Previous Entry Share Next Entry
Diaspora might actually be the social network that gets it right
So the Diapora project finally penetrated my consciousness today. I must say, I'm very intrigued. It's just a small bootstrap project at this point, but it looks like they might actually get it: they're a small group focused on starting up a truly open-source, truly distributed, truly pluggable social network. We'll see if they get the pieces right (or at least, right enough to get it off the ground), but they are saying all the correct things so far...

  • 1
I looked at it yesterday, and they seemed to have the right idea, but I didn't see anything that suggested they had thought about the hard parts.

Possibly not -- we'll see. The nice thing about the project is that, simply by dint of its distributed architecture, it may *force* them to deal with the hard parts. A lot of the common shortcuts that cause problems arise directly from the fact that, at root, most social networks are just a single database, so it's way too easy to just shrug and cheat when you get to the complicated problems. If you take distribution seriously, it makes it much likelier that you will get privacy and identity right. (Or at least, that the protocols will *allow* you to get it right.)

They're quite explicit that they're doing a fast first pass, and expect changes to be needed as they go -- I can respect that. Starting with distribution as an upfront requirement just makes it likelier that it'll be possible to wind up in the right place.

(I should note that, in the later days of CommYou, I was moving to an actor-oriented architecture. That was partly for scaling and performance reasons, but it was also so I could make the system fully distributed down the line -- and because if you do actors right, they kind of force you to not cheat. Same idea...)

it may *force* them to deal with the hard parts

Good point.

  • 1

Log in

No account? Create an account