be quite a way behind. I truly wish that were not the case, but it is.

/pulls out a stick and some marshmallows.

Re:Well, what's the DESKTOP killer app?

[irony nazi]

I'd switch OS's just to get the pagination to work!

Maybe this was your point, but I can't even get pagination to work under the same OS. Just when you have the pages properly, then change printers and blamo --- everything's screwed up again. Have somebody open your file on a different computer (but same printer), then blamo -- pagination is screwed up again.

Having good pagination in any MS-Office ap is akin to an unstable critical point in an ODE/PDE.

The funny thing is, that with LaTeX and GnuPlot/Matlab, I've NEVER had to worry about pagination. Occasionally I would make the margins narrower in order to make my advisor happy, but that's it. In my current job, I waste ~30 minutes per day on formatting related issues. Sure I've learned how to quickly fix formatting mistakes and reprint, but I rather waste those braincells on arbitrage pricing theory (the content of the reports) or drugs (just kidding).

Business desktop or home desktop?

[Kjella]

From time to time it appears people think they're one and the same, but for linux, they're not. I think Linux should focus on taking the business desktop, for several reasons. The problems with the home desktop are:

1. Free vs. free. Let's face it, that's the way it is for most home users. Either that or it's a sunk cost from when they purchased their machine, and people don't mind violating copyright. Unlike companies BSA is unlikely to pay a visit to them too.

2. Rapidly changing interface, particularly in the graphics area (DirectX, OpenGL). The interface for business application changes far less often.

3. More legacy applications. Companies generally have more legacy data, which can be converted. Recreating the API for running the apps is considerable more complex and buggy.

4. Faster application turn-over. Most business applications are continous developments, while games are released, then left for a sequel. By the time Linux game comes out everybody's waiting for the sequel, while people would be interested in Linux Officepack 2 even if Windows Officepack 3 is out.

5. Fewer competent users. Having a bunch of Linux admins who work full-time with Linux is better than a bunch of home users, even with many powerusers. Of course they are there to work and not do Linux development, but qualified people identifying, analysing and working around problems (one way or the other) still helps more than "I click and it doesn't work".

Kjella

Re:Well, what's the DESKTOP killer app?

[Syks]

compatibility?? i have a housemate that changed from WindowsME to XP and most of his drivers had to be updated, and that caused loads of problems on our network...

right now.. actually people need Linux to be quite simpler.. and more user friendly..it seems that everyone is afraid of using it, thinking that it is too complicated to use..

thats what people told me when i wanted to start using linux..(about 4-5 months ago)..when i asked about a linux partition question.. my answers were "you shouldn't be using linux if you cant partition" "you shouldn't be using linux if youre a beginner"... its a good thing that i didn't listen to them, with the help of the linux community, IRC chats and stuff.. now i can move myself around...and i love linux...thanks...

Re:Grrr! Stupid mouse remarks!

[pressman]

The user shouldn't need to change they way they work to fit the hardware. If the hardware doesn't do the job the user needs its the hardware thats broken not the user.

Software should allow the user to modify the way that input devices and peripherals interact with the hardware.

I can make my Wacom tablet pretty much behave like a mouse if I want to and that is definitely not the original intent of that device. I really don't care for multi-button mice, so i only ever use the left button and still use keyboard modifiers. Let's say someone is completely used to using a multi-button trackball? If the machine only comes with a 5 button mouse and not a 3 button trackball, does that mean that the hardware is broken? No, it simply means that the user has to go out and purchase a new input device to better suit their needs.

Apple makes a one button mouse/trackpad because it suits the needs of the vast majority of their trarget market. Wintel machines ship with a multi-button mouse because it satisfies the needs of the vast majority of their target market. Linux and *NIX users seem to need mice with 2 to 5 buttons and sometimes a scroll wheel. Seems kind of strange that the people who need a GUI the least are the ones who are pickiest about mice.

Re:It's a long shot.

[Eloquence]

Compare this with installing Linux. Even setting up Linux to see NTFS drives is a pain,

Possibly (though some distros recognize both FAT and NTFS partitions automatically and put nifty icons on your KDE desktop). However, setting up Windows to see ext2 or ReiserFS drives is much more painful.

ls | mc

[Animats]

Anyone remember when the UNIX file listing utility "ls" was separated from the columnization utility "mc", so that you wrote "ls | mc" for a multi-column file listing? Now that was modular. And how long did that last?

The Games Myth

[Watts Martin]

I see this "the PC industry was built on games" line frequently, and with all due respect I think it's dubious at best. If you measure the personal computer revolution using applications as roadmarks.

