Minutes: XDI TC Telecon Thursday 8-9AM PT 2008-02-21

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

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

Date:  Thursday, 21 Febrary 2008 USA
Time:  8:00AM - 9:00PM Pacific Time

Event Description:
Weekly unofficial call of the XDI Technical Committee.

ATTENDING

Bill Barnhill
John Bradley
Victor Grey
Drummond Reed
Nat Sakimura
Markus Sabadello


1) CALL TIME

As Nat was able to join us, we had the opportunity to discuss the telecon
timing. Drummond mentioned that another international workgroup had
concluded that 9PM GMT (1PM Pacific) was the only time that worked for
Europe, the U.S., and Japan continents -- it's 6AM in Japan and 10PM in
Italy, but still allows us to all get on the same call together.

There was agreement that having a consolidated call would be worth it, and
Nat confirmed that 9PM GMT would work for him and the other NRI team member
he would like to have join TC soon. Bill also said the time would work
better for him as it would not conflict on alternating weeks with the OASIS
TAB telecon.

# DRUMMOND to send a message to the list about moving the telecon to 9PM GMT
(1PM Pacific) on Thursdays.


2) WORKGROUPS

In mailing list discussion of the minutes of last week's telecon, there has
been a general consensus that the TC should form workgroups to "divide and
conquer" the work we have to prepare for the XDI 1.0 specifications.
Drummond started the following wiki page to track proposals:

	http://wiki.oasis-open.org/xdi/WorkGroups

Discussion:

* Victor said he wanted to make sure that workgroups did not splinter the
work of the TC, since he wanted to stay involved in several of them. There
was consensus on this point.

* Drummond explained that there are two basic options: "heavyweight", which
would mean using formal OASIS subcommittees with their own chairs and
mailing lists, or "lightweight", which would mean self-organizing workgroups
that used wiki pages and threads on the main TC mailing list. There was
consensus that the lightweight workgroup option was preferable.

* There was also consensus that we should align workgroups with our
specification deliverables. 

* We discussed initial strawman proposals on the wiki page. The consensus
was that we should not try to produce one monolithic XDI Dictionary
specification, but instead produce small, moduler specifications covering
specific applications of XDI dictionary semantics. The SAML specification
suite was used as an example.

* Bill specifally advocated how XMPP has organized their specs. See:

	http://www.xmpp.org/extensions

* Bill pointed to XMPP extension proposal 211 is a way to roll up all the
extensions into different clients -- a bare-bones client, a mid-level
client, and a full-featured client. Bill feels that's the reason XMPP has
more adoption than SIP. 

* Further discussion produced the following initial outline of
specifications:

	1) Overview and Conformance
		* Basic XDI Client
		* Advanced XDI Client
	2) XDI RDF Graph Model ("Core")
		* Abstract API
		* Contexts
		* Serialization
	3) REST-based API
		* Messages
		* Operations
		* HTTP Binding
	4) Asynchronous API
	5) Dictionary
		* Typing
		* Comparison Operators
		* Transformations
	6) Versioning
	7) Link Contracts

* Bill suggested that we have an information document about known XDI
Implementations.

* John felt the modular approach would also make it easier to take current
implementation experience and feed it into the specific areas.

* Drummond said that if there was consensus on this set of deliverables, we
could finally update them in our long-awaited charter clarification.

# DRUMMOND to update http://wiki.oasis-open.org/xdi/WorkGroups to reflect
these notes. (DONE)


3) ATI MODEL DISCUSSION

There was a brief discussion of the direction of the ATI model. John
explained that ooTao, the company he is working with, has a commercial
implementation of the ATI model with a customer base.

* Bill said he'd like to have that work continue to be recognized by the TC.

* John pointed out that the ATI model also supports RDF.

* Drummond agrees strongly that "XDI RDF" is somewhat of a misnomer -- while
it is more directly based on the RDF graph model than ATI, both of them can
be transformed into conventional RDF.

* John suggests that if the two efforts stay together, we should emphasize
the similarities rather than the differences.

* Drummond acknowledged the depth of the ATI thinking and development and
said "the two genetic strains have 80% common genes".

* Markus asked for more information on the ATI model, as newer TC members
are not familiar with it.

# JOHN to post a wiki page on the ATI model that has a link to the key
resources.

* There was consensus that John should talk with Andy and Steve about what
direction they would like to take with the ATI model, and then we will
schedule this for further discussion next week.

# JOHN to develop proposed ATI direction with Andy and Steve.

# DRUMMOND to schedule this topic for next week's agenda.