Charlie Wiecha, IBM
Leigh Klotz, Xerox (minutes)
Nick van den Bleeken, Inventive Designers
Philip Fennell, MarkLogic
Steven Pemberton, CWI/W3C (chair)
Uli Lissé, DreamLabs
Steven Pemberton: It's under beta
test at the moment.
Nick van: SIP?
Steven Pemberton: Yes.
Charlie Wiecha: And Skype?
Steven Pemberton: Yes, Skype works
already.
Nick van: Skype Out.
Steven Pemberton: We'll look at this for next week.
Steven Pemberton: Leigh said that
the bext XSLTForms version has some bug fixes.
Leigh Klotz: I have some patches in it
as well which Alain hasn't yet incorporated. I've sent it to
Michael Sperberg-McQueen and I'll send it to you as well.
Steven Pemberton: Who is likely to
turn up for our four days, beside me.
Charlie Wiecha: Probably not.
Uli Lissé: I will.
Nick van: There is a small chance that
I may have a conflict.
Philip Fennell: I can't attend.
Steven Pemberton: The OWL WG uses
Wikis to create their specs. They don't have anything to release to
us yet.
Steven Pemberton: Here's an example of
a set of specifications produced together: http://www.w3.org/2007/OWL/wiki/Round_8
Steven Pemberton: Here is an example
of a produced PR http://www.w3.org/TR/2009/PR-owl2-profiles-20090922/
Steven Pemberton: And here's the page
that the editors produce: http://www.w3.org/2007/OWL/wiki/Profiles
Steven Pemberton: There's also some
supposedly perl code to put an existing draft into MediaWiki.
Charlie Wiecha: That's HTML or
specxml?
Steven Pemberton: HTML. Then their
program produces HTML.
Charlie Wiecha: Not SpecXML.
Leigh Klotz: I'm sure it could be
hacked on.
Steven Pemberton: The spec page looks
recognizable.
Leigh Klotz: View Wiki Source on it
looks fairly clean.
Leigh Klotz: Charlie was going to do
the syntax and I was going to do the program but we've got them
both now.
Steven Pemberton: Are we happy?
Charlie Wiecha: Maybe some Team
cross-wg support for this?
Steven Pemberton: It would fall on the
systems team and they're overloaded.
Leigh Klotz: We should talk about
where we need to use SpecXML as an intermediate step.
Charlie Wiecha: We should at least
share code.
Steven Pemberton: This is produced by
a W3C Team Member and is in SVN.
Action 2010-09-1.1: Steven Pemberton to follow up on OWL wikispec generator.
Steven Pemberton: I think we don't
have a problem with the dates.
Leigh Klotz: Then a formal reply to
Art would be good, so I'll do that now.
Leigh Klotz: We're still adding Orbeon-style parameters and I'll report back when we're done.
Philip Fennell: I've plugged in the
XForms 1.1 schema into the oXygen 11 NVDL and got it working. I'll
create an XProc pipeline that runs the test cases through
NVDL.
Leigh Klotz: What XProc are you
using?
Philip Fennell: Calabash, although we
could use others.
Leigh Klotz: This all sounds great to
me. It's what the ISO/IEC committee said, to use NVDL instead of
just RNG-RNG integration.
Philip Fennell: There's a Calabash
extension that uses NVDL and XProc. I'd also like to integrate XSLT
into the mix.
Leigh Klotz: For Schematron
validation?
Philip Fennell: For doing
XHTML+XForms+XSLT host document. You could use Schematron by using
an XSLT transformation actions inside one of those implementations.
Or you could run an XSLT1 implementation of Schematron inside an
XForms implementation to validate instances.
Leigh Klotz: So you'll send out a
status message or some files?
Philip Fennell: Yes, I'll zip up what
I have and then start on the test suite through nvdl.
Leigh Klotz: You might be able to use
Nick's XML control files for describing the test suites.
Philip Fennell: I've seen the HTML and
XForms files.
