See also: IRC log
Date: 19 April 2012
<scribe> Meeting: 213
<scribe> Scribe: Norm
<scribe> ScribeNick: Norm
<alexmilowski> There is nothing to hear as of yet.
-> http://www.w3.org/XML/XProc/2012/04/19-agenda
Norm was going to add the zip questions, but in Jim's absence...
Accepted.
-> http://www.w3.org/XML/XProc/2012/04/12-minutes
Accepted.
No regrets heard.
<scribe> No progress reported on any actions #sadpanda
-> http://lists.w3.org/Archives/Public/public-xml-processing-model-wg/2012Apr/0024.html
Norm: There's been some email; we could go there or if Murray, there's anything you're looking for...
Murray: I'd like to walk through the use cases and see which, if any, are thought to be incomplete.
Norm; Murray, lead on!
Alex: This is the one I'm going to write about, but haven't yet.
<scribe> ACTION: Alex to write about 5.7, extracting MathML [recorded in http://www.w3.org/2012/04/19-xproc-minutes.html#action01]
Alex: Depends on what we mean by persistence.
<scribe> ACTION: Alex to write up a description of persistence for 5.12 [recorded in http://www.w3.org/2012/04/19-xproc-minutes.html#action02]
Alex: That could be interpreted many different ways
Vojtech: We can construct pipelines in the browser and submit them to the server, you could do that on a handheld.
Norm: We could drop it.
Alex: It's underspecified.
Norm: We could satisfy the letter of this req. along the lines Vojtech decided, but it'd take half a day and we could do other things with that half day. Let's drop it.
Alex: I agree.
<scribe> ACTION: Murray to drop 5.13 [recorded in http://www.w3.org/2012/04/19-xproc-minutes.html#action03]
<scribe> ACTION: Alex to write a pipeline to demonstrate 5.14 [recorded in http://www.w3.org/2012/04/19-xproc-minutes.html#action04]
<scribe> ACTION: Alex to write a pipeline to demonstrate 5.15 [recorded in http://www.w3.org/2012/04/19-xproc-minutes.html#action05]
Alex: Can we do this in XSLT?
Norm: I think you can.
Alex: I think we'd need to make the "collection" step explicit.
<scribe> ACTION: Norm to write a pipeline to demonstrate 5.16 [recorded in http://www.w3.org/2012/04/19-xproc-minutes.html#action06]
Vojtech: I think 5.17 and 5.12 are the same.
<scribe> ACTION: Jim to write a pipeline to demonstrate 5.17 [recorded in http://www.w3.org/2012/04/19-xproc-minutes.html#action07]
<scribe> ACTION: Norm to write a pipeline to demonstrate 5.18 [recorded in http://www.w3.org/2012/04/19-xproc-minutes.html#action08]
Murray: We don't have a step to read non-XML.
Alex; You did some work in this area, yes?
Vojtech: Yes, but it's not version 1.0 compatible.
Alex: I think that's ok.
Vojtech: I can do this one, I can add a pointer to the paper
Norm: I think this is a clear requirement for XProc V.next
Alex: You could do this today with p:data and using XSLT to parse out the data.
Norm: Yes, I think we should be able to do better in V.next
Murray: I think we should not write ths up then, it's a V.next requirement
<scribe> ACTION: Vojtech to send a pointer to his XML Prague paper to Murray [recorded in http://www.w3.org/2012/04/19-xproc-minutes.html#action09]
Some discussion of various approaches, EMC & MarkLogic included.
Vojtech: There's nothing special about the pipelines, it's just behind the scenes a connection to a database.
Alex: I thought about this with
the MarkLogic steps. There's no reason why they wouldn't work
with eXist.
... It's pretty generic.
... Would there be value in trying to have a standard way to
connect at least to XML databases.
Norm: This sounds to me like a
requirement for a new atomic step or set of atomic steps.
... I worry about the details, but we can certainly put a stake
in the ground and look at designing atomic steps.
Murray: We should be able to design the steps, whether or not they'd be satisfactory is an open question.
Henry: In the distant past with
the Markup Engine, I ended up implementing a very simple
gateway to a relational database. It turned out that there were
a number of tasks we were exploring where you wanted to be able
to look up an atomic item in a relational database.
... I think that's worth exploring. There are a bunch of
existence proofs, whether there's a sweet spot is an open
question.
Murray: They seem like they'd be easy.
<scribe> ACTION: Vojtech to write pipelines to demontrate 5.21 and 5.22 [recorded in http://www.w3.org/2012/04/19-xproc-minutes.html#action10]
Norm: I think this duplicates 5.12 and 5.17
Vojtech: Well, there's some conditionality in here. I can take a look.
