Minutes of Special XRI TC Telecon 1PM PT Wednesday 2007-05-09

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

Following are the minutes of today's special XRI TC telecon (see
announcement message appended below).

ATTENDING

Marty Schleiff
Steve Churchill
Peter Davis
Drummond Reed
Victor Grey
Les Chasen

* = discussion point
# = action item


1) DIRECT CONCATENATION

* We reviewed the new wiki page describing this proposed feature of XRI
Syntax 2.1.

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

* Peter suggested that this page could use some more detail in explaining
how subsegments are all "integral units" now, i.e., how under the Draft 14
ABNF on http://wiki.oasis-open.org/xri/XriCd02/XriAbnf2dot1, the fundamental
unit is "subseg", which is not further parsed. We agreed we need more
documentation of this key proposed feature of the ABNF.


2) SOCIAL VULNERABILITY OF =/@ CONCATENATION

* Steve summarized the evolution of discussions on this point, which served
as a good introduction to the options on the new wiki page:

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

* Steve broadly characterized the 8 options on this page as:

	#1: Prevent =/@ concatenation at the syntax level
	#2: Allow it as the syntax level but don't allow it in resolution
	#3: Use a special non-standard resolution w/out reference
verification
	#4-#8: Perform standard resolution w/a form of reference
verification.

* Steve and Drummond prefer at least #2 or higher since the Draft 14 syntax
proposal is so clean and elegant an abstract model.

* Les feels that option #1 should still be considered if we don't come up
with a solution to the social vulnerability issue.

* We talked about the proposed "grammatical" difference between a global
xref. In short, it is that a global xref like "=drummond" should be
interpreted as "THE =drummond", i.e., the same resource referenced by the
global XRI independent of the context, while a local xref like
"*(=drummond)" or "!(=drummond)" should be intepreted as "A =drummond",
i.e., a local resource that uses this identifier but is not necessarily the
same resource as identified in an independent global context. Les likes this
clear distinction; Marty expressed concerns about it. 

# DRUMMOND (who understands Marty's concerns) to propose a slight revision
to this "grammatical rule" intended to satisfy both Les and Marty.

* We discussed option #3 and clarified that it does not perform any type of
reference verification; it simply resolves the child subsegment
independently of the parent. Steve and Drummond expressed concern that this
breaks the standard left-to-right resolution model, and does not actually
confirm any form of relationship at all between the two authorities.

* We did not have time to discuss options #4-#8 in detail, but in summary
form they all perform standard resolution with a type of reference
verification. 

* Option #8 stands out because it proposes a standard algorithm for
reference verification, dubbed "cross-resolution", that in practice works
very much like CanonicalID verification. It will work for any XRI authority
segment containing global subsegments and scales to all sizes of registries
and communities.

* Peter asked for this cross-resolution model to be documented in greater
detail, including sample XRDs, so it would be easier to analyze and discuss.

# DRUMMOND (and any other editor who wants to help) to document
cross-resolution in greater detail.

***********

-----Original Message-----
From: Drummond Reed [mailto:drummond.reed@cordance.net] 
Sent: Monday, May 07, 2007 1:46 PM
To: xri@lists.oasis-open.org
Subject: [xri] Special XRI TC Telecon 1PM PT Wednesday 2007-05-09

XRI TC Members:

As the XRI Syntax editors try to close the last key issues remaining before
producing XRI Syntax 2.1, we will be holding a special telecon open to all
members of the TC this Wednesday at 1PM Pacific time. This call will be in
addition to our normal call on Thursdays at 10AM PT.

We will send out an additional email with an agenda and wiki links before
the telecon, but we want to invite all TC members to put in on their
calendars.

We will use the standard TC telecon number (thanks to NeuStar for hosting
this):

TO ACCESS THE AUDIO CONFERENCE:
    Dial In Number: 571-434-5750
    Conference ID: 5474

=Drummond