Nathan Vack wrote:
OK, I'll bite:

* We build a registry mapping OpenID providers to OpenURL resolvers.
Yes, I've been thinking along exactly these lines too.

One logical place for this registry to live is in the already existing
OCLC Registry which already includes institutional link resolver
registration. They just need to add a component for individual OpenID
registration to one of the already existing link resovler registrations.

I tried to explain this to someone at OCLC, but they didn't seem to
understand what I was talking about, or the need. Perhaps I was talking
to the wrong person.


* A user comes to our tool for finding licensed material (eg, a
LibraryFind implementation)

* If (by IP, OCLC's link resolver) we know the OpenURL resolver,
rewrite URLs to point at that resolver.

* Otherwise, we punt to an OpenID login form, and look them up in the
OpenID -> Resolver registry, and use that resolver when rewriting links.

Now, anyone whose institution has both has an OpenURL resolver and
provides OpenIDs can use our tool, without making any interaction
with us.

The really nice thing is that (at least for us) the OpenID resolver
handles trust issues, proxying requests if necessary. The resolver
doesn't need to be OpenID-aware -- though it would make for a nicer


Jonathan Rochkind
Sr. Programmer/Analyst
The Sheridan Libraries
Johns Hopkins University
rochkind (at)

Reply via email to