I first used Subversion about five years ago while writing my second book. It was an early milestone, but it was easier to compile, install, and configure than CVS was, and it was much more powerful than RCS. Since then, I’ve used Subversion to host every book I’ve written or edited. (See The Making of BSD Hacks.)
Though distributed version control systems are gaining in popularity (and though I use SVK atop Subversion), Subversion is still tremendously useful. Try being a F/OSS developer without knowing how to work with a project in Subversion, for example.
The preponderance of support with free hosting providers for F/OSS projects, as well as the quality of documentation and implementation, mean that I spend a lot of time working with Subversion. I’ve rarely had any trouble (even as far back as milestones 19 and 20). My work would have been much more difficult without this project, so thanks to everyone who’s contributed in any way.
We appreciate the appreciation! :-)
I agree with you that the distributed systems are gaining popularity, and they are better for some applications; the Subversion project is trying to learn from them and take features where it can. But I think our focus is likely to stay on being the best centralized version control system around, even as we absorb some distributed features (I'm not speaking for the other developers here, just myself).
I don't seem to be able to link directly to my comment on this post, but just search for my name there and you'll find a longer discussion of this question. (By the way, the "begrudingly" in Michael's post's title is wholly inaccurate, but to his credit he later realized that and made a followup post apologizing for the tone of the original.)
I eagerly await the release "Essential Distributed Version Control". :)