Torvalds On Desktop Linux's Slow Uptake

javipas notes a Wired piece summarizing a two-part interview with Linus Torvalds that's up at linux-foundation.org (part 1, part 2). In the second part the creator of the Linux kernel gives his view on the limited success of Linux on the desktop. "I have never, ever cared about really anything but the Linux desktop... The desktop is also the thing where people get really upset if something changes, so it's really hard to enter the desktop market because people are used to whatever they used before, mostly Windows... better is worse if it's different."

Comments

Re:Simple reason enough

[eneville]

People invest a lot of time learning a certain UI, the way it does things and the interface. For technical people like us, it's not that difficult to learn a new UI (since we have an appreciation of the underlying works). But for non-techies, learning a new UI (particularly one that makes as much use of the terminal/command line as most Linux distros do) can be a major hassle. It's just not worth it for most people, just for some nominal security benefits and to save the $100 or so that Windows adds to the typical computer.

but going from xp -> vista is also quite a "learning" investment.

Totally wrong

[Reality Master 101]

Why is this so hard for so many people to understand? The reason Linux doesn't get adopted has nothing to do with how the desktop works. I have news for you: Linux, Windows and the Mac are effectively identical when it comes to operating them.

I shout this from the rooftops everytime this comes up: PEOPLE USE APPLICATIONS, NOT OPERATING SYSTEMS.

Applications are EVERYTHING. Microsoft has long understood this. Why are people so upset at Vista? It's not because of the popups... it's because of the compatability problems. People want absolute, "it just works" compatability. People want to be able to walk into Best Buy, grab a box off the shelf (software OR hardware), and install it. No muss, no fuss. That's why the Mac has long had single digit adoption rates. People don't to figure anything out, they just want to buy a damn box and load it on.

Linux will be adopted with a) it has nearly perfect Windows compatibility, or b) the major companies start producing Linux version of their commercial software.

And yes, I understand that there are typically free versions of various commercial software. But again, people don't want to figure anything out. They want to know that if they see a box, it will work. If they buy that fancy computerized sewing machine (such as my mother-in-law), it will work.

Re:Totally wrong

[junglee_iitk]

Thanks for saying this.

I have korean project partners who are angry with me because I won't install Office 2007 and make them save it in some older format. They are even playing games, claiming that they are using some essential parts (read Microsoft Equation Editor) which they cannot convert to old format.

I have given up on explaining the morality behind not using pirated copies... I have given up the rational that Office 2007 adds no new mission crucial functionality. I simply say that I don't have a computer and I work in my office. I don't even tell them that in my office I use OpenOffice, on Linux.

People are not masters in this area. There is a very simple thinking behind all this:
1) Expensive is good, more expensive is better, even when it is pirated.
2) I use bla/I _like_ bla, can you do it? (Until we can do it by clicking here and there, we are argument-less in their eyes.

Re:People don't like change

[ProppaT]

People also don't like crappy UI's, programs with really absurd/dorky names that make no sense to anyone but nerds who get the inside joke (if there even is one), and O/S's that don't support their favorite software. Honestly, I'd say it's about 100x's more likely that OSX gains significant ground to the point where it makes sense for apple to source out OSX to third party system builders than it would that Linux gains any significant headground. You know, unless the Linux community understands and finally makes strides to make Linux a) look like a program you would actually go out and spend your hard earned money on and b) make the UI and naming convention on the included software logical.

Re:Simple reason enough

[gsslay]

But for non-techies, learning a new UI (particularly one that makes as much use of the terminal/command line as most Linux distros do) can be a major hassle. For non-techies, the UI is the computer. So if techies want to understand what an upheaval it can be; imagine learning a new operating system that works to three state bits, stores its configuration in jpegs, uses venn diagrams and tonal whistles instead of WIMP and communicates with hardware not by interrupts, but by a "alphabetical sort queue" principle.

Scared? Now you're getting the idea.

It's easy - just make it better

[edmicman]

One thing I don't understand is that there seems to be a consensus that Apple got it right, UI-wise. Unix underpinnings, but an elegant interface (from what I hear, anyway....I haven't used a Mac since ~1994). The knock against linux seems to be that the frameworks are there, it's just sort of a kludgey interface a lot of the time. "Too much command line needed". In my experience, things like Ubuntu have made it a lot better, but it still seems like a bastardized version of Windows. Sure somethings are prettier sometimes, but a lot of times other parts aren't remotely close. So my question is....

