url -> name

[Brad Fitzpatrick]

On Mon, 23 May 2005, Ben Hyde wrote:

> On May 23, 2005, at 3:41 PM, Imran Ghory wrote:
> > going over the protocol I notice there's a slight discrepancy in the
> > overall description and the protocol spec, the end result of the
> > protocol is verifying the user is say brad.livejournal.com but the end
> > result of the description is verifying the user is Brad from
> > LiveJournal.
> >
> > So how do you get from URL to username ?
>
> See if can get this right.
>
> All the spec does for you is tell you the url of a page this user
> controls.  The spec does not currently specify a way to get other
> attributes for that user.  That page be full of useful seeming
> information, for example it might link to a FOAF file.
>
> Is that right?

Correct.

> The mention of FOAF in the current spec is a vestigial organ, right?

The way it's mentioned in the spec is correct:  the helper on the
consumer's site can do whatever it wants... maybe it returns/parses a FOAF
or hCARD or XFN... whatever.  Once you know a user does indeed own a URL
tree, you can trust any document underneith it, including FOAF and such.
If their FOAF link is to a document outside of their owned URL tree, then
you shouldn't parse it.  If you do, you'll have to double-check it has a
"rel=me" pointer to something they own, but I'm not sure the exact syntax
for that in FOAF.

- Brad