Pete Hopkins (This Pirate Kills Fascists) rounds up some recent posts about feed subscription in Telling An Aggregator To Subscribe.
Besides my recent entry on this subject, he summarizes a Kevin Burton (NewsMonster) post called Aggregator Subscription Mechanisms and one by Scott Johnson (Feedster) called A Suggeston for Aggregators.
Pete supports the prepended URI approach and says
I really like Kevin Burtonï¿s suggestion of prepending feed: to the HTTP URI of the feed file. For my site here, the URI would look like:
feed:http://pirate.typepad.com/blog/index.rdf. This solution makes it explicit that you’ll be using HTTP to get the RDF file, and presumably that you could use HTTPS and other protocols if necessary. … (I believe that Feedster has the clout to make a de-facto standard if they’re first out of the gate.) Aggregators can let the operating system know in whatever manner it takes that they can handle the new scheme, and then real-working subscription links can be added to web pages.
BTW, to most people this all sounds like gibberish: URI, aggregator, RSS, ping. It’s crazy talk. The whole point is we’re trying to figure out a way to put flesh on the mechanism so that people can punch up what they want and ignore how it works.