mart at degeneration.co.uk
Wed May 25 02:53:52 PDT 2005
Brent 'Dax' Royal-Gordon wrote:
> Actually, this sounds like a much larger idea than just OpenID. A
> module that takes an arbitrary URL, retrieves all the metadata
> associated with it, and sets it up in a usable, format-agnostic form
> would likely have applications elsewhere. Suggested name:
> WWW::Profile or WWW::AuthorProfile.
I was thinking the same thing. Make a generic module (what about
WWW::ProfileInformation ?) which can parse and return all of the
personal profile data for a given document, and then create the
OpenID-specific module (which accepts a VerifiedIdentity object) which
is just a thin wrapper around it.
However, the advatage of it being bound to OpenID is that it can borrow
the generic caching functionality from the consumer. I guess that
indicates that it's a good idea to separate that out into a separate,
non-OpenID-specific module as well.
The OpenID consumer has also already fetched all of the auto-discovery
URLs while it was getting the OpenID server, so there would need to be a
mechanism to throw in already-discovered URLs as well as the "get
whatever information you can about this URL" interface.
This suddenly got a lot more complicated!
Perhaps for now we should just make the OpenID-specific module for the
sake of getting things working quickly. We can always refactor it into a
generic module later, and with it being OpenID-specific it shouldn't
take more than a few hours work.
More information about the yadis