Log in

No account? Create an account

Previous Entry | Next Entry

Posted This On Our Internal Blogs

This morning I attended the second Internet Identity World Conference, this one being a session focused on developers (http://www.socialtext.net/iiw2005/) wanting to get into the identity space. I spoke, lighting ten minute rounds though, mainly focusing on OpenID, but also touched a bit on YADIS. Was a group of about half people I had met before, and seen Brad’s previous presentation, as well as about half of a new group of people. Presentation is at: http://www.danga.com/words/2005_iiw_developer_lunch/

This evening Brad and I then met with Dick Hardt of Sxip. Sxip 2.0 is largely based on the model OpenID uses, mainly due to Brad telling Dick how lame Sxip 1.0 has been for a few months. Sxip 2.0 is now based on URLs as identifiers and is no longer centralized or relying on a certain service to be up.

Wrote the following on the ID Workshop mailing list about it:
I had the chance to see John Merrell's presentation on Sxip 2.0 this morning and then a chance to talk with Dick along with Brad Fitzpatrick this evening. I don't know if the large announcement should be "Sxip is going URL based" or "Dick is listening". I know Brad has been putting him through the wringer with Sxip 1.0 often expressing things as being "lame" or "retarded", but I don't think neither he or I found faults in the 2.0 proposal. Dick even had answers for us when talking about transactions which take place without a browser being involved, ie web service to web service.

Without going into the nitty-gritty, they seem to have found an extensible balance between being "lightweight" and providing the value services on top. A single user-experience, that makes sense, has been worked out and there are many easy optimizations which can be built on top of the base spec.

One different between YADIS and how Sxip 2.0 will do capability discovery is that Sxip does discovery against the homesite whereas YADIS does it against the actual persona. Both have pros and cons, though it would be nice to see convergence on this between both specs. Also allowing a membersite to support someone entering either their homesite or full-blown persona URL. Brad and I were speaking about this afterward and it really doesn't seem that hard to do, though does add a bit of overhead to both protocols.

All in all, I'm very excited to see what Sxip 2.0 ends up looking like and think it will continue to bring great value and innovation into this space.


( 2 comments — Leave a comment )
Dec. 14th, 2005 07:32 am (UTC)
I don't recall ever calling Sxip 1.0 "retarded". My biggest complaint was always that it was tied to sxip.net, so nobody could trust it staying alive/etc.
Dec. 14th, 2005 07:41 am (UTC)
Didn't mean to make it sound like you called Sxip 1.0 retarded, though can see how it could be interpreted that way. Sorry.
( 2 comments — Leave a comment )