[OpenID] is openid 2.0 a lightweight identity system?

Gabe Wachob gabe.wachob at amsoft.net
Fri Feb 9 21:02:55 UTC 2007


You've made the assertion that supporting XRI adds a LOT of complexity. I
just don't see it that way. 

>From the OpenID point of view, one can see XRI resolution simply as "add
http://xri.net at the beginning + some query options at the end" and do an
HTTP request [1].  Other identifier schemes could be resolved in a similar
manner if someone deployed such resolution infrastructure
(http://ssn.net/<socialsecuritynumber> - Lets hope not!). In any case, you
get back a descriptor document. 

To the extent that it's more complicated, it is something that's either an
implementation problem on the xri.net proxy side [2], or something the
inames community should address to make it easy for OpenID developers.  

Using HTTP as a resolution protocol and getting back an XML document with
some basic semantics about a service endpoint is what's new (this is what
XRI early on - YADIS is expansion of that concept with HTTP URLs) - I think
that's mostly the extent to which an OpenID RP developer should be saddled
w/r/t new identifiers.

For what its worth, we (the XRI & inames communities) are trying to bridge
this understanding/complexity "gap" - I've begun (was derailed for a while
by wiki hiccups) a busy developer's guide:


[1] The one wrinkle is the concept of canonical IDs - but as Steve Churchill
said, the mechanics of discovering and confirming these IDs is something the
proxy resolver (xri.net) should be doing. 

[2] BTW, xri.net is based entirely on the OpenXRI codebase - you can run
your own proxy - nothing special about the xri.net instance except that its
well known and run as a bootstrapping mechanism by the same folks who run
the root registry.

> -----Original Message-----
> From: general-bounces at openid.net [mailto:general-bounces at openid.net] On
> Behalf Of Robert Yates
> Sent: Friday, February 09, 2007 6:01 AM
> To: Recordon, David
> Cc: general at openid.net
> Subject: Re: [OpenID] is openid 2.0 a lightweight identity system?
> On 2/9/07, Recordon, David <drecordon at verisign.com> wrote:
> >
> > As you can see in the wild by OpenID 1.1 implementations taking
> advantage of
> > Yadis (XRDS) based discovery.
> >
> Just to be clear, I do believe that the introduction of Yadis adds
> real value.  My real concern is mandatory support for i-names in the
> core, this adds a LOT of complexity.  Could i-names support, instead,
> be done as an extension and the core modified so that alternative
> valid openid URI schemes can be defined later?
> I don't think a modification to the core spec. allowing for future
> alternative URI scheme definition would be that hard to do. Extensions
> defining alternative URI schemes simply need to define discovery rules
> and possibly some different verification steps. It could add real
> flexibility to the spec as it seems fairly limiting, in my mind, to
> restrict openid URI schemes to only http/https or xri.
> Rob
> _______________________________________________
> general mailing list
> general at openid.net
> http://openid.net/mailman/listinfo/general

More information about the general mailing list