Tuesday, June 21, 2005
Well, I'm sitting here in an afternoon session. So far, this whole experience has been quite fun. After getting my Stanford parking permit situation straightened out this morning, I was able to catch the end of Jeff Shrager's How Lisp will Save the World session. Jeff did a great job of discussing Lisp in a biology context. I'm not a biologist and I still understood the overall direction of Jeff's talk, a great testament to him as a speaker. Additionally, he had a lively presentation style that kept me quite awake, even for an 8:00 AM session.
I attended three breakout sessions following Jeff's great session and I wished that I had gone to the other breakouts instead. I won't mention names to protect the guilty. The speakers generally spent a lot of time on setting up the problem without really describing the guts of the talk. Then everything was compressed into the last 10 minutes with lots of slides being skipped. One tip for presenters to any conference, please make sure that you provide good concrete examples of what you are talking about early in your presentation (like the second slide after the title slide would be good). If I'm not familiar with your problem domain, it's very hard to grok a lot of theoretical information in a 40-minute time slot. With a simple example in-hand, I can easily make the connection. A second tip is to quickly answer the "What's in it for me?" question that's going through everybody's head. ILC has a slightly academic feel to it and perhaps this is the normal style for such a conference, but as a newbie to the conference, this would let me enjoy even talks where I don't have extensive background.
I spent lunch with Peter Seibel, William Bland, and Ben whose-last-name-I-forget (sorry, Ben, I'm horrible with names). The lunch discussion was great. Nice discussion about Peter Gabriel's talk from the previous day, which I had missed. (Update: Okay, okay, Richard Gabriel. Sheesh.)
Following lunch, I attended Matthias Hölzl's talk, A Framework for Dynamic Service Oriented Architectures. In contrast to some of the weaker morning sessions, Matthias did a great job of getting through the material and communicating the main points. This was a challenge because his project actually uses Dylan and not Lisp, and so he included a short Dylan tutorial in the beginning such that people could follow the examples.
Finally, I just attended Paul Dietz's The GNU ANSI Common Lisp Test Suite. Paul is also a great presenter and did a great job of explaining the test suite and the rationale behind it. Personally, it's obvious that Paul's test suite is the greatest gift to the Common Lisp community in decades. The test suite is finding lots of bugs in all the various Common Lisp implementations, both commercial as well as open source. Assuming those bugs all get some attention from implementation developers, that's a great thing for application writers who are looking for consistent behavior. Unfortunately, to write a sizeable program in Common Lisp, you're still going to have to stray from the safety of the strict ANSI spec into implementation-specific libraries. In spite of this, the quality of CL implementations is being improved. Well done, Paul!
Post a Comment
Links to this post: