Woo hoo! Got one monkey off my back, finally.
I wrote a 12 page paper about how our distributed hash harvester actually works. It's not on the web yet, because it needs to go through a review process, but I hope it will be soon.
It's not the most spectacular use of Perl. Having the same piece of code running on 6 PCs doing loosely parallel hashing, monitored by a master cron job and a couple database tables isn't rocket science - rockets are someone else's job.
Using sexeger was the best improvement we've made so far. It halved the time we were spending in our slowest section of code, and lowered our overall run time by 1/3.
I plan on going to YAPC::Europe::2002 and I submitted an abstract; hopefully I can give a presentation, but if not I'll certainly want to pick up as much input as possible. There's so much more I should be doing with benchmarking, hashing and ODBC modules.
More later, and hopefully a URL for the paper in the NSRL website.
