Re: an OpenID GRDDL profile

On 18/01/07, Dan Connolly <connolly@w3.org> wrote:
>
>
> FYI, I just cooked up an OpenID profile and a few slides to present
> it...
>
> Using GRDDL to ground OpenID in URI space
> http://dig.csail.mit.edu/2007/id/talk19


Great stuff.
Although OpenID had been beeping on the edge of my radar for a while, I
didn't get the idea until reading Sam Ruby's post [1], which also prompted
me to get as far as registering with one of the providers, although I
haven't yet followed through in using the thing...

Anyhow -

MS InfoCard uses URIs - didn't know that, good for them.

openid.sreg.* - my oh my, they are familiar

Last slide : "Use HTML profiles to ground OpenID link relationships in URI
space"

- spot on.

Back a few : "Are values of rel= all that different from tag names?"

Well, as you suggest, no. But it can be argued that its a constrained
extension point, a lot more controlled than arbitrary tag names. Narrowed to
this space and with obscure strings, it's difficult to counter Tantek's
Chicken Little argument on naming clashes.

So why should anyone use @profile? Your material on claims works, but
although the path through GRDDL to RDF works for RDF folks, I'm not sure how
well this would come across to folks that see URIs as merely part of those
clickable things that show in a browser.

Last slide again :

"Use GRDDL to look at OpenID links as RDF statements"

Works for me.

"Explore interaction between FOAF, hCard, XFN, and OpenID"

There may be a good leverage point for encouraging @profile in the latter 3.
I feel imagination-challenged on what's possible around OpenID and FOAF, but
I'm sure there's good stuff there.

It's not as much of a solution to world hunger as I thought it would
> be, for one thing...


Give a man a string and you have fed him characters for a day. Give him a
URI...

Cheers,
Danny.

[1] http://www.intertwingly.net/blog/2007/01/03/OpenID-for-non-SuperUsers




-- 

http://dannyayers.com