So the first actual day of the conference was fantastic.
I spent the morning in Damian Conway's "Best Practice Perl" tutorial, which was quite worthwhile. His major point was that you should make a concerted effort to write code such that it's easy to understand. There may be more than one way to do things in Perl, but that doesn't mean you have to use some of the odder ones in your day to day coding. He also had a number of tips on specific problems to avoid and useful idioms that you should be aware of. All in all, a worthwhile talk. Best of all, he's apparently working on a book about the same topic, for O'Reilly.
Then I went to grab lunch with Rich Bowen, which is always enjoyable. We talked mostly about writing technical books and the various problems we've both had going through various parts of the process. It's odd that so many authors seem to have the same problems, yet they still don't seem to have found good ways to avoid them.
The afternoon was spent in the "Cross Platform Rendezvous" tutorial, which was given by Stuart Cheshire from Apple, the head of the zeroconf working group and the creator of most of this technology. Much of what he had to say was about the philosophy behind the creation of the zeroconf technology, and why he thinks it has succeeded in a space where a huge number of other systems have failed in the past. Perhaps the most remarkable thing is how simple the underlying technology is. It's quite impressive how much functionality they've managed to get while still fitting pretty seamlessly into the existing infrastructure. He also talked about some of the plans for extending zeroconf techniques outside the local network and onto a company wide WAN scale network.
I'm itching to start coding on some changes to Subversion to allow it to advertise repositories via zeroconf so that you can easily find out the available repositories on a network. The idea's been in the back of my mind for a while, but it's always seemed like something of a toy project since it's not overly useful on just a local network. WAN scale zeroconf would make it quite useful though, so hopefully I'll have time to make it happen. I also talked with Stuart a bit about some of the particular issues I saw in using zeroconf and Subversion and he seemed quite receptive to the ideas I had for extensions to the current functionality to make it more useful for Subversion.
After the tutorials finished up the real fun began. I finally got a chance to meet Brian Fitzpatrick, who I've known online for years via the Subversion project. After giving Greg Stein enough time to extract himself from the traditional pile of people asking questions at the end of a talk the Subversion and Apache crowd went out to dinner at a nearbye chinese resturaunt, which was a lot of fun.
After dinner a number of us retreated to VQ, a nearbye bar, and proceeded to spend the rest of the night drinking and talking with some of the ThoughtStream guys, who brought along Ward Cunningham, of Wiki, TDD, and OO fame. I must admit that it's nights like this that make coming to these kind of conferences truly remarkable. You simply can't find a group of people like that in other environments.
The bar finally closed and kicked us out, so I made it back to my hotel room around 3AM, already fearing the hangover that was waiting for me.