[dstone]

Have you thought about using Google Reader as a backend instead of keeping your own database of feeds? It seems like that would cut down on your processing of the feeds and enable you to just work on the stuff you care about (the filtering and UI aspects of it).

[conesus]

Yes, I have thought about that. (Ohh, how many times I wanted to switch to this while developing.) But that means I would lose a whole lot of control, as well as the ability to perform real aggregation and intelligence on those feeds.

That, and I get to build a full-on distributed feed fetching and processing service, which has been a brutal, expensive, and wholly worthwhile experience.