Users of Berry Bloglines should notice snappier performance after a port to Ruby on Rails from the older CGI scripts, whose overhead had grown too high under increased load.

Tech notes:
The new implementation uses ruby-bloglines. Compared with the Universal Feed Parser, Ruby Bloglines currently has some limitations. It doesn’t handle ill-formed XML and rather than meaningful objects it returns nested hashes with some inconsistent structures.

Unfortunately, Bloglines is an example of a site that doesn’t generate valid XML! For instance, ampersands in feed titles or descriptions are not escaped as & causing a non-lenient XML parser to blow up. Although I won’t complain as harshly as Tim Bray does about invalid XML, I do hope the Bloglines team can clean this up as it makes life difficult.

A better RSS parsing option might be Ruby FeedTools or perhaps something built on top of RubyfulSoup, but the bloglines web services interface would need to be modified to use it.