XML Processing Model WG

15 Dec 2005


See also: IRC log


Norm, Jeni, Paul, Henry, Alessandro, Andrew, Eric, Michael, Richard, Rui, and_Alex
Norman Walsh
Norman Walsh




<ht> Norm, are you going to try to be like our hero PaulG and chair _and_ take minutes?

<ht> Or will we need a scribe?

ht, I'm going to try today at least :-)

<ht> OK, I'll be available if you need me as fallback

<scribe> Scribe: Norman Walsh

<scribe> ScribeNick: Norm

Date: 15 Dec 2005

Welcome and introductions

Scribe decides not to attempt to record introductions.

Norm asks if the WG minds having Rui as a guest this week as he hasn't finished the invited expert process. No objections.

Chair thanks the members for their introductions.


Zakim: http://www.w3.org/2001/12/zakim-irc-bot

RRSAgent: http://www.w3.org/2002/03/RRSAgent

RRS Agent: http://www.w3.org/2002/03/RRSAgent

<alexmilowski> port issue?

<AndrewF> http://cgi.w3.org/member-bin/irc/irc.cgi

Face-to-face meeting at the Tech Plenary?

<MSM> http://www.w3.org/2005/10/tp6groups.html

<PGrosso> http://www.w3.org/2002/09/TPOverview.html is the only URL for the TP info to date.

We have two days of f2f space at the Tech Plenary outside Cannes on M/T the 27/28 Feb 2006

Regrets from Jeni for the f2f

Proposed: to hold the f2f meeting at the Tech Plenary as planned.


Meeting on 22 Dec 2005?

Meet on 22 Dec?

<MSM> telcon regrets Richard, Alessandro

Regrets for 22 Dec: Richard, Ale1

Next meeting: 22 Dec 2005

In general we'll meet weekly at 11:00 EST on Thursdays

<alexmilowski> Are there definitions of the WG acronyms on http://www.w3.org/2005/10/tp6groups.html ?

Eric wonders if we can change times. Norm proposes we stick with this until the f2f

<scribe> Cancelled: 29 Dec 2005; next meeting after 22 Dec will be 5 Jan 2006

<richard> http://www.w3.org/XML/Group/Processing.html

Group page: http://www.w3.org/2000/09/dbwg/details?group=38398

Review of charter and deliverables

Charter: http://www.w3.org/2005/10/xml-processing-model-wg-charter.html

<MSM> Charter: http://www.w3.org/2005/10/xml-processing-model-wg-charter.html

Public group page: http://www.w3.org/XML/Processing/

<scribe> ACTION: ht to update administrative page with pointers to relevant docs (Charter, etc.) [recorded in http://www.w3.org/2005/12/15-xproc-minutes.html#action01]

We'll do most of our work on the public page

Norm takes a run through the charter

<Zakim> ht, you wanted to wave a big stick

HT points out that there are two kinds of WGs (broadly) ones that are inventing new stuff, ones that are standardizing existing technology and working on interoperability. We're in the latter group and are well placed to work quickly.

AlexM wonders how others feel about the charter and if there's consensus regarding the goals

AlexM: I have two hot buttons: extensibility and streaming pipelines

Richard: Scripting language has two parts: having a way to specify a sequence of processes and to talk about what form the data is in as it passes through the process. It would be nice if we could do the former without constraining the latter
... One could imagine describing a pipeline without talking about data models and have different implementations that did the same thing with different data models

<MSM> Richard, do you mean one implementation that passes SAX events at the process boundaries, and one that passes a DOM around, and one that passes XML (or NSGML) around?

<richard> yes, that sort of choice

<MSM> +1

Eric: Richard has a good point; those can be separated and we shouldn't enforce a particular data model
... Wonders if the use cases we have today can be met with this kind of implementation
... It's not necessary that V1 be able to do everything in every pipline language. We should try to get the bulk of the use cases and let individual vendors provide extensions for the rest.

Rui: I think there should be some kind of registry so that we can have common components that do the same thing in different implementations
... different implementations should be able to get the same results with the core components

ht: If I understand Rui correctly, it addresses one of the weaknesses of the original Sun pipeline note which is the awkardness of having to have indirection to identify what each stage of the pipeline does

<rlopes> exactly

ht: If you had a registry, you could identify small names for the processes that do things, like registering "XSLT" for the XSLT 1.0 process

richard: proposes that URIs could be used to identify such components; there will be some standard (and some non-standard) components and they will need to be named
... it would be nice if you could abstract away exactly how the data is provided to these components
... on the one hand you have a scripting language and on the other hands you have components that describe the kinds of inputs that they accept

ht: I'm going to signal early that I think the hard question is going to be conditionals and whether or not we have some form of conditionality. The driving force behind answering that is probably going to be error handling. I'm not sure I have a clear opinion.

MSM: one goal for the first deliverable: I'd like to propose that the spec be able to be short; no more than 15 normative pages.
... The Algol 60 report is 17 pages.
... It ought to be implementable by a desperate perl hacker in a week or two.
... I'm nervous about the second deliverable. I think it may not be possible. While I agree that there are contexts, like browsers, where it makes sense to have a default processing model, but there's a lot of pressure to force XML away from a declarative semantics and towards an imperative semantics and its going to be difficult to resist that pressure if we aren't resolute.
... I think it's a trap to assign imperative semantics to the !DOCTYPE declaration or the existinance of a schema location hint.
... Those are declarative statements, not requests that processing occur
... It's very difficult to describe a default semantics without becoming imperative.

<PGrosso> http://www.w3.org/TR/2004/NOTE-proc-model-req-20040405/

Chair encourages WG members to contribute use cases that they feel the should be solved by the core language

Richard asks why the Core WG requirements document isn't in the charter

Norm suggests oversight.

<Jeni> Thanks Norm!

<Jeni> bye

<ht> Norm, presume you will link the minutes from the public page?

<ht> Or just email them. . .

<rlopes> bye

<ht> Michael, some status info on #xmlschema for your delectation . . .

Summary of Action Items

[NEW] ACTION: ht to update administrative page with pointers to relevant docs (Charter, etc.) [recorded in http://www.w3.org/2005/12/15-xproc-minutes.html#action01]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.127 (CVS log)
$Date: 2005/12/22 16:08:58 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.127  of Date: 2005/08/16 15:12:03  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: i/agendum 1./Topic: Welcome and introductions
Succeeded: i/agendum 2./Topic: Administrivia
Succeeded: i/agendum 3./Topic: Face-to-face meeting at the Tech Plenary?
Succeeded: i/agendum 4./Topic: Meeting on 22 Dec 2005?
Succeeded: i/agendum 5./Topic: Review of charter and deliverables
Succeeded: s/regrets:/telcon regrets /
Succeeded: s/Accepted./Resolved/
Found Scribe: Norman Walsh
Found ScribeNick: Norm
Default Present: Norm, Jeni_Tennison, Jeni, PGrosso, Ht, +1.408.891.aaaa, Ale1, +1.650.491.aabb, AndrewF, ebruchez, richard, RuiLopes, alexmilowski, msm
Present: Norm Jeni Paul Henry Alessandro Andrew Eric Michael Richard Rui and_Alex
Agenda: http://www.w3.org/XML/XProc/2005/12/15-agenda.html
Found Date: 15 Dec 2005
Guessing minutes URL: http://www.w3.org/2005/12/15-xproc-minutes.html
People with action items: ht

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.

[End of scribe.perl diagnostic output]