ILC 2010 part 1
So, the International Lisp Conference is over, and I've somewhat recovered from travel, so here are a few things about it.
First, it was great to meet people in person whom I had previously only talked to online. It was nice to talk to Peter "gigamonkey" Seibel and he signed his book for me. He's hard at work getting Code Quarterly started, and you should go work on the latest code challenge.
I also wore my Quicklisp shirt the first day and several people came up to me and said nice things about it (Quicklisp, not the shirt).
On the first day I opted to sit in on and show support for the CL tutorial session instead of going to the SPLASH keynote. It was kind of a mistake, as the pace was glacial (it was scheduled for 3 hours spread out over 3 days) and as an amateur CL language lawyer it was hard not to nitpick. I think almost everyone else there already knew CL and was there for some purpose other than to learn CL. (Someone said they were there to learn how to give a CL tutorial. I'm not sure the session was a good way to learn that.)
Larry Hunter gave a talk about computational biology and the goal of producing something AI-ish to extract useful results from the tsunami of data available in the field. Though he expressed affection for Lisp, the talk was pretty high-level and was almost a recruiting pitch to get programmers to help solve big, sexy computational biology problems.
I was pleased to see a fellow Mainer, Professor Roy Turner of the University of Maine, talk about a simple literate programming system for Common Lisp. He discussed the original Knuth-style systems and how he a chose different implementation approach because of a different programming style offered by Lisp.
Faré and Robert P. Goldman talked about ASDF2, and the slides provided rationales for some of the ASDF2 design features that weren't clear to me before. For example, ASDF2 broke pathnames in existing systems in some contexts, and one slide in particular explained the unpredictable nature of ASDF1's pathname handling meant that while the systems might have worked for me on SBCL and Linux, working on other implementations was not as predictable. So ASDF2 uses an incompatible system that has the advantage of behaving much predictably across implementations and operating systems.
The first night a dozen or so of us headed off to the Great Basin Brewing Company down the road from the hotel, and managed to get split up into several tables. I dined with Dan Weinreb and Paul Tarvydas and heard a lot of interesting stories from Dan about MIT, Symbolics, Object Design, ITA, and more. It's the kind of stuff you can only get by spending some face-to-face time with people; it's just not the same reading about it on a blog or in an email.
I'll write a little bit more about it tomorrow.