Why not rip off the other guys? Rather than chase Windows, chase freakin' OS X. If Apple can make a glamorous OS based on Unix, why can't anyone make a glamorous OS based on Linux? Is it because Apple has those magical UI fairies? FOSS vs commercial shouldn't matter - people are ultimately the ones that make the stuff. Are you telling me there are no more best and brightest out there working in the FOSS world, that they're all snatched up and locked down for commercial project?

I love a lot of the aspects of the Linux desktop, but it just seems like the vast majority of FOSS projects' tagline should be "almost as good as the commercial counterpart, but it's free!". IMHO there are only a few major projects that have actually *improved* on their commercial counterparts and made a *better* product. And those projects are the ones that succeed. For Linux On The Desktop to actually work, it needs to stop trying to be the "free alternative to Windows or Mac" and actually be a *better* alternative, for more reasons than just not having to pay for the software.

Re:Simple reason enough

[at_slashdot]

Sorry, but UI is a red herring, it's hardware compatibility and software availability (AKA "lock in") nothing else. KDE and Gnome are pretty much Windows like point an click interfaces.

Perspective

[Kenshin]

My experience (and probably many of yours) is starting with a computer from the Apple II, Atari, Commodore era. Wrote high school term papers on a typewriter. In college I did amber screen work and wrote papers with a dot matrix printer... The rise of the Linux desktop feels comfortable to me.

This middle-aged woman at the office listens to the "E-Z Rock" radio station. That's because it feels comfortable to her. She grew up with stuff like that.

Me? I turn that shit off the moment I get the chance.

Re:Linux on Desktop? Ha

[at_slashdot]

Linux is ready for desktop, hardware and software vendors are not ready for Linux. The are few reason beside hardware and software lock in for which people would not switch to Linux. "Oh My God, do you mean that I have to click only once!!!!11!!1!!"

Re:Simple reason enough

[Anonymous Coward]

Maybe if you're configuring hardware, or setting up firewall rules, but for the average user, I just don't buy it. You click on the start menu, select the program you want, and it works the same (besides the transparent window decoration). Just because it's a pain in the ass for the /. crowd to learn the new control panels doesn't mean there's a massive leap for most end-users...

Re:I disagrrree

[Anonymous Coward]

I agree. Linux on the desktop over the past 2 years has taken spectacular leaps forward, and the next couple of years are going to be just as bold as Linux starts maturing.

It's been a combination of several factors:
* The rise of FireFox and, to a lesser extent, Safari means that the web doesn't require, nor mean, Internet Explorer.
* The release of Vista and the negativity surrounding it has been key - people are pondering alternative OSes, including both Linux-based OSes and Mac OS X.
* The rise of Ubuntu as a 'standard' has helped solve the confusion of multiple competing desktops to new users and driven increased users. It's also improved support - UbuntuForums is a fantastic resource - and increasingly made GNOME as the "de facto" desktop environment.
* Improved driver support, which is going to keep improving. It's still far from perfect, sadly, but it is most definitely getting there - when Intel and ATI are both releasing open source specifications to get proper open source drivers written, it's a good sign.
* Eye-candy. It sounds silly, but people like eye-candy and Compiz Fusion delivers it.
* Vendor support. Big names like Dell are now taking steps towards Linux presumably as there is some demand. Hardware manufacturers are going to have to soon start touting Linux support for many OEMs to go onboard.

It's still not perfect, but neither is XP, Vista nor Mac OS X and I'm looking forward to a Linux-using future.

Re:People don't like change

[cp.tar]

People also don't like crappy UI's,

Luckily, KDE's Kickoff menu is lightyears ahead of Vista's Start menu, and Linux UIs in general are of pretty high quality.

programs with really absurd/dorky names that make no sense to anyone but nerds who get the inside joke (if there even is one),

Actually, they don't care much about names either way. As long as they can make the program do what they want it to with as little hassle as possible, they couldn't care less about its name.
Besides, KDE, for one, shows a short description of the program right in the menu, so you don't even have to memorize it.

and O/S's that don't support their favorite software.

