What is “The Linux Desktop” anyway?
In my last blog I mentioned there really isn’t such a thing as "The Linux Desktop". The topic was mentioned in a forum today so I thought I’d post a follow-up with some related thoughts and hallucinations – you can decide which is which. I apologize if this posting isn’t organized very well. I’m blasting this one out quickly and thoughts don’t always come in nicely organized paragraphs.
The Linux desktop is on about 2% of the world’s computers, and that’s with an estimated increase of 60-80% over last year. To anyone buying, selling, or using desktop applications on a regular basis I think it’s safe to say there isn’t much of a Linux desktop market now/yet. It’s a shame that Red Hat and other companies aren’t driving to make Linux a desirable desktop platform – that’s a dumb non-move on their part, IMO.
I think the Linux desktop is a good idea for anyone who just uses their computer for documents, email, and browsing. That’s a significant market and there is plenty of desktop software for Linux for these applications. I’m inclined to think that Linux gear-heads are spending a bit too much time thumbing their noses at Microsoft and not enough time Marketing what’s available to the average computer user / consumer. As with many tools of convenience, when people start using Linux at home they’ll probably want it in the office too.
Can people use it at home? Not really. While there are some apps available and a number bundled in each Linux distro, Linux is not a consumer-oriented product. Maybe this is the whole point and precisely what needs to be changed. Installing software on Linux requires the right underpinning libraries to be installed, and the right versions – you’re in trouble if you upgrade software X to load software Y and software Z relies on a prior release. Installing applications and the underpinning libraries is not trivial either. I’ll come back to this in a couple minutes.
Let’s assume it is used at home. Who can use it? Mom n Dad? The Kids? Grandma and Grandpa?
Younger people not only want the browser, they want their games, and game companies don’t normally target Linux. They target XBox, Playstation, PC if we’re lucky sometimes, and Linux is a distant last. So kids will be able to do some of their homework but outside of that the box is useless. And while Linux guys are fond of getting all the drivers they need for their new hardware, the application companies aren’t inclined to produce software that will use the devices. So what’s the point?
You may notice that when I talk about kids, I mention homework and browsing but not email. I dunno about you but I’m finding kids don’t do a lot of emailing, they IM/chat and send messages through MySpace and other social networks. Are we seeing the death of email in the eyes of the next generation? Well, at least there are a number of IM clients to choose from for Linux.
Grandma and Grandpa can use the Linux desktop because they probably won’t be installing software, and it’s nice to be able to email or IM with everyone else. Hey, maybe that’s a nice slogan for Linux desktop marketing : "Grandma and Grandpa just love it!"
For the rest of us in the middle, Linux just isn’t compelling enough. This comes back to marketing. I suppose what’s required is a comparison chart / site where people can enter the names of all of the applications they use in Windows, and after clicking Submit they get a list of suitable Linux replacements. The site can publish the hit% of apps found that match the requests, and when people see that something like 90%+ of the apps they want have an equivalent in Linux, we might see more people trying Linux.
There’s yet another side of this Linux desktop debate, and the real focus of this blog entry. What is "The Linux Desktop" anyway? There really isn’t such a thing. There are many X Window desktop environments but incompatibilities between them make it difficult for developers to support more than one, and just as difficult for consumers to pick one. The above service to show equivalent apps would need to show results for at least KDE and Gnome, maybe others too. Have a look at this Wikipedia page for some idea of the complications, especially down around the default programs packaged. Consider that not only are all of the programs for each manager different, but each one is going to have a specific set of features which it supports, lacks, and which don’t work because of some bug.
What? Bugs? Fie on ye for such blasphemy! Yes, just because software runs on Linux doesn’t make it perfect. It might be free of cost and free for all to look at the source, but none of that implies bug-free. So when we get our KDE apps and we find something we don’t like, do we switch over to Gnome? No, can’t really do that.
