Re: [xdi] Minutes: XDI TC Telecon Thursday 2007-01-04

[laurie.rae@cordance.net]

[So here we go. First of many minutes. Please let me know if these are too
detailed or if I missed anyone.]

Following are the minutes for the joint unofficial telecon of the XRI and
XDI TCs at:

Date:  Thursday, 2nd February January 2007 USA (Friday morning Asia)
Time:  4:00PM - 5:30PM PT

ATTENDING

Gabe Wachob
Drummond Reed
Laurie Rae
Steven Churchill
Marty Schleiff
Bill Barnhill
Paul Biciunas


AGENDA


1) PROPOSAL TO APPOINT NEW SECRETARY

It was agreed that Laurie Rae will take over responsibility as the
Secretary for both the XRI and XDI TCs. She will distribute weekly minutes
and agendas for the unofficial joint XRI XDI TC meetings.



2) XRI FORMS AND TRANSFORMATIONS LADDER

We agreed to the changes proposed by Marty and Drummond at:

        http://wiki.oasis-open.org/xri/XriCd02/FormsAndTransformations

		* Action Item:The rest of the steps for XRI-Normal Form need to be
		completed by Marty based on the rules currently listed in the
		Canonical Form section of XriCd02/Xri2dot1Formats. These should
		also include the new rules in XriCd02/CompactSyntax.



3) XRI SYNTAX 2.1 COMPACT SYNTAX PROPOSAL

There was extensive discussion about the following proposal:

            http://wiki.oasis-open.org/xri/XriCd02/CompactSyntax

The discussion on this item continued until 5:20, when we decided to
review the RDFX proposal (see below).

Key Points:

	* Questions were raised as to whether we should have optional syntax
	at all.

	* Many questions were raised around nested cross-references.
	Does the compact syntax mean that certain nested x-refs are not
	supported? The compact representation of nested x-refs seems problematic,
	especially wrt to trying to "uncompact" a nested x-ref.

	* Stephen raised a concern that dropping the delegation symbol is too
	extreme. Why not have =example*+blog? He agreed that this was a trivial
	issue for machine processing, but for human readers, it requires a
	context switch.

	* Gabe added a problems section to the wiki page. Everyone added their
	own concerns to the list as the meeting progressed. The detailed list is
	provided at the end of this document.

Outcome: It was agreed that the proposal comprises two separate proposals:
1) a proposal about compact form 2) a proposal that compact form be normal.
Further discussion is required.


4) XDI RDFX PROPOSAL

We did not have a lot of time for this item, but extended the meeting by
another half hour to review the proposal.

Key Points:

	* Bill Barnhill sent a proposal for "X3" notation.

	* Concerns were raised that we were inventing a new syntax for RDF?
	It was suggested that we look into TriX and other existing serializations.
	It was argued that the addressing syntax in a certain way simplifies by
	using a single xri for representing any resource.

	* It was mentioned that the uris in RDF don't identify instances of that
	use of a predicate in a statement. That is, the identifier isn't reifying
	the use of the predicate.

	* It was also suggested that we use ($IsA), ($HasA) to express the
	relationships.

	* It was also suggested that with RDFX, you wouldn't be able to have
	anonymous nodes, which is something that RDF doesn't provide.

	* It was also suggested that we consider using the data URL scheme. It was
	agreed that this required further consideration.

	Action Item: Spend some time considering how the RDF Reification
	vocabulary might be used instead.
	We will also explore an RDF XML serialization of the proposed RDFX Model.

	Action Item: An updated version of the proposal will be discussed with
	Mark Wahl on Wednesday. The results of this discussion will be covered
	during the unofficial



Appendix A) Problems posted to
http://wiki.oasis-open.org/xri/XriCd02/CompactSyntax:

	=example*(+tag1*(+tag2)) becomes the same as =example*(+tag1)*(+tag2)

	this implies I cannot extract (+tag1*(+tag2)) as a unit, but

	=example*(+tag1*( http://foo.com)) is not the same as =example*(+tag1)*
	( http://foo.com)

	So, the ability to extract (+tag1*( http://foo.com)) as a unit is
	preserved, but I cannot extract (+tag1*(+tag2)) as a unit

	Double - xref: =example*((+tag1*(+tag2))) ? (Gabe)

		Solution proposed by Drummond to this problem: Have the delimiter
		apply until have the delimiter apply until you reach the next
		parameter

	Does =example.name*(tag1)*(tag2) become =example.name(tag1)(tag2) ?
	 (Marty)

		Suggestion: the better options is to use double parens.
		This is still under discussion.

	So the =example.name*(+foo)*(+bar) first becomes
	 =example.name(+foo)(+bar), and then becomes =example.name+foo+bar? (Bill)

	How about =example!(+foo) ? (Steve)

		Decision: You always have to keep !. Therefore, delegation
		characters can only be eliminated in the non-persistent case.

	if "+tag1*(+tag2)" becomes "+tag1+tag2" then "=example*(+tag1*(+tag2))"
	 should equal "=example*(+tag1+tag2)" and should equal
	"=example+tag1+tag2", so I don't think you can logically say you can't
	compact an xref containing another xref. (Marty)

	Issue: Shouldn't it be changed so that the outer parentheses disappear?