October 27th, 2010

ILC 2010 part 3

The last day of the International Lisp Conference started off with the SPLASH keynote by Kenneth Stanley. He spoke about some of the results he got from his website, picbreeder, where people were discovering interesting and complex images, but arriving at them only accidentally. His delivery and cadence reminded me a little of a polished version of the rural preachers I saw as a kid. On the goal-less nature of evolution he had the line "What is the goal of evolution? What is the pinnacle to which it aspires? Is it you?" (Calvin might think so.) There was a lot more to the talk, and he kept the audience rapt for an hour. The SPLASH page has a decent thumbnail sketch of it.

Back on the ILC front, Lowell Hawkinson talked about some of the success he achieved with the Lisp-based G2 real-time expert system from the company he founded, Gensym. The company's clients included NASA, Exxon, the Panama Canal, Biosphere 2, Nabisco, and more. He had a lot of amusing anecdotes about different clients. A satellite company rewrote years worth of their control software in G2 and invited Lowell and another executive to be present for the first launch. "We felt a little like hostages. What if it didn't work?" But everything worked fine. At Nabisco, he saw the engineers were using the live plant to test changes, mostly because it was cheaper to have an occasional minor cookie disaster than to set up a separate test environment. NASA once urgently called to get a renewed license code - they had replaced a board in a computer, and it triggered the G2 DRM and locked them out, 7 minutes before a scheduled launch. Gensym support was able to expedite the request and the launch went as planned.

He also talked about his new project, EXP DB and EXP READ, but I had a harder time following it. His ILC speaker entry covers it a bit.

(Lowell also used my DisplayPort-to-VGA adapter to hook up his Mac to the projector. The same thing happened at the July Boston Lisp meeting when Slava Pestov spoke. I'm a walking Apple Store for Lisp gatherings!)

After lunch, Didier Verna talked about his Clon library, which is for processing command-line arguments. I was conflicted about the talk. On one hand, it was strongly in my area of interest: practical, useful stuff written in Common Lisp. On the other hand, it looked very complex — intimidating, in fact — to configure and use. The latter impression might have come from the desire to show all the useful features in the talk. Maybe simple things are simple the complexity kicks in only when you want to do fancy stuff. I dunno. I'll have to read the manual. "Guys, always write manuals." Amen!

The official part of ILC closed with 90 minutes scheduled for lightning talks and "open forum." There was no formal signup for talks, and I only heard one person mention giving a lightning talk, so I thought it would be one quick talk, and then I'd get up and do a Q&A about Quicklisp for an hour or so. My impression was way off, though, as one person after another got up to quickling talk about stuff they were interested in or working on, and there was a lot of positive energy and interaction to the whole process. Robert Goldman, John Maraist, Faré, Paul Tavyrdas, Didier Verna, Kuroda Hisao, and maybe others (sorry if I left you out) all got up to say a few rods.

I wound up with about 15 minutes to talk about Quicklisp at the very end. I screwed up at my first time using LispWorks 6 as a demo platform, but people literally cheered when I got a web server (Hunchentoot) up and running on a bare Lisp install in about 90 seconds. There were good questions and I got a lot of positive feedback from the process.

For dinner about 20 of us went to the Mexican place down the road, and despite our unwieldy headcount the staff were very accomodating, rearranging half a dozen tables and generally taking good care of us. At the end, the ALU generously and unexpectedly picked up the tab. Thanks, ALU!

I'll write about some miscellaneous impressions and ideas tomorrow.