RE: [xri-comment] XRI Resolution

["Wachob, Gabe" <gwachob@visa.com>]

Loren-
	Let me respond inline. Its GREAT to hear about this implementation effort!!

	Forwarded to the xri main discussion list as well. Hope you don't mind.
<snip>

> It seems that the thing embodied by the <IdentifierAuthority> 
> may not be an instance of an Identifier Authority as defined 
> in the glossary.
> 
> It MAY be an Identifier Authority, or MAY be a local access 
> endpoint, or BOTH.  It seems that when you ask for resolution, 
> you're asking to resolve an XRI Segment.  What you get back is 
> information about that Segment.  If the Segment represents
> an Identifier Authority, you get information on how to 
> communicate with it.  If the Segment represents a local
> endpoint, you get information on how to communicate with
> the endpoint.
> 
> Unless I'm misunderstanding something, I'd like to recommend
> we change <IdentifierAuthority> to something else.
> ResolvedSegment, ResolvedXRI, SegmentIdentifier, or something
> closer to what the instance represents.

Your intent is correct, and I understand the confusion. I would be happy with "ResolvedSegment" or something to that effect, because thats really what we are talking about (well, ResolvedIdentifierAuthoritySegment but thats a big pain ;-)


> ALSO:
> 
> If the XRI segment DOES represent an identifier authority, 
> it looks like you know that by the presence of <NextAuthority>. 

Well, remember that the XRI segment can be BOTH!! So its not an either or - the presence of the NextAuthority element merely means that if you want to continue with Authority Resolution (ie you have more "dot" or "colon" segments), you can. For example, if you have xri://@a.b.c.d and you've resolved @a.b.c, you would expect there to be a <NextAuthority> element - if there weren't, that would be essentially a "unresolvable identifier authority" error for ".d" (I made up that error, btw. We don't specify errors explicitly). 

> When reading the definition of <NextAuthority>, it doesn't seem 
> to describe the NEXT authority, it seems to describe THIS authority 
> (or how to communicate with the authority represented by this
> Segment).

OK, I see your point of view. It really describes *this* thing as a IdentifierAuthority (as opposed to the thing identified by the next subsegment). 


> If I understand this element correctly, it seems like it
> should be labeled <IdentifierAuthority> vs. <NextAuthority>.

Agreed. 
 
> Example:  @foo.bar - When requesting resolution for ".foo", you
> are asking "@" (the Identifier Authority for .foo).  You might 
> get back the following:
> 
> <ResolvedSegment xmlns="xri:$r.s/ResolvedSegment">
>     <Resolved>.foo</Resolved>
>     <IdentifierAuthority>
>         <URI>http://resolver.foo.com</URI>
>         <URI>https://resolver.foo.com</URI>
>     </IdentifierAuthority>
>     <LocalAccess>
>         <Service>xri:$r.a/X2R</Service>
>         <Type>application/rddl+xml</Type>
>         <URI>http://xri.foo.com</URI>
>     </LocalAccess>
>     <LocalAccess>
>         <Service>xri:$r.a/X2R</Service>
>         <Type>image/jpeg</Type>
>         <URI>http://pictures.xri.foo.com</URI>
>     </LocalAccess>
>     <Mapping>xri:@:1:2:3</Mapping>
> </ResolvedSegment>
> 
> I now know how to ask @foo to resolve ".bar", and I 
> also know how to ask @foo to get some jpeg pictures.
> @foo is both an identifier authority (for .bar) and
> an endpoint for accessing resources identified by XRI.
> 
> Apologies for the long mail.  I wanted to provide enough
> information to see if I understand the resolution process.

You make good points. The question we have to tackle now is whether or not we should fix these things and have a revote. I'm inclined to say we should. But that requires rev'ing the spec (1.01?) and revoting. Drummond???

We can do this in very short order. I think we had hoped that feedback like yours would come back before the beginning of our vote, but I suspected we'd get good feedback like this more than a few days after thanksgiving ;-)

	-Gabe