[OpenID] No benefits of XRI i-names/i-numbers as OpenIDs(was:isopenid 2.0 a lightweight identity system?)

Kevin Turner kevin at janrain.com
Tue Feb 13 23:01:48 UTC 2007


Les Chasen wrote:
> [David Fuelling wrote:]
> > Why should the OpenId spec coronate a given vendor/endpoint?  A "good"
> > spec should be more agnostic (like it is now).
>
> I think it is.  It is just pointing people to a public server that can
> resolve an iname so that folks don't have to install one themselves.
> Xri.net is not the authoritative source for an XRI.

But everyone assumes we're talking about i-names, not fully general XRI
resolution.  I've yet to hear anyone really advocate using, say, an XRI
with a community root with OpenID.  That is, you don't want OpenID to
support XRI in order to use xri://(your-uri.example.com)*johndoe as a
personal identifier.

So we're just talking about i-names in the = and @ registries, which
_do_ in fact each have a single authoritative source, equal.xri.net and
at.xri.net, if I understand the GSS on that point.  I'm not trying to
say that's a terrible thing -- the existence of the GRS is necessary to
support many of the properties that you've designed i-names to have --
but I think it's a little confusing to try to argue that it doesn't have
this centralized, single-authority property.


and, stepping back from this particular detail for a moment,

I still think the best response I've seen to this debate is the message
from Scott at the beginning of this year,
http://article.gmane.org/gmane.comp.web.openid.general/3633
(03 Jan 2007, "Re: [OpenID] Anti-XRI FUD").






More information about the general mailing list