• The first drive from mainframes to "non-programmer's programs" predates PCs of any sort: the IBM DisplayWriter, a word processing terminal.
• The programs that brought the first microcomputers out of the hacker realms? Certainly not Adventure and SubLogic's FS1 (the predecessor of Microsoft Flight Simulator): try WordStar and dBASE on CP/M and Visicalc on the Apple II. The first two were revolutionary in the sense of bringing mainframe-like functionality to cheaper hardware, and the third was revolutionary, period.
• The program--not interface paradigm--that really made the Macintosh? Aldus PageMaker, of course.
• The field that the Amiga dominated long after its demise? Video editing, particularly with Newtek hardware/software. Despite being sold as a game machine, this niche non-game market kept the Amiga not only alive but undead. For some years after Commodore went away, Newtek was still selling their own branded Amigas!

The "Windows PC" is largely carrying on the CP/M heritage. Games only sell machines to hardcore enthusiasts. For the majority of computer buyers, a range of applications sell the machine and games are just icing on the cake. (Games arguably sell video cards for PCs.)

The Linux gaming world is likely to always be like the Mac gaming world. It's there, but people clearly aren't going to the platform to play games that they can also play on Windows. They're going to be going to the platform for something they need to do and that, objectively or subjectively, is better on that platform.

The enthusiasts will come to Linux already (they already are, and most of them are on Slashdot). To get regular users as Linux desktop users for its own sake, appeal to their sense of need with something that done more elegantly, effectively or more easily on Linux than it is on other platforms. That's why Linux is doing well on the server side--and it's a major component of all things Macintosh.

Re:Whose desktop are we talking about?

[TheAJofOZ]

This itself is part of the problem. Everyone expects a very complex system to be EASY. Computers inherently are NOT easy!

From The Humane Interface by Jeff Raskin:

Complex tasks may require complex interfaces, but that is no excuse for complicating simple tasks.

There are many tasks that can be done on a computer which are inherently simple and the UI should reflect that.

I don't think however that you need to dumb-down the distro. Linux should do this, IMHO: On install, after you pick the install type (Workstation, Server, etc.), pick the install type (basic or advanced).

Again from The Humane Interface:

As a user of a complex system, yhou are neither a beginner nor an expert, and you cannot be placed on a single continuum between these two poles. You independently know or do not know each feature or each releated set of features that work similarly to one another.

Raskin then goes on to explain in further detail that the Beginner-Expert Dichotomy is false. He adds almost at the end of the section: a well-designed and humane interface does not have to be split into beginner and expert subsystems.

There is substantial proof that UNIX can be user friendly and it comes in the form of OS X. Despite what some other posters have said, OS X adheres to UNIX traditions extremely well (they changed /home to /Users). The only other major changes were using NetInfo for OS management instead of plain text files and the system init procedures. Both those changes were made for technical reasons (right or wrong) and are not used directly by the great majority of users.

The real problem that Linux has with being user friendly is that it is being created by people who are hopelessly unqualified to do user interface design (note that I fit this category) and that it has no standard for the way a user interface should look and feel. Now, the GNOME and KDE projects are making great headway and I'll bet that they've picked up a lot of talented designers over the years so that the brains trust is now adequate to solve the first problem. However, these people need to take the plunge and completely change the interface if it is required (a point echoed by Raskin).

The second problem will be extremely difficult to solve - it does not have to involve ridding the world of either KDE or GNOME but does involve treating them as largely different OS's that are compatible. There should be pressure on developers to develop both a KDE interface and a GNOME interface so that the user experience with either desktop is consistent. Another option would be to define a set of APIs that maps to both GNOME and KDE as native interface elements depending on which is currently in use.

This is the kind of thing that Linux has to do to be userfriendly and hence be successful on the desktop. If you think that these suggestions are unrealistic or impossible, you're saying that Linux can't make it on the desktop - I however disagree. Linux has achieved many impossible things before and in time I think it will achieve end to end user friendliness if the developers are serious about achieving that.

Re:Linux On The Desktop Is Viable Today

[ryantate]

Linux is ready for the desktop, and has been for years. Are migrating desktop users willing to learn how to use Linux on the desktop? Not yet.

Oh I see, it's the users' fault.

As soon as they stop acting so st00pid Linux will take over on the desktop.

People don't want to learn a new environment unless there's a concrete benefit to doing so. Not supporting "Right click/Properties" just to be different or for the convenience of developers makes Linux a little bit harder to learn. Add up a bunch of these subtle interface incongruities and your users will feel furstrated, upset and angry [joelonsoftware.com].

Linux brings more to the table than a re-arranged GUI with different colors and fonts. So why not try to match the familiar (Windows) interface on commodity OS elements (copy/cut/paste keystrokes, file property menu location and other common right-click behavior, basic File menu commands and Edit menu commands, even task bar/dock and start/apple menu) in as many cases as possible while adding subtle, unobtrusive improvements to interface components where Linux can excel (paticularly software upgrades, since most apps upgrade for free under linux and can almost always do so online, and doo dads like a graphical uptime monitor that highlight Linux' strengths).

Re:Whose desktop are we talking about?

[JohnsonWax]

Don't forget:

Mac OS X:
Drag application to Applications folder

You usually even get all localized versions. Anyone want to compare uninstalling?