Implementing YADIS with no new software

Drummond Reed drummond.reed at cordance.net
Mon Oct 31 14:54:30 PST 2005


>Ben wrote:
>Ok. I'm a bit confused here. From my understanding, multiple XRI specs
>define:
>1. XRI as URIs.
>2. A resolution protocol. That is, translates an XRI (possibly to another
>XRI) to a concrete URL.
>  2a. Part of which provides XML output (XRI Descriptors) for the
>resolution.
>3. Metadata that can be embedded into an XRI. Which the spec says "allows
>XRIs to be embedded within other XRIs in order to share identifiers across
>contexts".
>
>What are we talking about in the XSL case? To (2a) or (3)?

(2a)

>Secondly, what is the scope of the XRI and OpenID discussion? Are we
>thinking of adopting XRI resolution as a means of service discovery, or is
>it just to be used as Identification URIs?

Just using XRIDs as the service discovery description format. Nothing else.
That way both XRI resolution and OpenID/LID URL service discovery would use
the same format. The latter would disregard any XRI resolution info in an
XRID (which is only a few lines) and only use the capabilities description.

>And a bit off topic:
>My gut reaction to seeing XRIs is, "EDI X12 syntax ugliness." That is, I
>have to blink a few times in order to focus past the special characters and
>concepts of segments and subsegments.

You wouldn't be the first one ;-) But you'd be surprised how much power is
there given the few extra characters added to XRI syntax (which all works on
top of URI syntax). Many developers have had that first reaction and then
dug down a little and start to see entire continents of new functionality
that can be delivered with smarter identifiers and a simple XML resolution
protocol.

There's still not a simple document that explains the exact enhancements
that XRI syntax adds to URI syntax (and IRI syntax), but the XRI TC is
working on one. I hope that should make it much more approachable.

=Drummond 



More information about the yadis mailing list