<scribe> ACTION: Vojtech to consider writing a pipeline for 5.23 [recorded in http://www.w3.org/2012/04/19-xproc-minutes.html#action11]
Norm: Maybe we should combine 5.20 and 5.24 into a more wholistic requirement
Discussion suggests that this is similar to 5.12, 5.17, etc.
Alex: Maybe we need a more general: pipeline performs query and formats results.
Murray: This is something we can do today.
Alex: Well, sortof.
... I wonder if we want to combine this with the non-XML input
example.
<scribe> ACTION: Alex to write a pipeline to demonstrate 5.28 [recorded in http://www.w3.org/2012/04/19-xproc-minutes.html#action12]
<alexmilowski> Veronika cries about DSDL
Norm: I think the salient issue here is whether we want to allow multiple numbers of inputs and outputs
Alex: Do we have any other use case that has that feature requirmenet.
<scribe> ACTION: Norm to write up a description of the split step and the eval step to inform discussion of multiple inputs and outputs [recorded in http://www.w3.org/2012/04/19-xproc-minutes.html#action13]
Henry: I doubt that the complexity is worth the functionality, but we'll see.
Alex: I can write this up. This is a streaming use case. Nothing prevents p:viewport from doing this, but your implementation may not work that way.
<scribe> ACTION: Alex to write up 5.29 [recorded in http://www.w3.org/2012/04/19-xproc-minutes.html#action14]
Norm: Isn't 5.30 just an example of 5.29?
Alex: Yes.
Vojtech: There are a couple of different ways to interpret this.
Norm expresses concerns about the multiple output document.
Norm: I'm all for writing it up, but let's leave out multiple output documents because it's a side issue
<scribe> ACTION: Vojtech and Norm to collaborate on writing up 5.31 [recorded in http://www.w3.org/2012/04/19-xproc-minutes.html#action15]
Vojtech: The pipeline will fail at runtime not compile time.
Norm: I can see how a long run that failed because of the last step would be annoying.
Alex: But we have step-available, so this is solved.
Norm: I agree.
<scribe> ACTION: Alex to write a pipeline for 5.32 [recorded in http://www.w3.org/2012/04/19-xproc-minutes.html#action16]
Murray: I only recently came
across the mail that says the aims of this effort is at
low-hanging fruit.
... You can see from this document that I didn't know
that.
... When we're talking about resource management, are we being
entirely opportunistic?
... Or is there room for something like a webdav.
Norm observes that we did the low-hanging fruit thing in order to be done quickly.
Alex: Why can't we just add a requirmenet for webdav?
Norm: I'm all for adding the requirement
Alex: The deeper question of resource management is still an open issue I think.
Norm: On the resource manager, I think we agreed to talk about it but not put it in the critical path.
Alex: where are the use cases?
<scribe> ACTION: Norm to write a few resource manager use cases [recorded in http://www.w3.org/2012/04/19-xproc-minutes.html#action17]
Murray: The other thing is
debugging.
... Who's the champion for debugging?
Norm observes that he'd do it if he had a coupleof years salary in the bank
Alex: Can we engage a tool vendor to help in this area?
Murray: I gotta hope there's more
to the story than log messages.
... I know we can do better. What I'm finding it hard to find
in the spec is all the ways I can get information.
Alex: I sort of divided the world into debugging vs. tracing a profiling. Should we address this in the spec?
<scribe> ACTION: Norm to put "debugging" on the agenda for next week [recorded in http://www.w3.org/2012/04/19-xproc-minutes.html#action18]
Henry: I was surprised and
impressed that the XQuery implementations got as far as they
did. Here's a use case: can you write an XProc processor in
XProc?
... I don't think so.
... Another way to put it is, is there an "evaluate this step"
step?
Norm: Only in extensions.
Adjourned.
This is scribe.perl Revision: 1.136 of Date: 2011/05/12 12:01:43 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/relation/relational/ Succeeded: s/treiv/triev/ Found Scribe: Norm Inferring ScribeNick: Norm Found ScribeNick: Norm Present: Vojtech Alex Henry Norm Murray Regrets: Cornelia Jim Agenda: http://www.w3.org/XML/XProc/2012/04/19-agenda Found Date: 19 Apr 2012 Guessing minutes URL: http://www.w3.org/2012/04/19-xproc-minutes.html People with action items: alex jim murray norm vojtech[End of scribe.perl diagnostic output]