Ah, a good start to the day: I got up nice and early so I didn’t have to rush through the morning routine, and joined a few other geeks at Pia and Jeff’s apartment for breakfast. After downing our daily intake of leet-bix, we went with high expectations to Damian Conway’s keynote, titled “Sex and Violence: Technical and Social Lessons from the Perl 6 Development Process”. As I’ve mentioned in previous posts, Damian’s a brilliant speaker, and today was the best talk I’ve seen that he’s given. Not too many public speakers I know can take pokes at everyone from Guido van Rossum to type theorists and get laughs from the people he’s paying out, weave in slides featuring S&M pictures and Web pages from What Would Satan Do, give good, practical advice to people managing any sort of project, and inspire the audience about Perl 6, all at the same time.

Along with Gus, I was lucky enough to have a good chat to Damian afterward about the interaction between the Perl and Haskell communities, the awesomeness of the amazing Audrey Tang, and thank him for taking Perl 6 in what I consider to be the right direction. One of the greatest things about these conferences, and Linux.conf.au in particular, is that you do get to rub shoulders with the best and brightest in the world.

(Warning to my non-geek friends: geek content in this paragraph!) The rest of the day turned out to be the most educational one so far: Conway’s keynote taught me many useful things about project management, and I learnt plenty from the other talks too. For revision control geeks who somehow missed svk on their radar: go check it out. All I have to say about it is that you can work in a distributed manner (e.g. offline) with any existing Subversion repository, in a very simple, darcs-like way. Very nice indeed. Van Jacobson’s talk about optimising the networking stack in the kernel actually drew a standing ovation from the crowd, and he proved that he’s still a networking God after 20 years: he had some incredible tables and graphs to show just how well he managed to optimise networking performance on Linux. (I was quite amused how their optimisations made 10Gb Ethernet peak out at 4.3Gb/sec because that was the peak limit of DDR333 RAM…) Interestingly enough, Shane and Jaq pointed out that the guys at the Gelato project at the University of New South Wales have been researching and advocating this approach for years, except that they did their research on that evil M thing that’s so hated in the Linux community (“microkernel”). Not to discredit Van Jacobson in the least, but it’d be nice to see the KEG, NICTA and Gelato guys get some due credit for their work too!

So, after yet another afternoon of more fun geekness, it was time for the LCA conference dinner. This year’s affair was pretty standard for an LCA dinner: that is to say, very good, with lots of drinking, plenty of socialising and chatting, and, of course, an auction. This year, they auctioned off the the John Lions book on the annotated source code for the AT&T Unix Version 6 kernel, autographed by some of the most renown UNIX hackers in history: Dennis Ritchie, Ken Thompson, Marshall Kirk McKusick, Eric Allman, Peter Salus, Linus Torvalds, Van Jacobson, and all the speakers and conference organisers for LCA 2006. I am proud to say that thanks to the great generosity of the students, lecturers and alumni at the University of New South Wales, UNSW won the bid for the book. As such, the Lions book will return to the shining halls of UNSW in its rightful resting place, where John Lions lectured and inspired so many students for so many years. Kudos to all those who contributed in the bid for this piece of history, and thanks to Benno and John for organising it!

That wrapped up the night quite nicely for me and set me in a hell of a good mood for the rest of the evening. Thankfully, I decided that I had quite enough drinking for the evening (you know, since I actually had to pay for my drinks that night), and that more than four hours of sleep before another full day on Saturday might be an idea…