Minutes of Encoding Task Force, 30 October 2001.

Current list of action items
-- Paul to setup next telcon for tuesday Nov 6, 6-7p Nice/France time
-- Asir to write up algorithm (for issue #1), by friday
-- Asir to post revised algorithm to dist-app (assuming no show stopper
issues raised by ETF members) monday afternoon
-- David to add dist-app-posted algorithm to XMLP WG agenda for
consideration at their Nov 7 telcon
-- All: Evaluate Jacek's proposed resolution issue #18, see [2] (in time
for next telcon)
-- All: discuss when/how/why use Schema to help resolution of issue #17
-- Asir to summarise "sparse arrays" email thread to identify any new
issues, by thursday Nov 1
-- All: review Jacek's proposed resolution for issue #144 [4] (in time for
next telcon)
-- Jacek to ask DanBr/RDF and BarbaraP/XMI (and cc dist-app) whether his
proposed text for resolution of issue #29 satisfies their requirements,
this time to set a deadline of close-of-business monday, Nov 5
-- David to start prioritising ETF issues

Asir Vendamuthu
Marc Hadley
Paul Cotton
David Fallside (scribe)
Jacek Kopecky

1. Review of action items
-- we have not heard anything back from RDF nor UML/XMI, decided to resend
request for comments with a deadline of close-of-business monday Nov 5

2. Discussion of "full/partial algorithm" [3], see Jacek's summary email
[1] and Asir's suggestion [2] for background
-- general agreement that neither the partial algorithm nor the full
algorithm are entirely suitable
-- general agreement on using the "full" algorithm (but without xml
escaping) applied to the localName part of a prefix:localName namespace,
and  where the constraints on prefix are as specified in XML Namespaces
(normative reference)

3. Discussion on issue #18, see Jacek's summary [4]
Point 1, whether or not to disallow forward references (fRefs).
-- Marc indicated his developers are in favour of fRefs. Asir indicated his
developers are against fRefs  because they require two-pass processing by
the receiver.
-- This split is typical of what Jacek observed in the SOAPBuilders.
-- Paul noted that the status quo is to allow fRefs, and disallowing them
now would result in backward incompatibility
-- Asir said this could be handled by versioning, Paul thought this a very
weak argument
-- David noted that the ability of a receiver to handle forward refs,
perhaps due to footprint etc, was something that should be understood
through some out-of-band mechanism, and senders who want to use fRefs
should not be prevented from doing so
-- Asir is going to take this issue back to his developers for further
consideration. We will take it up in the next ETF telcon, but if there is
no resolution then, we will take it to the WG.

[1] http://lists.w3.org/Archives/Public/xml-dist-app/2001Oct/0231.html
[2] http://lists.w3.org/Archives/Member/w3c-archive/2001Oct/0071.html
[3] ftp://sqlstandards.org/SC32/WG3/Progression_Documents/Informal_working_drafts/wd-xml-2001-06.pdf
[4] http://lists.w3.org/Archives/Public/xml-dist-app/2001Oct/0231.html