XML Processing Model WG

18 Nov 2010


See also: IRC log


Norm, Alex, Vojtech, Paul
Henry, Mohamed


Date: 18 November 2010

<scribe> Meeting: 184

<scribe> Scribe: Norm

<scribe> ScribeNick: Norm

Can you hear me? I can't hear you?

I can't here you. I meant.

<Vojtech> I don't hear anything

That's weird I heard you say that. I'll dial in again.

Alex, can you hear me?

<alexmilowski> I'm on skype without a headset, so we'll see how this goes

<alexmilowski> No.

<alexmilowski> I can hear others.

Vojtech, I heard you say "The silence is almost absolute" then you went away

Vojtech/Alex can you hear each other?

Apparently I can hear Vojtech and Alex

Accept this agenda?

-> http://www.w3.org/XML/XProc/2010/11/18-agenda


Accept minutes from the previous meeting?

-> http://www.w3.org/XML/XProc/2010/11/04-05-minutes

As amended yesterday.


Update on LC draft of processor profiles

Henry reports he'll have something for us next week.

Next meeting: telcon, 9 Dec 2010?


Review of p:document-template note

Norm: I think the only substantive issue is how to parse curly braces

Vojtech: XQuery handles curly braces by watching for quotes.
... I think doing it like that would be more consistent for users, but harder for implementors.

Norm: Well...

Vojtech: On the other hand, maybe it's not really that hard.

Norm attempts to summarize:

When you see a "{" (not doubled), switch to xpath-mode. When you see a ' or " in xpath-mode, switch to quote-mode. In quote-mode everything is quoted until the closing quote.

At the closing quote, go back to xpath-mode.

In xpath-mode, a "}" whether it is doubled or not, immediately ends the expression.

Some discussion of what happens if you leave a quote out. You'll read the whole content of whatever text node you're looking at.

Norm: Are those the rules everyone prefers?

Vojtech: What happens if you use character references for quotes?

Norm: No. We'll only see the expanded references.

Alex: I like it.

Norm: In the common case, you'll be able to write { concat('{', $foo, '}') }
... That probably is easier than doubling all the curly braces

Vojtech: What about the rules for doubled braces?

Norm: We could say "{" is an error in XPath mode.

Vojtech: What happens if XPath 3 introduces "{" in XPath?

Alex: We'll have to change the parsing rules.

Norm: Given that XSLT and XQuery both use curly braces to bound expressions, it seems unlikely to me that they'll decide to use them for something else in XPath V.x

In regular-mode, {{ is a {. In regular-mode, { (undoubled) starts xpath-mode.

In xpath-mode, ' or " starts quote-mode. In quote-mode no characters are

special except the matching quote that returns us to xpath-mode.

In xpath-mode { is an error. In xpath-mode } ends the expression and returns

to regular-mode (after inserting the result of evaluating the expression).

<alexmilowski> yes

<Vojtech> yes

Norm: Is that what people like?

General agreement.

Norm: Any other discussion about document templates or the note?

GRDDL step?

Alex: I was looking at GRDDL and think it's something we should do as a standard step, but I haven't looked into it yet.

Vojtech: I looked at the spec and from what I read, it seemed to me that parts of it we could do already.

<alexmilowski> e.g.

<alexmilowski> <html xmlns="http://www.w3.org/1999/xhtml"

<alexmilowski> xmlns:grddl='http://www.w3.org/2003/g/data-view#'

<alexmilowski> grddl:transformation="glean_title.xsl

<alexmilowski> http://www.w3.org/2001/sw/grddl-wg/td/getAuthor.xsl"

<alexmilowski> >

Vojtech: It just checks for attributes and fetches a stylesheet. It looks like something that should be possible with XProc.
... There's also the merging of the RDF graphs.

Alex: That's what I was thinking about. One interpretation is that you'd output RDF directly. Another is that you output Turtle or something like that.
... We should consider what would be useful for people doing something like harvesting.

<scribe> ACTION: Alex to review GRDDL and return with a suggestion. [recorded in http://www.w3.org/2010/11/18-xproc-minutes.html#action01]

Alex: one outcome might be a standard pipeline but implementors could do it natively.
... That leads to an interesting possible enhancement: the ability to say "here's the fallback implementation of this step" but with the explicit provision that a native implementation could be used instead.

Vojtech: The GRDDL spec says that the transformation can be any process, typically it's XSLT. But XProc is also a possibility now too.
... I wonder if we do the GRDDL step if we want to say something about what languages are supported for the transformations.

Alex: Right.

Norm: Alex, as you consider GRDDL, please keep those things in mind.

Possible erratum: definition of an XProc Processor

Vojtech: We don't have a definition of an XProc Processor.

Alex: How did we do that?

Norm: Does anyone think they can write a concise definition of an XProc processor?

Vojtech: There are all sorts of rules in the spec, it has to apply the steps in the right order, etc.

<scribe> ACTION: Vojtech will review the spec and propose a definition of an XProc Processor. [recorded in http://www.w3.org/2010/11/18-xproc-minutes.html#action02]

<alexmilowski> [Definition: The software responsible for transforming source trees into result trees using an XSLT stylesheet is referred to as the processor. This is sometimes expanded to XSLT processor to avoid any confusion with other processors, for example an XML processor.]

Norm: Yes, I think we'll want something like that.

Any other business?

Vojtech: I was wondering if having a p:sort would add any value.
... something like split-sequence that takes an XPath expression to use for ordering the documents.

Norm: I'm not opposed. I never even thought of doing it.

Vojtech: Maybe it's better to put something like this in EXProc.

Norm: Yeah, we can leave it there and see if we get requests for a better job.

Vojtech: My concern is that it might not be sufficient for more complex use cases.
... if the sort criteria are not easy to express in an XPath expression, for example.

Alex: I'm just not sure what the use cases are.

Vojtech: The use case that I had was that I was trying to implement OAuth and you have to sort the request parameters before you hash them.
... For that I needed a sort. It was simple string sort so you could do that with XSLT or XQuery.

Adjourned. See you in two weeks.

Summary of Action Items

[NEW] ACTION: Alex to review GRDDL and return with a suggestion. [recorded in http://www.w3.org/2010/11/18-xproc-minutes.html#action01]
[NEW] ACTION: Vojtech will review the spec and propose a definition of an XProc Processor. [recorded in http://www.w3.org/2010/11/18-xproc-minutes.html#action02]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.135 (CVS log)
$Date: 2010/11/18 16:41:57 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.135  of Date: 2009/03/02 03:52:20  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/brace/braces/
Succeeded: s/preferes/prefers/
Succeeded: s/speciall/special/
Found Scribe: Norm
Inferring ScribeNick: Norm
Found ScribeNick: Norm
Default Present: Norm, Vojtech, alexmilowski, PGrosso
Present: Norm Alex Vojtech Paul
Regrets: Henry Mohamed
Agenda: http://www.w3.org/XML/XProc/2010/11/18-agenda
Found Date: 18 Nov 2010
Guessing minutes URL: http://www.w3.org/2010/11/18-xproc-minutes.html
People with action items: alex vojtech

[End of scribe.perl diagnostic output]