This post originated from an RSS feed registered with Agile Buzz by James Robertson. Original Post: Sunday Feed Title: Travis Griggs - Blog Feed URL: http://www.cincomsmalltalk.com/rssBlog/travis-rss.xml Feed Description: This TAG Line is Extra Latest Agile Buzz Posts Latest Agile Buzz Posts by James Robertson Latest Posts From Travis Griggs - Blog

I can't seem to get BottomLine working from MacOSX here at the conference, so I'm actually having to VNC to my Linux box back at work over a VPN to do these posts. The performance is .... stellar.

Today is tutorial day for me. I started out the morning with a session by Alistair Cockburn, regarding "Tuning your Agile Methodology." It was interesting enough. The swimsuit metaphor was quite entertaining. I liked the soundbite "we should means we won't." Some good ideas about people issues. Don't worship any one methodology. Meat regularly to update your methodology. Consider throwing something out for each thing you add.

In the afternoon I went to a "CLOS/MOP exploration" tutorial given by Pascal Costanza. This was really cool. Really, I just went to get some exposure to Lisp. I know, this is worse than embracing a weird technology like Smalltalk. But I've never met a real Lisp person who knew other languages who didn't just love it. So somewhere out there, along with being accomplished at Smalltalk, I want to be able to do CLOS. I had downloaded the Franz Lisp demo version in preparation. It wasn't helpful. Luckily, Pascal handed out demo CDs for LispWorks (runs on Linux, OSX, and Windows). This was a much more approachable environment. Pascal took us through definining a simplist class/metaclass model. And then generic functions (i.e. multimethods). And then we threw that out, and switched to CLOS where he derived a Python object model inside of the Lisp environment. Lisp seems to me like Smalltalk. Simple principle of execution to swallow, huge library of things to learn to actually really be productive. One of the things Pascal demonstrated was how you could use MOP to create a class where the "instance variables" were not actual object slots, but database field accessors instead. This was really cool. I spent a while thinking about how one might do this in Smalltalk. Seems to me you'd have to allow the compiler to "consult" the target class environment when it goes to do an ivar fetch/store. By default you get the normal bytecodes, but a given class could extend the behavior to do interesting things.

Took a break in the evening to go watch "Serenity". Cool movie.