Group items matching

in title, tags, annotations or url

I’m sure some folks are wondering exactly what this means. Even though I was close to the decision making around this, I believe it is still too early to tell. Personally, I share Marc Canter’s skepticism about Dataportability.org given that so far there’s been a lot of hype but no real meat.

As far as I can tell, Dataportability.org seems like a good forum for various social software vendors to start talking about how we can get to a world where there is actual interoperability between social software applications. I’d like to see real meat fall out of this effort not fluff. One of the representatives Microsoft has chosen is the dev lead from the product team I am on (Inder Sethi) which implies we want technical discussion of protocols and technologies not just feel good jive. We’ll also be sending a product planning/marketing type as well (John Richards) to make sure the end user perspective is also being covered. You can assume that even though I am not on the working group in person, I will be there in spirit since I communicate with both John and Inder on a regular basis.

Ultimately, the problem represents an opportunity for Internet companies. "In the end, whoever moves first to be truly open will have the advantage," Saad said. J. Trent Adams, founder and chief innovator at Matchmine, another vendor pushing for data portability, concurs. A stealthy startup could develop a clever application or service built on data portability standards and force other vendors to respond. "All of a sudden, it's the right idea, and the big boys will have to react," Adams said. In other words, Internet companies who insist on locking up their users' data to protect their businesses might soon find themselves instead stuck in the mud.

When Facebook joined the DataPortability.org Workgroup a few weeks ago, the press described the move both as a “bombshell” as well as “brilliant PR”. In order to understand what Facebook’s decision to join actually means a little bit better, I spoke with Chris Saad, Co-Founder and Chairperson of DataPortability.org.

IF: What does it mean for companies like Facebook to “join” DataPortability.org? CS: It means they agree to engage in the conversation and work towards a blueprint for maximum interoperability between applications.

IF: Has Facebook promised to implement any particular functionality by any particular time? CS: Not yet - but once the blueprint is done we can then start asking vendors to implement things. Many other vendors have already moved quickly - in the last few weeks and months lots of vendors have been implementing OpenID, etc - these things are not unrelated.

MySpace and Google demonstrated an interesting mashup of the MySpace Data Availability API, oAuth and the iGoogle gadget specification at the oAuth Summit a couple of weeks ago. The application, which pulls the core MySpace feature set into iGoogle, is not yet publicly available, although MySpace has said to expect in in August.

Chris Saad, cofounder and chair of the nonprofit ­DataPortability Project, notes that many current methods of transferring data expose users to huge security risks. For example, it's a common practice for social sites to ask users to submit the usernames and passwords for their Web-based e-mail accounts when they first sign up; an automated service can then search the network for people listed in their address books. "The door is open right now for any application that scrapes your Gmail address book to go ahead and scrape your shopping cart as well, or scrape your searches, or keep your username and password and pretend to be you," says Saad. "It's a nightmare of security, and it's something we need to solve sooner rather than later."

The bigger downside of Friend Connect is that Websites using it cannot mash up the data with their own to make compelling new applications. Glazer confirmed that the data will be sent to third party sites via an iframe rather than directly through a set of APIs (as Michael speculated on Friday). However, Glazer also says that he wouldn’t be surprised if eventually Google or somebody else makes it possible for Websites to combine the Friend Connect data with their own.

MOAT Server is a application that serves tag meanings (in HTML, JSON or RDF/XML, using content negociation) for any Tag you request. The server is shipped without any data, which means that users have first to define meanings for their tags (thanks to the clients they use, that will then update the server information) to efficiently use it, using data available on the Semantic Web, especially from the Linked Data initiative.