Minutes: XRI TC Telecon 2-3PM PT Thursday 2009-03-12

["Drummond Reed" <drummond.reed@cordance.net>]

Following are the minutes of the unofficial telecon of the XRI TC at:

Date:  Thursday, 12 March 2009 USA
Time:  2:00PM - 3:00PM Pacific Time (22:00-23:00 UTC)

ATTENDING

John Bradley
Brian Eaton
Will Norris
Eran Hammer-Lahav
Drummond Reed
Nick Nicholas
George Fletcher
Nat Sakimura

REGRETS

Markus Sabadello 
Nick Nicholas


1) XRD 1.0 - LINK HEADERS/HOST META/LRDD

	http://tools.ietf.org/html/draft-nottingham-http-link-header
	http://tools.ietf.org/html/draft-nottingham-site-meta 
	http://tools.ietf.org/html/draft-hammer-discovery

Eran said there has been no further feedback yet; drafts are currently
frozen for the upcoming IETF meeting.


2) XRD 1.0 - SCHEMA

	http://wiki.oasis-open.org/xri/XrdOne/XrdSchema
	http://lists.oasis-open.org/archives/xri/200903/msg00022.html 

Our main topic was discussing the questions Will posed (and Drummond started
to answer) in the second link above.

Questions:

	A) How closely is <XRD:Link> is supposed to map to Link-Header?

Eran: individual child elements of <XRD:Link> have the same semantic meaning
as Link-Header. For example, with Link-Header, you can have separate
Link-Headers all with the same rel attribute value.

As far as XRD goes, having multiple <XRD:Link> elements with the same
<XRD:Link:Rel> element value but different <XRD:Link:URI> element values has
the same semantic meaning as having one <XRD:Link> element with one
<XRD:Link:Rel> value and multiple <XRD:Link:URI> element values.

	B) Cardinality

The cardinality of <XRD:Link:URI> and <XRD:Link:URITemplate> are zero or
more.

	C) <XRD:Link:URI> and <XRD:Link:URITemplate>

The two elements MAY be combined under one <XRD:Link> parent.  Eran
explained that there are use cases where you could have both a "hard-coded"
<XRD:Link:URI> element and a "soft" <XRD:Link:URITemplate> element.

You can also think of the <XRD:Link:URI> element as a <XRD:Link:URI>
template without any parameters.

There was consensus that in this case, it is essential that the priority
attribute work across all instances of both.

We discussed whether the two elements could be collapsed into one with a
template attribute whose value indicates the template type. Eran prefers
keeping them in separate elements.

Will felt that it would be relatively easy to implement priority across
multiple <XRD:Link:URI> and <XRD:Link:URITemplate> elements.

We also discussed the layering of the functions of the discovery library
such as the one Will is building. Eran pointed out that it's important for a
discovery library API to support multiple <XRD:Link:Rel> elements in the
same <XRD:Link>.

	D) URITemplate Pattern

Eran thinks we should use the simple one from LRDD. We had a long discussion
about the URI templating mechanism. 

Eran believes that an application will first ask an XRD resolver to obtain
an XRD or a specific Link, and then if it receives a URITemplate in that
link, the application will call a template library and supply the template
and a dictionary. The dictionary is where the input URI has been parsed into
the component pieces for the dictionary.

This same function applies to XRIs, which are designed for that kind of
decomposition.

	E) Implied Relationships and Semantic Equivalence

Two or more <XRD:Link> elements that contain different <XRD:Link:Rel> child
elements are only semantically equivalent if the <XRD:Link:URI> or
<XRD:Link:URITemplate> child elements are identical.

Three <XRD:Link> elements with three different <XRD:Link:Rel> elements but
identical <XRD:Link:URI> elements can be collapsed into one <XRD:Link>
element with three different <XRD:Link:Rel> elements and one <XRD:Link:URI>
element.


	F) XSD and RelaxNG Schemas

Will asked if the TC had produced XSD and/or RelaxNG schema definitions yet.
We have not. Drummond pointed out that although XRI Resolution 2.0 published
schema definitions in both, it specified RelaxNG as the authoritative
format. There was concensus we should do the same.


3) XRD 1.0 - TRUST TEAM

	http://wiki.oasis-open.org/xri/XrdOne/XmlDsigProfile 

Brian reported there are no open issues outside of the questions that have
been discussed on the list and that Dirk is proceeding with an
implementation as part of the Step2 open source project to produce an
OpenID/OAuth hybrid.

John and Drummond have not done any further work on a Synonym Binding
Profile yet, and will likely not until the XRI 3.0 Syntax and Bindings
specification Working Drafts are posted.


4) XRI 3.0 - SYNTAX AND BINDINGS UPDATE

Drummond reported that he had reviewed Nick's first draft of the info:
binding, and also Nick's proposed info: subscheme registration. He proposed
that both XRI bindings (http(s): and info:) use the same section names, so
the next action is to develop those.

The timing on when to submit the info: subscheme registration is still TDB.


5) NEXT CALL

Regular time next Thursday 3/19, 2-3PM PT (22:00-23:00 UTC). NOTE THAT THE
US HAS SWITCHED TO DAYLIGHT SAVINGS TIME.