Minutes: XDI TC Call Wed 10/5 9AM-10AM Pacific 12-1PM Eastern

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

Following are the minutes of today's XDI TC telecon. Action items marked #.

ATTENDING
Victor
Andy
Peter
Owen
Steve
Drummond
Jason

1) BRIEF XRI 2.0 SPEC STATUS
Drummond gave a brief report that XRI Syntax 2.0 Committee Draft 02 is
expected to go to a vote on Oct. 14. XRI Metadata 2.0 Committee Draft 02
should be ready shortly thereafter as the only proposed addition is a $t
space for identifier types. The biggest chunk of work is on XRI Resolution
2.0 Committee Draft 02 as it needs to fully incorporate HTTP proxy
resolution, which may require need another month or two of work. The goal is
to have it at Committee Draft 02 status by Dec. 1.

2) CHARTER REVISION
It was agreed that the priority for Working Drafts remains 1) XDI Schema and
Addressing, 2) XDI Protocol & Bindings, 3) XDI Service Dictionary.

# Drummond has the action item to produce a revision of the XDI TC charter
and send it to the list for comment. 

We discussed the need for implementers to begin to reference XDI service
endpoints in XRI Descriptors (XRIDs). There were several proposals for how
this might work while the XDI specifications are in Working Draft status.

# Andy will send a proposal to the list to begin a discussion thread.

2A) POLICY REFERENCES
Victor also asked a question about how policies and policy references will
work and how they can be made extremely simple. Drummond explained that Doc
Searls had been given the action item at a Berkman-sponsored meeting on the
identity metasystem at Stanford last week to make a proposal for how a
"Creative Commons" style approach could be taken to policy usability for
identity and data sharing transactions. This fits perfectly with the ability
for XDI to use policy references rather than requiring actual
machine-readable policy definitions in XDI link contracts.

There was consensus that this approach would be the fastest route to early
adoption.

Andy clarified that even if policy references are used, XDI link contracts
will still need a handful of other standard resources, such as the date, the
effective period, signature, etc. Victor agreed. These need to be defined in
the XDI Service Dictionary.

3) PRIORITY ATTRIBUTE DISCUSSION
Drummond explained that work on the issue of element priority in the XRI
Resolution 2.0 specification raised the question of whether the XDI TC
should also incorporate the same "priority" attribute in the XDI schema.
Specifically, if a Resource or Link has two or more XRI elements, should the
XDI document author be able to include a priority attribute that gives a
consuming application guidance on which synonym has top priority? Or should
this be handled in a more XDI-model fashion, meaning using a $priority
resource.

The discussion was that while this might be a good addition, there is not
yet an explicit use case for it, so let's not add it until we need to.

Steve brought up that he also anticipates the need to explicitly declare the
order of multiple child elements (specifically the Resource children of a
Resource element). Drummond said the same issue had come up a year ago and
the pattern that had been suggested was the use of a $order element.
Potentially this element could be an addition to the XRI Metadata
specification, patterned on the $v entry.

4) SIMPLE SCHEMA IMPLICATIONS
Drummond raised this topic to point out that the current XDI schema proposal
(http://www.oasis-open.org/committees/download.php/14620/draft-xdi-schema-v1
3..xsd) makes the structure and navigation of XDI documents significantly
simpler. It does this partially because it loosens some of the
constraints on a strict three-level model (authority, type, instance).
Specifically, it does not inherently constrain a XDI type to be one level
(the way authority is inherently constrained by XRI syntax to one
level/segment), but instead relies on a pure dictionary model (i.e., a
cross-reference to a type definition) to define the relationship of a type
and an instance.

We discussed this using a sample XDI document in the v13 schema (see
http://lists.oasis-open.org/archives/xdi/200510/msg00001.html). It was
agreed that this increased flexiblity was an asset - Andy compared it to the
flexibility of using normalized vs. non-normalized databases.

*** END ***