bub =~ s/pub/hub/
PubSubHubbub is a sensible step in the direction of a more distributed, less micro, real-time blogging system. Subscribers to a feed register with a third-party hub that then pushes updates to them. Publishers help out by telling hubs when they’ve updates to publish, and recommending hubs in their feed. Hubs take on the complexity of managing subscriptions, calculating and caching updates, and the accordant scaling issues - so publishers now have an easy route to real-time.
This is yet another technology with Brad Fitzpatrick somewhere behind it. I’m continually impressed with his pragmatic approach to politically viable simple pieces of webware, even if they don’t always catch fire. Being part of the Googleplex doesn’t seem to have hurt his open bias, and PubSubHubbub seems plausibly neutral (although for 100% certainty, I guess there’s Dave Winer’s RSSCloud.)
It occurred to me that I have a few syndicating and cacheing jobs that I want to provoke to run asynchronously every time I publish a blog entry, and that I might use the protocol to loosely join the small pieces involved, with the task runners as subscribers to the feed.
Conceptually neat, but because I use external services to host my blogs, I already have to manually intercede to ping the hub; and actually that’s then a very natural point for me to hook in all the provoked tasks, especially since I can take the opportunity to sub-edit syndicated content (e.g. choose which 140 characters make the best summary tweet.) So I’ve ended up just publishing to the default open hub (as you should be able to see in the feed.)
