Minutes: XDI TC Telecon Friday 2012-07-27

[Drummond Reed <drummond.reed@xdi.org>]

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

Date:  Friday, 20 July 2012 USA

Time:  10:00AM - 11:30AM Pacific Time (16:00-17:30 UTC)

ATTENDING

Bill Barnhill

Les Chasen

Markus Sabadello

Drummond Reed

Joseph Boyle

REGRETS

Phil Windley 

Giovanni Bartolomeo 

THE ETHERPAD LINK FOR TODAY IS:

     http://piratepad.ca/wQiz31Mbav

***** NEWS & UPDATES *****

--- MIKE SCHWARTZ

Per his message to the TC, Mike is stepping away from XDI for a period. He relayed in his last talk with Drummond that he intends to remain part of the community but needs to shift his development priorities at this time.

***** DECISION POINT QUEUE REVIEW *****

See the top of the wiki home page:

   https://wiki.oasis-open.org/xdi/FrontPage

We are beginning to make progress through the queue. A quick poll of the priorities of the TC members on the call:

• Bill: simple querying - the first step in full queries and the creation of a test suite (see below)

• Markus: next conponents of the API spec (see below)

• Drummond: the multiplicity proposal (already written up) and dictonary proposal (ready but not yet written up)

We briefly discussed the API spec components. The proposed components include:

• Simple HTTP transport (done_

• Message requirements

• Message operations

• Error messages

# MARKUS volunteered to begin writing these up.

***** DECISION POINTS FOR THIS CALL *****

This week's decision queue is the following set of proposals:

--- LITERAL CONTEXT NODES

   https://wiki.oasis-open.org/xdi/LiteralContextNodes

This proposal is foundational to the multiplicity and dictionary proposals.

--- JSON SERIALIZATION RULES

   https://wiki.oasis-open.org/xdi/JSONSerializationRules

This proposal aggregates all our deliberations about JSON serialization for the last 2.5 years.

--- LITERAL DATATYPING RULES

  https://wiki.oasis-open.org/xdi/LiteralDatatypeRules

This proposal documents how to uniformly identified literal datatypes in the graph.

On this proposal, Bill asked whether there would be any ambiguity when a literal was an XRI. Drummond explained why there was not. However Bill suggested that in the specs we should give very clear guidance that as a rule a literal should not be an XRI, because if it was, it should be an XDI context and not a literal.

The other TC members agreed that we should try to avoid the kinds of problems that catch developers simply because the spec does not adequately warn them to anticipate over avoid them.

DECISION: The above 3 proposals will be accepted after the standard two-week review period.

Bill also suggested that we create a single consolidated dictionary page of all XDI $ words.

# DRUMMOND to create such a page.

---- NON-NORMATIVE RULES

We did not cover the following two non-normative proposals:

   

   https://wiki.oasis-open.org/xdi/XdiDisplayFormat

   

   https://wiki.oasis-open.org/xdi/SerializationStatementOrder

# DRUMMOND to send a message to the list requesting review if TC members have time.

***** WIKI CLEANUP *****

We discussed and agreed on taking the following steps to complete the cleanup of the XDI TC wiki:

* Complete creating the Category auto-listing pages

* Create the CategoryFeature page

* Organize all proposals

* Put obselete proposals in a separate category

* Organized archived content so it is clearly not current content

***** IMPLEMENTATION-INDEPENDENT TEST SUITE *****

We discussed the topic of a test suite that can work against all implementations and is completely vendor- and project-neutral.

The TC members widely acknowledged and thanked Mike Schwartz and the OpenXDI development team for their work on the OX test suite, which blazed the trail in this area.

IWe agreed a copy of the test suite must live at OASIS. But Bill proposed that the tests be separate from the test harness. The tests themselves could consist entirely of a set of XDI documents.

Each test should consist of:

• The input message

• The associated metadata about the input message

• The set of queries and the expected results for each query that the response message must pass (similar to what Schematron does for XML documents using XPath queries)

We agreed we don't currently have enough bandwidth on the TC to create this test suite, but we also agreed it should be a goal of the TC.

***** NEXT CALL *****

The next call is next week at the regular time.