Minutes:Joint XRI & XDI TC Telecon Thursday 2007-08-16

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

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

Date:  Thursday, 16 August 2007 USA
Time:  10:00AM - 12:00PM Pacific Time

Event Description:
Weekly unofficial joint call of the XRI and XDI Technical Committees.

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

PRESENT

Gabe Wachob
Wil Tan
Les Chasen
Drummond Reed
Victor Grey
Peter Davis


AGENDA

1) TREATMENT OF SEMICOLON FOR MEDIA TYPE PARAMETERS IN PROXY RESOLUTION

This issue is summarized in the following message to the XRI list:

	http://lists.oasis-open.org/archives/xri/200708/msg00013.html

The consensus was to specify that the semicolon needed to separate media
type parameters included in HXRI query parameters MUST be percent-encoded.


2) NODEFAULT PARAMETER & MULTI-VALUED MEDIA TYPE PARAMETERS

The current design is to use the nodefault parameter multiple times. The
consensus is to change this to separate Boolean parameters called
nodefault_t, nodefault_p, and nodefault_m.


3) BOOLEAN PARAMETERS

All boolean parameters will accept "true" or "1", and "true" is
case-insensitive. ANY OTHER VALUE IS FALSE.


4) MISSING ELEMENT IN SEP SELECTION PSEUDOCODE

In implementing an XRI proxy resolver in Ruby, Victor found one missing
element in the SEP selection pseudocode -- a nodefault check is needed in
the Match.x loop. Drummond will fix.


5) SYNONYMS IN ED04

A new proposal was posted for discussion on this issue:

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

This wiki page was updated with additional information and feedback shortly
after the call, so it reflects the discussion on the call. The main points
of the discussion were:

* There is concensus that all claims of equivalence in an XRD are just
claims, and the scope of verification of identifier claims by XRI resolution
infrastructure is limited to at most correlation of identifiers across XRDs.

* There is consensus that a CanonicalID should be the persistent global
foreign key for a resource in the context of the authority that assigns it,
and thus that it should never need to change.
* There is consensus that Canonical ID verification is needed for security
purposes in order to detect roque XRI authorities and prevent spoofing of
CanonicalIDs.
* The new proposal reverts to the WD10 definitions and usages of LocalID,
CanonicalID, and Ref.
* The new proposal actually allows Refs to carry even less semantics, and
only assert "resolution equivalence" instead of "identification
equivalence". Refs only indicate that services may be shared across XRDs.
* The new proposal uses EquivID, MapToID, and MapFromID elements to handle
all types identification equivalence assertions.

See the wiki page for further information. The action item is for all TC
members, and especially XRI Resolution 2.0 editors, to review and send in
any feedback ASAP, as the updated proposal will go into ED04 next week. 


6) OTHER FEEDBACK ON ED03

We stepped through the remaining open issues as listed at:

	http://wiki.oasis-open.org/xri/Xri2Cd02/ResWorkingDraft11

The conclusions reached were:

* Issue #33 (Warning about append= attribute): This proposed warning (about
the use of the append= attribute in URI elements) is now obviated by the
conclusion on issue #34, below.

### DRUMMOND to add text in ED04 pointing out that consuming applications
SHOULD use the uric=true subparameter if they do not want to deal with HXRI
parsing.

* Issue #34 (Proposed uric= attribute): There was consensus to add support
for the uric=true attribute as it lessens the work necessary by consuming
applications to use XRDs and makes it easier for service providers to use
the append= attribute.

### DRUMMOND to add this to ED04.

* Issue #44 (Path matching case sensitivity): After a long discussion of the
pros and cons, the consensus was to maintain backwards compatability and
have path matching be case insensitive as it is with authority matching.

### DRUMMOND to note this to ED04.

### WIL to send DRUMMOND examples of path matches to include at the end of
section 10.3.7.

That concludes all open issues.


7) ASSIGNMENTS AND SCHEDULE FOR ED04

We reviewed editor assignments for ED04 (listed on
http://wiki.oasis-open.org/xri/Xri2Cd02/ResWorkingDraft11). We agreed that
the final step of updating the XRD schema and the reference API in Appendix
H should wait until ED05, which will be week after next. So the goal is
produce ED04 with all other action items done and issues address by next
week's call.

### WIL to send his action items to DRUMMOND.

### DRUMMOND to compile a list of Gabe's action items and send to him.