Actually, it's the other way round: application vendors do not support certain operating systems.
There is little Linux people could do to support Photoshop, except create an emulation layer or something like that...

I'm truly fascinated with the way things are reversed in the computer world, and how natural it seems to most people... operating system developers should support applications, web designers should support browser rendering bugs... Get a grip on reality, will you, people?

Honestly, I'd say it's about 100x's more likely that OSX gains significant ground to the point where it makes sense for apple to source out OSX to third party system builders than it would that Linux gains any significant headground. You know, unless the Linux community understands and finally makes strides to make Linux a) look like a program you would actually go out and spend your hard earned money on and b) make the UI and naming convention on the included software logical.

I, for one, find a bit more logic in the Dapper Drake --> Edgy Eft --> Feisty Fawn progression then in the Panther --> Tiger --> Leopard one.
I'd even go so far to say that Windows seems to have the most inane naming policy, yet it still dominates the market.
Not that I find that naming really matters. At all.

As far as the way Linux looks — have you seen Compiz Fusion?
Do you know how many people not only considered, but actually started using Linux based on the Compiz bling factor alone?

And get this: you don't even have to spend your hard-earned money on it.
I can get you a pirate version really cheap. ;)

different bugs

[bcrowell]

This is the argument Linus makes in the article. I agree with it to some extent, but I also think the way he presents it is a little misleading. He makes it sound like Windows and Linux are just different, so there's absolutely nothing the Linux community can do to encourage adoption of Linux on the desktop -- it's all a matter of users' ingrained prejudices. But Windows and Linux aren't just different by design, they're also different in terms of their bugs. If you use Windows as your desktop, you encounter bugs. If you use Linux as your desktop, you encounter bugs. For instance, I've just spent half an hour this morning dealing with an issue in CUPS where every time I boot my Linux box, it starts spewing page after page of raw postscript. (Deleting the job from the queue didn't help. It just reappared the next time I booted the machine.) Well, this is a bug that I know about now, and I have workarounds for it. (Delete the printer and then reinstall it in CUPS's web interface.) Bugs in the Windows desktop aren't a strong motivation for Windows users to switch to Linux, because they're used to those bugs, never really think about them much. But if they were to try Linux, they'd say, "Oh my god, this OS is a total piece of crap. Look at the printer spewing page after page of garbage, and it starts again every time I reboot. This is pathetic. I'm sticking with Windows." They notice the Linux bugs more because they're unfamiliar and mysterious, and also when you switch OSes, you get hit with lots of these new and unfamiliar bugs all at once.

So it's not just a matter of user preference, and it's not something that's outside the control of developers in Linux's OSS ecosystem. The quality of the Linux desktop sucks -- sometimes I think it sucks almost as much as the quality of the Windows desktop -- and it needs to be improved. If that happens, it will increase adoption of Linux on the desktop.

Re:People don't like change

[Lemmy Caution]

None of those issues are the real obstacle to the mainstream adoption of Linux, at least not directly.

These are the main obstacles:

1. a lack of unity of experience. The unix way is great for system admins and people who like "a lot of little tools doing well-defined small things well". That is exactly what a desktop user, generally, doesn't want. An end-user is interested in their work, not the computer's work: they (and since I left IT, that includes me) want my components to integrate smoothly. This means an address book that intelligently talks to my mail client, which is aware of my calendar. It means not only that the menu navigations are both consistent across applications and let me do what I want to do with information that is only one or two clicks away from being revealed to me, but that default settings generally work and that any customizations I do are unlikely to be harmful.

2. "the Linux community" is not a unified development team. There is no final decision maker. There is no unifying vision. This make Linux a great place to a. learn stuff, b. experiment, c. scratch unusual and idiosyncratic itches. There are, of course, distributions that try to introduce more discipline and restraint, but then they run afoul of the fact that 3rd parties are developing for "Linux," not so much for this or that flavor of Ubuntu or what-have-you. In short, distros are small neighborhoods.

3. Advertised and guaranteed hardware support. I have a MacBook. While not every peripheral in the world supports Mac, I can look at the packaging of a peripheral and see a "Mac" logo on it, and buy it without breaking a sweat. In Linux, not only do I need to Google, but I should probably check SKUs, versions, warnings, etc.