Been a busy period for me at the moment. Working full-time, taking three classes, repairing the house, and other family duties certainly isn't leaving much time for play! Luckily the light at the end of the tunnel is shining regarding UWF; although I'm really quite enjoying this semester's classes.
I've been delving into "Artificial Intelligence: Structures and Strategies for Complex Problem Solving" by George Luger. I'd previously been using "Artificial Intelligence: A Modern Approach" by Norvig but thought I'd check out logger's take on things. In particular I'm going over the machine learning section for Shoechicken (http://www.shoechicken.com & http://shoechicken.blogspot.com), and have thus far come up with some great ideas from the text. I hadn't thought about using a genetic algorithm for Shoechicken, but the more I think about it the more I like it. This will become a big source of interest over the next week or so.
At work we've been working hard to get a stable release of MAST (http://mast.ihmc.us) out. We've just switched to using InstallAnywhere and I have to say that I am suitably impressed! It's allowed us to package the software in a very professional looking manner. We'd previously had problems with licensing and using other software; in particular MySQL. Since MySQL is under GPL, although MAST is free to download and use, it's not GPL so were having the user download and install MySQL for the MAST Server. I suddenly had a brainstorm a couple of weeks ago, can't believe I didn't think about it before to be honest, why not switch to another db server????? We only need a lightweight SQL database and all database traffic goes through a proxy. We have now switched MAST over to use Hsqldb (http://www.hsqldb.org) and can report that life is gooooooood. We now have everything we need packaged with the installation, so it's just a single download.
We've also been testing, testing, testing, which of course means debugging, debugging, debugging. I think that this is the most testing we've done of MAST all at once. Matteo and I really have brought the product a long way in the last two weeks. The installation procedure has become quite robust, and MAST itself is looking much more stable. Our target is to have a final release out on the web that's been tested by a few different people, so watch this space to find out when it's ready for download!
