XML Processing Model WG

18 Oct 2007


See also: IRC log


PGrosso, avernet, Ht, +95247aaaa, MoZ, Andrew, Norm



Date: 18 October 2007

number Meeting: 88, T-minus 2 weeks

<scribe> Scribe: Norm

<scribe> ScribeNick: Norm

<ht> Norm, shall I start the call w/o you?

<ht> http://www.w3.org/XML/XProc/2007/10/18-agenda

<ht> scribenick: ht

HST: Agenda approved


HST: Minutes approved
... Next meeting 25 October
... NW apologies for 25 Oct and 1 Nov, HST to be in the chair _pro tem_
... Charter has been extended for 1 year

PG: F2F agenda?

NW: one day on comment processing, one day on the future, I guess

PG: I find a detailed agenda helps folk to be prepared

NW: Well, item 1 is "Read and be familiar with the details of the issues list"

<scribe> scribenick: Norm

Charter extension

Extended 1 year.

Review of action items

A-86-01: Alex to review XSLT streaming requirements before the face-to-face.

<scribe> Continued

A-86-03: Henry to reply to the commenter (non-string parameters; issue 30)

<scribe> Completed

A-86-04: Henry to craft the prose to cover the defaulted output case

<scribe> Continued

A-87-01: Norm to take a stab at reconsidering the default inputs feature applying it only to ports that are not primary

<scribe> Continued.

A-87-02: Alex to propose some text about imports and circularity

<scribe> Continued

A-87-03: Norm to attempt to incorporate Richard's draft text about step type scope

<scribe> Continued.

Comment 29: Determining whether a pipeline has a (defaulted) output

-> http://www.w3.org/XML/XProc/2007/09/lastcall/comments#029

<scribe> Continued pending Henry's action

Comment 6: Bindings for pipeline inputs

-> http://www.w3.org/XML/XProc/2007/09/lastcall/comments#006

<scribe> Continued pending Norm's action

Comment 18: Scope of step types

-> http://www.w3.org/XML/XProc/2007/09/lastcall/comments#018

<scribe> Continued pending Norm's action

Comment 24: Passing PSVIs between steps

-> http://www.w3.org/XML/XProc/2007/09/lastcall/comments#024

Norm outlines the issue and summarizes Jeni's observed options

Henry: I think we leave it impl. defined.

Some discussion of how defaulted attributes fit in: probably covered by the spirit of the best efforts clause.

Proposal: Leave it implementation-defined.


<scribe> ACTION: Henry will respond to the commenter on comment 24: passing PSVIs between steps. [recorded in http://www.w3.org/2007/10/18-xproc-minutes.html#action01]

Comment 7: Saxonica comments on sections 1 and 2

-> http://www.w3.org/XML/XProc/2007/09/lastcall/comments#007

Norm reviews the points in Mike's message.

Point 1: accepted.

Point 2: changed validate-* to validate-with-*; accepted.

<MoZ> http://lists.w3.org/Archives/Public/public-xml-processing-model-comments/2007Sep/0034.html

Point 3-6 are editorial.

Point 7: rejected, also made a separate issue

Point 8: discussion

Henry: I think we're pretty clear that we're not answering this question.
... We're trying not to be too precise.


Henry: No, that doesn't work. An XML document isn't an Infoset, the infoset is just a set of terms
... But maybe it's ok.
... Sure let's try this.

Norm: What about A.3?

Henry: Yes, that looks fine too.

Norm: Anyone unhappy with the editor's resolution of Mike's point 8?

None heard.

Point 9:


Norm: Anyone unhappy with the editor's resolution of Mike's point 9?

None heard.

That leaves some editorial clarifications, but I think we've covered the technical issues

Comment 12: Saxonica comments on sections 3 and 4

-> http://www.w3.org/XML/XProc/2007/09/lastcall/comments#012


Let's leave the "editorial" and "clarification" points to the editor unless they turn out not to be

Henry: I'm prepared to skip to 7.
... I think he's right.

Norm: I don't think these apply to the things *inside* the elements in the step

Henry: So you meant "children" not "contains"? If so, you'll have to repeat it endlessly.

Norm: Uhm...

Henry: Can't we just say that these rules don't apply inside p:inline?

Norm: Perhaps

Some discussion about what constitutes a step.

Norm: I think children would work.

Henry: What about giving the fifth and sixth bullets a parallel construction
... If any element in the XProc namespace other than p:inline, or any step, has text node children...

Norm: Sure, that would work for me.
... I think that resolves point 7.

Point 8:

Henry: I think replacing "within its container" by "immediately contained by that steps container"
... Or if we've formally defined subpipline as the immediately contained steps, then "the last step in document order in the subpipeline"

Norm: I'm happy to attempt to clarify that.

Point 11:

Some of this is editorail.

Norm: I think we have clarified that select only selects elements or documents.

Henry: I've always said that select needs the same namespace fixup we already described.
... We've already re-worded 4.2 so there isn't a double "each" anymore.

Norm: But that leaves "wrapped" and an explicit pointer to 2.6.1
... The select question points to 5.2 and aon through to p:input where it's covered.

Henry: No, I don't think we want to make the reference to 2.6.1 explicit.

<scribe> ACTION: Norm to see if the commenter agree's we've addressed his concerns. [recorded in http://www.w3.org/2007/10/18-xproc-minutes.html#action02]

Point 12:

Henry: I think we can replace "aggregated" with "concatentated" here.

Norm: I'm willing to do that and see if it helps.

Henry: The prose is still a bit terse.
... I think we should unpack it and make it more explicit.

Norm: Ok.

Point 13:

Norm: We probably need to make the distinction between match and select more clear.

Henry: Not that you need to make the point about match only matching element or document nodes here. There's no free ride.

Norm: The WG's intent is clear but the prose needs to be clearer.
... Should we make it an error to select a document node?

Henry: Yes.


Point 14:

Norm: I think it can be a static error.

Henry: Is this just a case where a processor could detect it statically if it wanted to?

Norm: It could, but why not make it static?

Henry: A select option to a step may be computed, so it has to be a dynamic error.

Norm: But for select on our compound steps and for test on when, then they should be static errors.

Henry: Let's try that and see if there's any pushback.

Norm: Any objections?


Any other business


Summary of Action Items

[NEW] ACTION: Henry will respond to the commenter on comment 24: passing PSVIs between steps. [recorded in http://www.w3.org/2007/10/18-xproc-minutes.html#action01]
[NEW] ACTION: Norm to see if the commenter agree's we've addressed his concerns. [recorded in http://www.w3.org/2007/10/18-xproc-minutes.html#action02]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.128 (CVS log)
$Date: 2007/10/18 16:06:49 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.128  of Date: 2007/02/23 21:38:13  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/scribe: ht/scribenick: ht/
Succeeded: s/yaer/year/
Succeeded: s/error/error to select a document node/
Found Scribe: Norm
Inferring ScribeNick: Norm
Found ScribeNick: Norm
Found ScribeNick: ht
Found ScribeNick: Norm
ScribeNicks: Norm, ht
Default Present: PGrosso, avernet, Ht, +95247aaaa, MoZ, Andrew, Norm
Present: PGrosso avernet Ht +95247aaaa MoZ Andrew Norm
Agenda: http://www.w3.org/XML/XProc/2007/10/18-agenda
Found Date: 18 Oct 2007
Guessing minutes URL: http://www.w3.org/2007/10/18-xproc-minutes.html
People with action items: henry norm respond will

[End of scribe.perl diagnostic output]