Current location:	Mountain View, CA
Current mood:	accomplished
Entry tags:	bugzilla, tech

10 Year Anniversary of the First Bugzilla Code Checkin
So, today (or tomorrow, depending on your timezone) is the 10th anniversary of the first Bugzilla code checkin. That would be the first time that Bugzilla's code was ever put into Mozilla's CVS repository. We can count this as the first time that Bugzilla really "existed" as a publicly-accessible, open-source project, though we have an even more important anniversary coming up on Sept 19 (the 10th anniversary of our first open-source release, Bugzilla 2.0).

Ten years is an incredible milestone for an open-source project. There actually aren't that many open-source projects that can say they are ten years old, outside of BSD, GNU, Linux, and Mozilla itself.

I'm sure that justdave will write more about it, but I did have a few thoughts about it.

I think it's amazing that none of the original developers are still working on Bugzilla, but it's still been a successful project for all this time. I myself have only been around for the last four years. The oldest team member, justdave, has been around since 2000 or so. I think that's a great validation of one of the best advantages of open-source software--even if the maintainers disappear, if the software is valuable to people, it will continue to be maintained actively.

Bugzilla has had a fairly interesting history. The first release, 2.0 (in 1998), I don't think anybody used but Mozilla, and I think things remained pretty much that way until about 2.8 (in 1999), when we first got checksetup.pl, our installation script--a beast that many years later I would become responsible for. Probably a few brave souls installed Bugzilla before then, but it was so difficult without checksetup.pl that the barrier to entry was too high.

Then, I know for a fact that several large organizations started with Bugzilla 2.10 (May of 2000). This was probably around the time that Bugzilla became well-known not just as "Mozilla's bug tracker" but as "the bug tracker that open-source projects use." There were some corporations using it too. Lots of people hacked up and customized their 2.10 installations, to the point that some of them only upgraded when we reached 2.22 or 3.0--the effort to bring forward their customizations was just too much until the Bugzilla Project had advanced to the point where the number of new features outweighed the difficulty of upgrading.

Bugzilla 2.14, in 2001, was where we had our first security audit. We had security advisories before then, but as you can see in the 2.14 security advisory, there were so many holes before that point that we couldn't even list them all.

Bugzilla 2.16 (2002) was, for a long time, our most popular release. Largely, this is because we took so long to get Bugzilla 2.18 (2005) out the door. As for why it took so long, I don't know all the details, because I wasn't around, but I do know that we engaged in several massive projects all at once: templatization of the UI, changing the groups system, and adding flags, and apparently we didn't want to release until they were very much done. The Bugzilla 2.18 development cycle was also the first time that we had several interim development releases before the final release. Until recently, there were lots of installations still using those interim 2.18 releases (numbered 2.17.1, 2.17.2, etc.), because they had such better features than 2.16, even though they were unstable. I hope that we never get into that situation again--development releases are great if you really know what you're doing, or you're in a testing environment, but they really should not be being used widely in production.

Then there was Bugzilla 2.18, which was the first version that I helped release. Actually, what happened around the time of Bugzilla 2.18 and 2.20 (2005) was that myself, Marc Schumann (Wurblzap), Shane Travis, and Frédéric Buclin (LpSolit) all showed up at once, and that gave a big kick in the pants to Bugzilla development. We called ourselves "MINB" as a little joke--the Minor Intra-Nation of Bugzilla.

I think for me and Shane, we had both been drawn to the project because we were using Bugzilla at our employers, and the whole "no new releases in a long time" thing was really affecting us. I personally wanted to see PostgreSQL support in Bugzilla, and Shane wanted various features, initially for the company that he worked for, and then later also just because we thought they'd be neat. (Shane was the original developer of the "General Preferences" system that you see in Bugzilla today.) LpSolit and Wurblzap might have similar stories, I don't know.

What's interesting about that is that it represents a pattern that I've long noticed in open-source--the two times that you get contributors are when your project is doing really well, and when it's doing really poorly but being used everywhere. When it's doing really well, there's just lots of potential contributors. When it's doing really poorly, the smaller pools of contributors become more motivated to contribute out of necessity. When you're just doing OK, there aren't a lot of new people who show up to help. You've got to be super-popular (like WordPress) or dismal-but-used (like very early Linux kernels).

Anyhow, back to Bugzilla. The short story from then on is that we focused on getting mod_perl to work, and that led to a lot of re-architecture. Well, either that or I'm an architecture junkie and mod_perl support was a really good excuse to do it. :-) The great advantage of all the re-architecture was that it allowed us to add new features really quickly, and on May 9, 2007, Bugzilla 3.0 was born with a zillion major new features, and was surprisingly stable for a major new release, too (thanks to our QA team).

Now we're just about ready for our next major release, Bugzilla 3.2, and I think that we're starting to get close to the super-popular level, because I'm seeing more and more new contributors come out of the woodwork, and I'm really happy about that. I'm also seeing more and more people complimenting the architecture of Bugzilla, appreciating its features, and saying generally nice things to us.

I think we've had a pretty good ten years. It's been a little bit of a bumpy ride, but we've stuck it through pretty well and I think that in many ways, the world is actually a better place for it.

-Max