In my last blog post about the real-world cost of Linux, I was struck by the contrast between my words and fellow InformationWeek blogger Alexander Wolfe's take on all this. He's lamented the broad variety of distributions out there, and found fault with the way those who create Linux distributions seem unwilling to accept what users really want. I figured I should lay out my own arguments.
First off, let me state my own position regarding Linux. On a day-to-day basis, I'm a casual Linux user. It hasn't eclipsed Windows on my personal desktop, and that's mostly because there's still a lot that I use Windows for in a work environment -- testing software and hardware, mainly. I like Windows for what it is, and I write tips for it regularly on my own blog, but I've also come to like Linux for what it is, too. Most of why I haven't moved to Linux wholesale is just practicality, but I remain curious and open to the possibility of eventually doing so. In my mind there's no contradiction between how I feel about both platforms, because I expect different things from them.
Now: What about Alex's two issues? First off, I don't think the broad number of Linux distributions out there is a bad thing, except in the sense that it makes it difficult to choose where to begin. This is definitely a problem, both for pros and beginners, but you could no more slim down the number of distributions out there than you could selectively turn off the stars. It makes more sense instead to provide people with a way to efficiently choose the right distro for the right job.
What we need is a better way to decide those things, and in fact I came across something that comes close called the Linux Distribution Chooser. However, I feel it starts off on entirely the wrong foot and asks too many questions to get the needed information. It's an area that badly needs more work.
Second is the perception of what the average user wants or needs. This can be a tricky issue to pick through for a bunch of reasons -- a major one being the pros tend to forget that less technically-savvy users are simply not as obsessed with the way the system works.
A regular user will say something like, "I'd like to deal with my email and browse the Web and write my novel." The exact way they get those things done isn't as important, as long as they can get them done in a relatively straightforward fashion. Windows and the Mac offer more deliberately closed-ended ways to get there, while Linux offers a plurality of ways to get there.
Again, choosing a distro is often the issue -- but, that said, the vast majority of distributions are so specialized that this is not as bad an issue as it might seem. It usually boils down to something like Ubuntu vs. Red Hat.
I do agree that there needs to be incarnations of Linux that are as comfortable as possible to use on the desktop -- something for Joe CD-ROM, as it were. But I don't believe the whole of Linux as we know it has to be remade to make that possible. Ubuntu has tackled the problem in a practical way, by building (and continuously improving) a desktop-friendly distribution that has drawn popular attention to Linux like little else before it. Sure, we could use more of that -- but the fact that we get it at all is something of a miracle, isn't it?