Friday, July 02, 2010

Diaspora: the $200k scam on free software

After a month of work, its bluntly clear the community has been scammed for $200k.

Four NYU students posted a video and solicited funding on the kickstarter website, promising to build an open source, pro-privacy alternative to Facebook. The community response was astonishing. Its too bad the people pledging didn't research these students, for example, checking to see if they have Ohloh pages with their past contributions to other projects.

The first and most severe, after a month of work, they have posted that they have running code but will not allow the community (including the people who paid for it) to view until its "complete". This is commonly referred to as the Cathedral model. People have invested a great deal of trust in them to develop the code they have promised to, the least they could offer is read only access to their VCS. Instead they posted a video demonstrating it, which poses the second problem.

Their claim to use the "latest and greatest in web standards" fell flat on its face in their choice of method to post their video. Instead of using HTML/5 <video>, which they could have even used Ogg with the Cortado java player as fallback for MSIE and Safari, they used a flash player. Perhaps they don't realize Flash is not a web standard, or that HTML/5 has built-in video support, in either case its a really bad sign.

The most blunt demonstration of their lack of skill and commitment is in their use of websockets. At this point, websockets is still in heavy draft state with the few browsers/servers implementing it all running a slightly different, incompatible variants, none of which will be the final draft which is (at least) months off. Moreso, they propose to implement chat and gaming over their own websockets-based protocol rather than using existing standards such as XMPP.

XMPP is being used by a wide number of social network services already including Google, Livejournal, and Facebook. Its an established, widely implemented standard recognized by the IETF and the community at large for federated, real-time exchange of chat and presence (status). Beyond chat, XMPP has support for PubSub, Data Forms, and Ad-Hoc Commands. There is virtually nothing Facebook, or Diaspora, aims to accomplish that cannot be implemented using these existing standards.

What these four students are building is lock-in, if they finish at all. Instead of replacing the walled gardens of Facebook and other proprietary social network sites, they aim to build a network of little sites running the same software, only able to federate with each other, and only able to be extended with their custom APIs. These sites will not federate with other distributed social network efforts being built such as movim and concordance.

Hopefully the community will learn a lesson from this. Pledge money to experienced developers with demonstratable code, not charismatic students pledging to work on free software only if they're paid to do it.