W3C

XML Processing Model WG

13 Sep 2007

Agenda

See also: IRC log

Attendees

Present
Andrew Fang, Paul Grosso, Rui Lopes, Alex Milowski, Michael Sperberg-McQueen, Henry Thompson (chair pro tem), Richard Tobin, Allesandro Vernet, Mohamed Zergaoui
Regrets
Norm Walsh
Chair
Henry S. Thompson
Scribe
Henry S. Thompson

Contents


Rollcall

As above.

Agenda

HT: Accepted as published

Next meeting

HT: We will meet next in two weeks, provided we get to Last Call this week

Minutes

HT: Comments on these minutes: http://www.w3.org/XML/XProc/2007/09/06-minutes.html
... Approved as they stand

Comments on the draft of 11 September

http://www.w3.org/XML/XProc/docs/langspec.html

AM: The appendix isn't there yet

HT: True, but as it's non-normative, it can be added later

AM: I have a draft for part of it, we could add it right away

PG: I'd rather not do that, let's get the LC draft out, and add that in a subsequent draft when it's complete. There's a time issue here, with the Tech Plenary coming up
... No objection to the idea of the appendix at all.

AM: Consensus was that we would have this appendix

Proceed to Last Call?

HT: Straw poll on 3 options:
... 1) Publish ASAP w/o any appendix
... 2) Publish same time with whatever Alex can supply by the time Norm needs it
... 3) Hold publication for agreed complete appendix

PG: We could publish as is for last call, and publish a separate WG note asap

AM: We can get the text I've written already in in just a few minutes

AV: Sounds like option 2 is what you want

<alexmilowski> Here's the text: http://lists.w3.org/Archives/Public/public-xml-processing-model-wg/2007Sep/0053.html

<alexmilowski> Minus the "general" bit

PG: (2) ; HT: (2); AM: (2) ; AV (2) ; RL (2) ; RT (2) ; AG (2) ; MZ (2)

HT: Unanimous straw poll result

Any objection to the editors being authorized to publish the 11 September draft as a public Last Call WD with the addition of a non-normative appendix giving guidance on Namespace fixup to the extent possible w/o delaying publication?

RESOLUTION: to publish the 11 September draft as a public Last Call WD with the addition of a non-normative appendix giving guidance on Namespace fixup to the extent possible w/o delaying publication.

Split the spec?

HT: Discussion -- could do it later, not a substantive question

AM: Prefer to keep it as one document, easier right now, and easier for consumers down the road

MSM: People say it's not substantive, but it does affect something crucial, namely the ability to say that you conform to the spec.
... If we split the spec., and version the parts separately, will people end up having to say "conforms to 1.n of the spec and 1.m of the library?"
... Also, splitting would make the framework very abstract, or we need to allow ourselves to refer to examples in the library 1.0
... Does the library of steps make sense outside the context of the XProc framework?

HT: Anybody prepared to argue in favour?

RESOLUTION: We will not split the spec. before going to Last Call

Test cases

HT: Some discussion about where they are going to come from has happened in email.
... Where is the energy going to come from for managing test collection?

RT: Implementors will produce tests
... Lets wait and see what they look like, and if we can put them into a framework

MSM: Would a task force help?

HT: Indeed, has worked some times

RT: Happy to work on test cases, but not until I need tests for my own implementation and am developing them

MSM: Last Call ends?

HT: 24 October

MSM: Only 5 weeks to know what to say our test input to the CR decision will be
... That's pretty soon, if we don't have any serious pushback on the spec. itself

HT: Two ways we could go -- push hard on tests right away, or lengthen the last call period

MSM: Or just expect we will have some gap between the end of LC and the beginning of CR

AM: This period is a really good time to focus on test coverage
... We can respond to questions by increasing test coverage
... an opportunistic approach -- test what seems tricky/controversial/novel to commentators

HT: Likes the idea
... I agree that the whole WG should be focussed on testing for the LC period

MSM: That's OK by me, if the entire WG is willing

HT: Anyone unhappy with guidance to the chair along these lines?
... So RESOLVED
... Congratulations all around


Minutes formatted by David Booth's scribe.perl version 1.128 (CVS log)
$Date: 2007/09/27 16:07:32 $