Title: Some XRI schema questions
We discussed the id attribute and honestly
I don’t remember the conclusion. I know its necessary for SAML Trusted
Res, but not sure it’s required otherwise.
Given that most people aren’t using
it now, I think that’s good evidence that it shouldn’t be required.
(ie should be optional).
-Gabe
From: Drummond Reed
[mailto:]
Sent: Monday, January 22, 2007
9:25 PM
To: 'Barnhill, William';
Subject: RE: [xri] Some XRI schema
questions
Bill,
I was catching up on back email and
noticed this hadn’t received a response. See ### inline.
=Drummond
From: Barnhill,
William [mailto:]
Sent: Sunday, December 17, 2006
3:22 PM
To:
Subject: [xri] Some XRI schema
questions
Hi all. This weekend I finally got back to doing
some XRI coding, and was adding unit tests to XRI4R (hopefully soon to be
release Ruby library for XRI, XRD, and XDI manipulation). I had coded the items
in the schema that are xrd:URIPattern as Strings, and changed that to return a
URI by parsing them with URI.parse(el.text), since schema has them as
xs:anyURI. It raised exceptions as it should since the XRDs I'm working with
have XRIs in those elements not URI-friendly HXRIs.
Q1. Am I working with invalid XRDs or should there be a xrd:anyXRI and
accompanying xrd:XRIPattern?
### No, the elements that
contain XRIs use the xs:anyURI datatype. For this, as long as an XRI value has
the “xri:” prefix, it should work fine.
Q2. Also, in looking at schema again it seems that <XRD /> is a valid XRD
(other than @id, see Q4 below). Is that intended to signify an unresolvable
XRI? I thought that was done with a status code on XRD/Status.
### No, it’s
allowed via the schema. We wanted to do that as some uses of XRDs (outside of
XRI resolution) may wish to allow totally empty ones.
Q3. If there is a minimum set of elements (my guess would be query, status)?
### Not by the schema.
XRI resolution does specify that certain elements are required for certain
responses, but that’s application logic and not schema logic.
Q4. Schema seems to require existence of an id attribute on XRD, but no XRDs
I've seen have that. Outdated XRDs, or does id attribute need to be tagged with
use="optional"?
### Good question. I
think it should be optional. Gabe?
Thanks,
Bill
-----Original Message-----
From: Drummond Reed [mailto:]
Sent: Sat 12/16/2006 12:06 AM
To:
Subject: [xri] Minutes: XRI TC Telecon Thursday 2006-12-14
Following are the minutes for the unofficial XRI TC telecon at:
Date: Thursday, 14 December 2006 USA
(Friday morning Asia)
Time: 4:00PM - 5:30PM PT
Event Description:
Weekly unofficial call that will continue until the end of the XRI 2.0
cycle.
ATTENDING
Gabe Wachob
Bill Barnhill
Steve Churchill
Drummond Reed
Les Chasen
Laurie Rae
Marty Schleiff
Wil Tan
AGENDA
1) XRI $ DICTIONARY ISSUES
The editors gave a progress report and brought up some issues around the use
of code within the documentation.
# LAURIE to upload notes from last week's XRI $ Dictionary f2f attended by
Laurie, Gabe, Marty, and Drummond.
* In the V2.0 dictionary, code samples will be used only for documentation.
* Bill pointed out that if the code is executable in any form (therefore
becoming mobile code, then that would trigger special requirements. There
was consensus that we don't want to consider those requirements.
* We discussed the possibility of using regular expressions instead of
example code, but they are harder to work with.
# GABE AND LAURIE have the action item to determine which scripting or
programming language (e.g., Javascript or Perl) should be used within the
spec.
2) XRI SYNTAX ISSUES
Drummond reported on discussions he had at Internet Identity Workshop last
week with several TC members (Victor, Gabe, Laurie, Marty) concerning
current XRI 2.0 cycle outstanding issues. He had not had time to post any
proposals yet. They should be ready by next week's call.
http://wiki.oasis-open.org/xri/XriCd02/XriSyntax2dot1
http://wiki.oasis-open.org/xri/XriCd02/Xri2dot1Formats
3) XRI RESOLUTION ISSUES
We discussed the new issues (#37, #38, #39) at:
http://wiki.oasis-open.org/xri/Xri2Cd02/ResWorkingDraft11
* We had a very long discussion of issues #37 and #39, including the
question of whether URIs should be part of the XRI resolution graph, for
example as values of XRDS Ref elements or CanonicalIDs. There was no clear
conclusion, but the preference was not to extend the XRI resolution graph.
* On issue #37, we analyzed if an XRI resolution request for an XRI with a
local part, which specifies a return type of application/xrds+xml or
application/xrd+xml, could return an XRDS or XRD document describing the
resource identified by the local part and not an XRDS or XRD document
describing the authority. This is a classic "meta" problem because
you are
using an XRDS document to describe the retreival of an XRDS document.
* No clear solution was identified, however the XRI resolution editors will
think further on it and make a proposal via the wiki.
