Mark Birbeck, x-port.net
Blake Jones, ViewPlus Technologies/DAISY
David Landwehr, PicoForms
Jan Kratky, IBM
Keith Wells, IBM
Joern Turner, DreamLabs
Leigh Klotz, Xerox (minutes)
Nick van den Bleeken, Inventive Designers
TV Raman, Google
Sebastian Schnitzenbaumer, DreamLabs
Steven Pemberton, CWI/W3C (Chair)
Ulrich Nicolas Lissé, DreamLabs
Jan Kratky: I'm leaving the WG and
being replaced by Keith Wells.
Keith Wells: I hope to be a good
contributor.
Jan Kratky: Keith has been working
with me on proofs of concept, and the test suite. I'll be moving on
to the work on the Rational products within IBM.
Steven Pemberton: Mark and I were
at both; Raman was at WWW-2007. I've come away with a good feeling
about the web direction. The compound document approach to the web
seems to be breaking through. There was a lot of talk about XHTML2
and XForms that was very positive. I was told that my XForms 1.1
tutorial was well-attended. There was a panel of 8 with 45 minutes
at XTech for the future of HTML. The BBC interviewed me and Mike
Smith straight afterward. During that panel, Dave Raggett, who is of
course a classic name in HTML, said that XHTML2 and XForms are
really great, and "we" (Volantis) are delivering XHTML2 and XForms
to mobile devices with no problem at all. Is Dave here? There was
lots of discussion and demo of RDF-A; six talks mentioned it. The
wonderful talk and demo by Joost. They're doing peer-to-peer
television. It's in beta testing. It's a neat application. They
talked about their architecture. Their architecture is compound
documents, SVG, XBL, and CSS.
David Landwehr: [IRC] Allan Beaufour
is working for Joost and it is based upon Mozilla
Steven Pemberton: It's in Mozilla and
everything is done with SVG and widgets.
David Landwehr: [IRC] it is actually
xul runner
Steven Pemberton: It's the post child
for our approach. Plus the Joost people started talking about
adding RDF-A to their website.
TV Raman: And the panel?
Steven Pemberton: It was the future of
HTML. At 45 minutes, there wasn't much time to say very much. We
had to introduce yourselves, and we were just supposed to say who
we were. Two of the HTML5 people immediately said how they much
they hated namespaces. So I said that perhaps that was the real
problem, and the issue isn't markup languages but something else.
Mark, anything else?
Mark Birbeck: From the floor, it
seemed that people wanted clarity on what was happening. Bystanders
seemed to be asking what was going on. The audience seemed to think
that XML was a done deal and were confused by it all. It does make
it clear the whole thing is a bit of a mess.
Steven Pemberton: There should be a
transcript. I'll try to post it to those who were interested.
TV Raman: The HTML5 WG has silently checked in the draft and they are integrating the Web Forms 2.0 and HTML5 documents. They've concluded the discussions.
Steven Pemberton: I also went to a
lightning talk by Anne von Kesteren. We've reduced the DOCTYPE to
<DOCTYPE HTML>. Of course, in XHTML2 we got rid of the
DOCTYPE. He also said they now have video and audio tags. It's not
clear why you need a video tag.
TV Raman: It seems to be an Opera
browser feature involving their codecs.
Sebastian Schnitzenbaumer: Ulrich will join from DreamLabs instead of Joern or me.
Leigh Klotz: The issue is that some
have found the text in the XForms 1.1 Basic Profile document "may"
about XML Schema subset is ambiguous. We have agreement now, I
think, that we need the test suite updated to set the XForms Basic
test flag to true for the XML Schema tests that that XForms Basic
should pass. Also, there are some minor issues about other tests
such as xsi:nil.
Jan Kratky: We don't have an action
item, though. We should give that to Keith.
Action 2007-05-23.1: Keith Wells to change XForms Basic test suite flag to true for XML Schema tests.
Steven Pemberton: When?
Jan Kratky: We can get it done before
the F2F.
Nick van: [IRC] I marked Jan Kratky to
provide support for XForms 1.0 Basic Profile test suite feature
list. as closed today, so we need a new action item or reopen that
one. In fact he already has 3, Jan's action items are transferred
to Keith ;)
Sebastian Schnitzenbaumer: I think
this should be an HCG issue. I did this in the HTML WG but nothing
happened.
Steven Pemberton: What do I report to
the HCG? You tried to set up the task force and failed?
Sebastian Schnitzenbaumer: The charter
said that architectural consistency is required. They say it is
architecturally consistent. Technically, it is the task of the task
force to come to this conclusion.
