During my vacation I got fed up with the lead-grey theme and decided to start anew. This is the first visible result, and I'll be changing more stuff around the site soon. Not all sections have been updated, but then, the photo album was a work in progress anyway, and I mean to revamp the cvs section as well.
Hope you like the results.
It seems there are more and more people working on this besides me. After reading Les Orchard's post about his own aggregator (which is, incidentally, also written in Python), I found Krishnan Nair Srijith's site and his Perl-based Intelli-Aggie aggregator.
Srijith's post on the suitability of Bayesian classification for processing RSS and Les's take on it make for an interesting read, and I think both are right, in a sense - Bayesian classification may not be the best way to tackle RSS feeds if you handle them exactly like e-mail (since, like Les points out, you use both kinds of information in a different way and you have already biased the results by subscribing to stuff you actually want to read), but it might be useful for separating items according to topic - or, at the very least, make sure some interesting items you might miss come out at the top of the list.