The Death of Twitter
Pubsubhubbub sounded cool, so I looked into plugging it into my publishing system. It turned out to be very simple: add one element to the news feed, and ping a “hub” when I have new content, by simply doing an http POST with the address of my feed. There is even a python module that does this for you.
It didn’t work at first; trying to find out why led me to discover that my atom feed was not quite valid, knowledge that itself was worth the excercise. After fixing that, it works.
By “works,” I mean that my test post popped up in Google Reader two seconds after I pinged the hub. This is fantastic, and what the multitudes have pined for.
Since news feeds can now be instantly transmitted, why would Twitter persist? Feeds have no character limit and can contain anything, including pictures, sounds, and advertising (obviously Twitter’s enforced simplicity has some advantages). Even more important is the fact that the pubsubhubbub mechanism depends upon no central server or company. It is a free and open protocol (like Google Wave), so anyone can operate a pubsubhubbub hub. If Twitter goes down, and you rely on it for your publishing, you are out of luck. If the pubsubhubbub hub you’re using goes down, there will probably be another one you can point to, or you can wire up your own.
Not that I really believe that Twitter’s fans will abandon it for something clearly better. Fashion trumps function, at least in the short to medium term. I do expect that the pubsubhubbub protocols will be the foundation of new services and ways of communicating that have yet to be built.