Minutes of unofficial XRI TC Telecon Thurs 2/24

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

XRI TC Members and Observers,

We had an excellent call today reviewing all four of the posted
specification drafts (Syntax O4, Resolution 04, Metadata 02, Primer 01).
Minutes follow below. We remain on track on our goal to hold the Committee
Draft vote on Thursday, March 10, at 2pm Pacific (Friday, March 11, 7am
Japan).

We will be holding one more scheduled review call next Thursday, March 3, at
that same time. Updated drafts will be posted regularly through this period,
so please check the mailing list for the latest draft prior to your review.

=Drummond 

*** MINUTES OF 2/24 INFORMAL XRI TC TELECON ***

Attending:

Dave McAlpin
Peter Davis
Gabe Wachob
Drummond Reed
Nat Sakimura
Mike Lindelsee (by IM proxy via Gabe)
Guest: Dallas Davam


SYNTAX

Dave reported that only pending changes that are not purely editorial or
"instructions to editor" were:

1) Checking the assertion in Section 2.5 the normalization and comparison
rules for URIs and IRIs also applies to IRIs.

2) Adding a section 2.4.7 titled "Leading Segments Beginning with a !
Symbol".

Mike has the baton and will hand it back to Dave who will post it tonight.

RESOLUTION

Dave has made extensive comments on the O4 draft. He is almost finished with
his review. He will hand the baton on his edited draft to Drummond by the
end of the day, who will return it to Gabe before Monday morning. Peter is
part way through his review and will send it to Gabe before Monday morning.
Gabe will then issue the O5 draft as soon as he can process all the
comments.

The issues discussed on the call included:

* "Resolved" element - it was agreed to make the cardinality "1" instead of
"0-N" because this simplifies behaviour.

* Returning HTTP Errors - there was a long discussion about why, in
lookahead resolution, HTTP error codes are returned along with a successful
result (when one subsegment resolves but there is an error later in the
chain). Peter's feedback was that this approach was unconventional, and Gabe
and Dave agreed with him, explaining that was their first reaction too.
However they had researched it and are satisifed that it is fully compliant
with the HTTP spec, and it is better to notify the client of an error in
lookahead than force the client to get the same error. Gabe will add some
text to the spec to explain why this design decision was made.

* DNS Names vs. IP Addresses for IRI Authorites - Dave brought up that an
authority may be contacted by an IP address. Gabe pointed out that in HTTP
1.1, there is a header element for passing on the DNS name, and he will mark
this required.

* Section 2.4.2.2 - Resolution Types - Dave felt that this section (which
was an appendix in 1.0) should not be included. It was agreed to remove it.
Drummond will proof on his review to see if a footnote is warranted.

* Proxied resolution - Dave felt that proxied resolution should not require
full resolution of the authority segment. Gabe agreed.

* Gabe also noted that he will fix the version number.

METADATA

Dave is planning his review tomorrow. Drummond asked for all reviewers to
complete their feedback by next Wednesday, so he can issue the proposed
release candidate by next Thursday's call.

PRIMER

Dave had a suggestion for improving the key graphic illustrating the uniform
abstract identification layer concept.

We reviewed the problem section, and Dave suggested that we add one more
about using XRIDs for "bootstrap discovery" of some basic attributes to
facilitate further interaction. Drummond will add this.

Mike is working on additional text. Drummond and Dave will collaborate on
the usage scenarios early next week so a complete draft is ready by next
week's review call.