I've been looking more into the API exposed by Amazon S3. And you know what? There's nothing so complex about this thing that couldn't be cloned in a weekend or two in PHP. I wonder how hard it would be to clone in the OPML Community Server? I could see a bit of a race between the Atom Publishing Protocol and S3 to see which gets adopted faster as the Way to Publish Files Over the Web. The OPML Community Server API is pretty keen too, though no one's seriously proposed it as anything other than the Way the OPML Editor Docks with a Mothership.

Good morning. Here's hoping that tunnel-end light ain't a train.