Sourceforge.net, the premier repository for open-source software, has more than 160,000 projects registered. Many of them will never reach the 1.0 revision marker. But is that really a bad thing?
Here's how the argument is typically phrased: Look at the sheer number of open source projects that spring up, engender a flurry of activity, then die off. Doesn't that say something about open source as a development philosophy? I've heard variations on this argument for a long time, and at first I thought there was something to it. Then common sense asserted itself.
Well, here's my answer, admittedly a tad snotty: Look at the sheer number of businesses that spring up, engender a flurry of activity, then die off. Doesn't that say something about capitalism as an economic philosophy?
The fact that many open source projects never get past an alpha stage is no more an indictment of open source as a development philosophy than the fact that businesses fail is an indictment of capitalism.
Many criticisms of Linux and open source, I think, revolve around mistaken perceptions of what they're for and why they're used. Open source doesn't guarantee anything except that the source code for your project will be readily available to others for inspection or for re-use according to whatever licensing you're providing it under. It doesn't guarantee that your project will be a success. It doesn't assure you that even if you're working on the most uninspired, copycat project imaginable (oh, look! yet another CD burning tool!), you'll enjoy popularity and re-use in many places.
What it does guarantee is that if you're no longer able to continue work on the project, other people aren't prohibited from doing so. It also guarantees that the work you do is documented in public, so that concerns about security or efficiency can be checked independently by others. And if you dismiss open source software out of hand for having sparked any number of unfinished projects, then you're also forced to dismiss all the other things that are roaring successes -- MySQL, Apache, Linux itself, or even little programs like 7-Zip (which I use daily).
I'm encountering an analog of this in another field. Every November, a great many writers around the world participate in National Novel Writing Month, an informally moderated challenge to write a 50,000 word manuscript in 30 days. (I've been participating for two years now.) Some of the sneering directed at the competition has been stupefyingly mean-spirited, to wit: most of this writing is total junk; you people are never going to get published so stop kidding yourselves; there's already too much bad writing out there; and so on. To all this I say: So what? No one's forcing you to read any of it. The point is not to write a masterpiece and make a million bucks, anyway; the point is to write, and maybe spark someone who does have talent to act on it.
There are other possible criticisms of open source, such as whether or not open source development makes it impossible to develop closed-source software at all (something I'll talk about separately). But this particular canard strikes me as nothing more than a peculiarly embittered variety of sour grapes at best, and a complete misunderstanding at worst.