[OpenID] Identifier persistence (was RE: Don't you think digital identity URIs should have aspecific TLD?)

[Drummond Reed]

>James A. Donald wrote:
>
>The concept of having reassignable identifiers and
>persistent identifiers is an example of Zooko's
>triangle, however XRI is only a partial and incomplete
>implementation of Zooko's triangle

Zooko's triangle (http://en.wikipedia.org/wiki/Zooko%27s_triangle) has been
discussed several times by XRI TC members. XRI architecture is actually
neutral with respect to Zooko's triangle because, depending on how you want
to use XRI syntax and XRI resolution, it can be used to implement any two
edges you want (human-meaningful, unique/secure, decentralized). Arguably
the synonym capability of XRI resolution (XRDS documents) lets you bridge
XRIs from different implementations, possibly overcoming the two-edge
limitation (but that's not an argument I care to make ;-)

One thing Zooko's triangle doesn't discuss, however, is persistence. It
perhaps be added to the discussion of the "secure/unique" axis, but it adds
another dimension that I don't believe the original analysis went into.

=Drummond