Minor digression / hallucination based on personal experience :
How "free" is Linux software when we’re so bound to specific releases, languages, and getting everything else "just right"? Hey, this chat program looks nice … oops, gotta find one that works with KDE … oops, that program on sourceforge hasn’t been updated in 3 years … oops, that requires glibc 2.3.6 and I’m running 2.3.2, dang, so close. Hey, I’ll upgrade glibc … oh oh, garbonzo v0.9 no longer compiles and I need that for platypus 0.4.6 which is required for printdweeb 4 and that’s required by sqlbuddy 2.5 and that’s required by my download manager. Bewm – why can’t I download anything anymore after I got this chat program? Gosh, I just spent 14 hours on it and I have all the source code, I should be able to make it work. Well, maybe I’ll just reinstall the system over the weekend…
Well, let’s just say it’s a good thing Grandma and Grandpa don’t need to change apps very often.
Ahem – the original point about the lack of a clear definition for "The Linux Desktop" was to point out that the Linux market is its own worst enemy. Once everyone agreed that Windows was the enemy they all ran to their quiet corners to write what they felt was good software. The result is 100 variants of what "good" means. Sure, diversity is good, may the fittest survive, and all the rest of that. But for years while Linux guys attempt to out-do one another, they’re all trying to duplicate and emulate what Windows already has, and they’re creating the same number of bugs, patches, and re-writes that we see in Windows. At some point a lot of these guys get turned off by the endless chase to keep up with the latest releases of libs and distros – and they’re replaced by more renegades who think they can do better – at least until they find a girlfriend, have kids, or get a job.
It seems to me the gear-heads just get in the way of their own future with perpetually competing desktops, tools like Qt vs Gtk, and a major lack of standards that allows software of varying releases to co-exist. C’mon kiddies, get out of the sandbox and play nice. If you want to know why The Linux Desktop isn’t here yet, look in the mirror. Microsoft may be the bastion of poor software and evil business practices to some, but the Linux market still looks like an immature gang of kids fighting with one another and throwing rocks at other people’s Windows just to blow off angst.
When Linux and its culture look stable compared to the alternatives, then we’ll start seeing more Linux on the Desktop. The first thing to do is define the desktop, get cross-desktop compatibility if and where required, and work on marketing that allays the fear and doubt that people have about losing what they have on their desktop already. You see, most people look at Windows and Apple, not so much as operating systems, but as desktop environments. The platform "is" the desktop. Linux started by being all OS and no desktop. For servers that’s fine but if you’re going to "have" a desktop you need to fight against platforms that to the common person "are" desktops. That’s a heck of an uphill battle and I don’t think the Linux market is prepared to fight that one or many of the others.
2 thoughts on “What is “The Linux Desktop” anyway?”
I complete agree with you Tony. But i think they will be a good corner for a desktop in Linux more than grandfathers. When i talk with TI people in mv sites, i always find comments as:It is imposible and very expensive to mantain all that windows boxes around the mv server.Actually, many many people still use in our market a Windows PC almost as and "Old green screen", with some data transfer to Excel, a little use of Word, e-mail, a web navigator and an EMULATOR with transfer capabilities and predefined keys.It should be very interesting to configure a Linux Desktop box with f.i.: Open Office Firefox Thunderbirdand…… the key, something like Accuterm or Wintegrate in Linux.With this, the effective cost of a seat in the mv environment will be much more cheap, with the same usability that the majority of the seats working now and with less problems.Who will write that Accu_WIntegLinux?
Good comments, Joseba, thanks. Pete Schellenbach is already working on AccuTerm for Linux. I know companies are cost-conscious and Windows can be a significant part of the budget. For some users who do not do anything but data entry, Linux is fine, and so is a thin-client appliance ($300US, 5 second boot, no software to maintain or mess up). I think average office workers are still too unfamiliar with Linux to put it on the desktop, so there is a training factor, but yes, that’s a good place to start getting people familiar with it too.
For guys like us, I don’t think the Windows desktop will go away even if we use a Linux desktop for common tasks. Clients send me VPN client software, Remote Desktop configurations, Word, Excel, and Visio documents, and from time to time I get other files and requests that require the use of Windows. I’ve been told by a couple of our Linux gear-head friends that they can’t open files that I’ve sent them. Well, I think they’re trying to make a statement about choice rather than capability, and hearing they can’t do something doesn’t endear me to their platform, it only keeps me away. But my point here is that companies that adopt Linux now should be prepared for some amount of "can’t do that" feedback. Personally I would really like to get a list of issues that people run into with Linux. I suspect the list is small but it would be good to have for anyone considering Linux desktop deployment.