FeedBurner Considered Dangerous

Robert's RFC for short URL auto-discovery reminded me the other day that FeedBurner could be responsible for the same potential mass link-rot problems as URL shorteners, depending on how aggregators interpret their feeds.

It doesn't seem like RSS has a concept of how to present an item's "real" (I guess canonical) URL, versus the URL of the item within the feed itself. Or if it does then FeedBurner doesn't use it.

For example, the link element in an item from Eschaton's feed contains "http://feedproxy.google.com/~r/blogspot/bRuz/~3/WwfzaA_LDo8/thursday-is-new-jobless-day_16.html" rather than "http://www.eschatonblog.com/2009/04/thursday-is-new-jobless-day_16.html", and (at least in Bloglines and Google Reader, the two aggregators I've used) that first one is the URL that you get when you try to copy the item's address to link to it in a post of your own.

The real URL is there, but it's in the origLink element in the FeedBurner namespace, which apparently not even Google Reader knows to look at. If you're not anal retentive enough to actually follow the proxy URL and resolve the redirect then you'll most likely end up linking to the FeedBurner link that's subject to all the same risks as a shortened URL. FeedBurner, being owned by Google, admittedly looks like a pretty reliable bet at this point, but who knows what the future holds.