Phillip Pearson - web + electronics notes

tech notes and web hackery from a new zealander who was vaguely useful on the web back in 2002 (see: python community server, the blogging ecosystem, the new zealand coffee review, the internet topic exchange).


More on decentralized social networking

Cool, Brad Fitzpatrick has revealed his thoughts on social network portability/interconnection. In short: create a single open über social network that aggregates all the rest, with cooperative ones (e.g. Brad's own LiveJournal) integrating directly, with browser plugins used to micro-scrape uncooperative ones (e.g. MySpace), taking great care not to do anything that would give them an excuse to ban you -- i.e. only ever scraping your own data.

The goal is to enable new applications that want to access "the social graph" (their users' relations on all other sites) to do so without having to try to collect the whole thing.

Very interesting, and it could be incredibly useful... quite a project. And Brad's in touch with many others in the industry, so this could actually work.

That said, I still prefer social network decentralization over portability. My thoughts are pretty much in line with this great statement of the point of decentralization vs portability by Dare Obasanjo: "when it comes to email which is more important? The ability to send emails to people regardless of what email service or mail client they use or the ability to import your contact list from one free email service into another when you switch service providers?"

People are focussing on being able to import friend lists between services, and perhaps on being able to move your profile around, but what about the many other functions of social networks? Private messaging is powerful; I'd love to be able to send messages between networks. How about friending someone on another network, or joining a group there?

Unfortunately it's a 'boil the ocean' problem -- you pretty much have to get everyone to support a common API for interconnection. Either build a whole new network from the ground up (something like what I roughed out in my 'DSN' post last week), or get a few suitably interesting networks to connect, with a published protocol, and some process to integrate with anyone else supporting the protocol.

Let's see how this plays out. Will the social network vendors be more open than the IM networks? I'm currently using Pidgin (previously GAIM) to access Y!, MSN, Gtalk and AIM. Y! and MSN are interoperable, and Gtalk uses an open standard, but the solution for me is still to use an 'aggregator'-type client.

(BTW - I really need something like what PubSub used to provide: a decent feed of incoming links to my blog. The Technorati feed still seems to send me a notification every time someone with me in their blogroll writes a post!)

... more like this: []