Decentralising Twitter - Centralising Me

Case in point. Scott Hanselman is looking at an open, distributed implementation of Twitter, and in so doing gives me a great excuse for another example of the illusion of the Centralised Me.

Play along at home.

Twitter is centralised, and so downtime has a massive impact. The data is centralised in their data silos and so it is part of the "Decentralised Me".

If we view Twitter as micro-blogging (another great buzzword), why should I use a 3rd party service when I already have my own blog? I can decentralise Twitter by centralising it on my blog. I've fixed the (service level) downtime problem, and now the data lives in the "Centralised Me".

And you could subscribe to my Tweets (terrible buzzword), and I could publish a list of all the people I was following. But by decentralising the service, I now don't know who's following me (assuming a standard RSS feed subscription - which we of course will decentralise by moving to the centralised FeedBurner service).

Twitter's centralised service provided a whole heap of very important plumbing. Not least of which was usernames. With a centralised list of usernames, I am able to address messages to people. And with a centralised message store, people can be notified when they are addressed by people they aren't following. A decentralised service cannot offer this without additional, dare I say it, centralised infrastructure (think global usernames and semantic search).

Yes, I'm clearly having far too much fun with this.

I just wanted to reinforce what I said last time - centralised and decentralised are simply points of view.

Which means the Centralised Me is either going to be an illusion, like FriendFeed, where your data is aggregated - read copied - from multiple data silos into one new data silo, or it's going to be something much more interesting. How about a data silo that's a superset of all the data silos you've contributed to?