finally, an up-to-date and quite good resource about rss: "how to syndicate web site news through rss feeds technology".

it contains a reference to another just-as-fresh article called "rss – lo-fi content syndication" which denotes some issues to look out for when providing rss feeds.

one of these is about bandwith and i think this will become (or probably already is) the crux: the idea of having a news aggregator on each desktop makes one single rss feed being fetched as often as there are subscribers which will be killing any lo-fi weblog (which might sound tautologic in most cases).

and it won't be enough to take etags and 304 responses into account. it will need a few centralized rss aggregator "backbones" that get a ping of each rss feed when it has changed. (not having said that decentralisation is a bad thing, generally. not at all. i just doubt that it will be working with news syndication.)

that leads me to my own little project called parss which should already be running on my new server. it's just that latter is still waiting to be moved to the right place in the flat where its noisy fans don't disturb my sleep.

so apologies for still teasing only (especially to chris), the time will come.

but i digress. parss, however, can do all of the things mentioned above. it has an xml-rpc interface that can handle rss pings, addition of and certainly request for feeds. it respects etags and its goal is to minimize the fetching-rss-by-time-based-guessing method.

furthermore, there is an antville module that connects this wonderful publishing software with a parss server and provides user-defined as well as merged feeds. uh, and karma voting. and direct-post™ technology, heheh...