[OpenID] Microsoft and OpenID Q&A

Christopher St John ckstjohn at gmail.com
Fri Feb 23 23:33:13 UTC 2007

On 2/23/07, Josh Hoyt <josh at janrain.com> wrote:
> OpenID 2 is not terribly far from OpenID 1. The points that add
> complication are:
> 1. Yadis
> 2. XRI
> 3. namespaces for extensions
> Other than that, it's pretty much the same as it was.

Incorporating other specifications by reference into the core
of a spec means, basically, tacking on the entire referenced
spec to the end of the OpenID spec.

XRI in particular is viewed in certain quarters as being a bit, uhh,
well, to put it nicely "unproven", and it seems a bit weird to be
incorporating it into the core of a spec that seemed to be taking
a "paving the cowpaths" approach of using as much proven,
standard technology as possible to get a very focused, well defined
job done.

I speak here as a spec implementer who went through the pain
of infinite spec recursion when implementing WS-Security and
others of the WS-* constellation-o-layered-complexity.

Christopher St. John

More information about the general mailing list