Nick van: I have a language that
automatically decides how to test, not how to validate. http://github.com/nvdbleek/com.orbeon.testsuite.w3c
Leigh Klotz: But we could start both
schema validation and test running from the same set of control
files.
Steven Pemberton: So you are saying
we should have more precise times.
Nick van: We had some agreement in the
last discussion.
Steven Pemberton: I think I agreed;
stronger types are better.
Leigh Klotz: Resolution and
action?
Nick van: This is part of my running
action item for XPath 2.0 model.
Steven Pemberton: Any objections?
Resolution 2010-09-1.1: We agree with http://lists.w3.org/Archives/Public/public-forms/2010Jan/0029.html on XPath 2.0 function signatures.
Steven Pemberton: We should wait
for Erik, but I had an idea that this was already thec ase in XML
Events 2:
http://wiki.orbeon.com/forms/doc/developer-guide/xforms-events#TOC-Filtering-on-the-event-phase
Steven Pemberton: I have no objection
but I don't know what the use case is. I don't know when you would
want bubbling without target.
Leigh Klotz: It says if you omit
ev:phase then target|bubble is the default.
Steven Pemberton: Absolutely. We
should wait for Erik.
Leigh Klotz: Charlie had a proposal
and we discussed adding in QNames for control. I contacted the
Badgerfish author who sent us a good link, but didn't answer when I
asked if he wanted it hosted by W3C.
Charlie Wiecha: We discussed a simple
mapping for the simple XML that is easily represented by JSON;
there's no notion of ordering or mixed content, so the mapping of
that subset of XML into JSON would be straightforward and ehave
good utility.
Steven Pemberton: So we can read all
JSON but not write all XML.
Charlie Wiecha: Yes. There are
cumbersome JSON mappings but they're not useful for the goal of
letting XForms processors talk to JSON in the wild. There are some
JSON property names have invalid characters, so there will be some
magic. It's considerably simpler.
Steven Pemberton: So we need a
concrete proposal.
Charlie Wiecha: We wanted a
declarative way for dealing with the base case and candidate
spec-ready text. There would be two levels: completely declarative
by MediaType with no event handlers or callback, and also an
extension mechanism to escape out and let people get out. The first
part would need a concrete proposal.
Steven Pemberton: Ideally, it would be
transparent (opaque) and you wouldn't see it in the form; the
system would see that it's JSON and convert it. The submission
would need to show it.
Leigh Klotz: We ought to serialize to
whatever we call it, "Badgerfish" or whatever, not "JSON".
Steven Pemberton: Right.
Charlie Wiecha: That's a good
forward-extension mechanism.
Leigh Klotz: Also we've acquired the
XSLT action now so we could transform the instance in XML to make
it suitable for serialization with the chosen JSON serializer. That
may make it clear how the "cleanup" works.
Charlie Wiecha: Good.
Steven Pemberton: Now we need to pick
which serialization we will use.
Leigh Klotz: Also we should get Alain
involved as there is the XSLTForms implementation already: http://www.agencexml.com/jsoncallback/wikipediasearch.xml
Steven Pemberton: Does he do
submission as well?
Leigh Klotz: No, just GET. For some
reason he uses submission/resource/@value also instead of a query
parameter.
Leigh Klotz: Need an owner for
this; was Mark Birbeck. It would be nice if someone from Ubiquity
would follow this more. It doesn't hit the back-end implementations
as much.
Steven Pemberton: John?
Leigh Klotz: We need to ask
John.
Steven Pemberton: And what abotu
Cross-site requests ; Uniform Messaging Policy, Level One?
Leigh Klotz: Let's put them
together.
Steven Pemberton: Who's interested
and what changes do we need?
Leigh Klotz: The biggest is probably
assertions.
Steven Pemberton: Since all we do is
include Schemas, does XForms need to do anything?
Leigh Klotz: I don't know.
Steven Pemberton: Who loves
schemas?
Philip Fennell: I'm tempted to say I'm
interested. I don't know how much time it will take.
Action 2010-09-1.2: Philip Fennel to review XML Schema 1.1 for XForms 1.2 and report what might need to be done.