Steven Pemberton: So you're requesting
to me that I report to the HCG and ask for this task force to be
formed?
Sebastian Schnitzenbaumer: I'm saying
that's our only option; we should decide as a group.
Nick van: The list is up to date. I sent a new version an hour ago.
Steven Pemberton: I suggest we go straight to these issues.
Steven Pemberton: Is Erik around today? Let's skip these.
Steven Pemberton: Nick?
Nick van: I need to re-read this. I
think you can never reach it point 2.b.
Steven Pemberton: Are we missing an
OR? How did we get here?
David Landwehr: I think John wrote the
text.
Nick van: Or Erik. I'm not sure.
David Landwehr: Nick, did you
implement this function for 1.1?
Nick van: I started for Chiba but not
yet.
David Landwehr: I found it was hard to
understand. I'm curious if you think the complexity is high.
Nick van: It's a bit harder than the
version in 1.0.
David Landwehr: And the ease of use
for the author? We had actions at one point but it seemed easy to
use. I'm curious.
Nick van: It is more powerful.
Steven Pemberton: It doesn't actually
contradict itself. The same thing happens in two places.
Nick van: It does seem to be code you
don't need, and why is it in the spec?
Steven Pemberton: I've got you.
Nick van: It's not a contradiction. Am
I missing something?
Steven Pemberton: The good news is
that it's not a contradiction and is just an editorial point.
Nick van: All of my questions are like
this, not contradictions.
Steven Pemberton: Then I'll consider
these comments and we don't have to discuss them and we can ask
whoever wrote this text to clear it up? Anybody object to that
resolution of these issues?
Sebastian Schnitzenbaumer: [IRC]
fine
Resolution 2007-05-23.1: http://lists.w3.org/Archives/Public/www-forms/2007Mar/0005.html http://lists.w3.org/Archives/Public/www-forms/2007Mar/0005.html are editorial and should be addressed by John Boyer and Erik.
Steven Pemberton: Was I hearing a
suggestion from you that it was simplified?
David Landwehr: I have thought this
the whole time and I have my doubts. I haven't sent a comment. It's
more a question than a comment.
Steven Pemberton: If it's an
implementation problem then that will show up during CR.
David Landwehr: The question is can
people really use it?
Steven Pemberton: We should write some
forms to see if it solves real problems.
David Landwehr: They solve real
problems, but we know how it works. Will someone from the outside
be able to?
Mark Birbeck: I agree with you. Maybe
I'm just dim, but I have to consult the spec or the tutorial and
scroll down to the bit I want to do and cut and paste the code,
because I can never remember how to get a template from one
instance to another. There's too many rules.
Steven Pemberton: Mark, would that be
a last-call comment?
Mark Birbeck: OK. I'm not sure what a
simpler approach would be other than duplicate and destroy, and
making insert and delete be operators on nodesets.
David Landwehr: The last call comment
should just say that insert and delete are too complicated to use;
you don't need a solution.
Steven Pemberton: If you, Mark, have
to keep going back to the spec, then it seems to me that the
solution has failed we have to simplify it.
David Landwehr: I can write it as
well.
Mark Birbeck: It's hard, especially
with attributes.
Steven Pemberton: Thank you
both.
Leigh Klotz: Does the resolution still
stand?
Steven Pemberton: Yes, as we still
have the text, and have agreed the text needs work. It may be made
moot in the future, but not now.
Steven Pemberton: Should we handle
issues from Erik if he isn't here?
Nick van: [IRC] Let's handle others
first.
Steven Pemberton: No value-changed
is fired to the controls after the instance is replaced.
Leigh Klotz: Joern is this is the
chiba-state-changed issue?
Joern Turner: Yes.
Leigh Klotz: I agree, then. If my
control listens for xforms-value-changed it gets it when a setvalue
or user changes, but not when the computation is moved offbox and a
replace instance happens.
Nick van: You don't get one when the
instance is initially loaded.
Joern Turner: ...
Steven Pemberton: So you would like
the xforms-value-changed to happen even when the instance is loaded
initially?
Nick van: We have a custom event when
any MIP changes including value.
Steven Pemberton: I think the instance
replace case is where this doesn't happen. Mark?
Mark Birbeck: I'm not sure whether it
doesn't change or whether we just happened to have done it. We do
fire it.
Nick van: And what about the other
MIPs?
Mark Birbeck: Yes, we do.
Leigh Klotz: It's a problem when it
doesn't work.
Mark Birbeck: I think that it should
just happen with the r* events, refresh.
David Landwehr: We fire the events as
well.
Mark Birbeck: I thought submission
said to do a rebuild on the model.
Steven Pemberton: My feeling from the
discussion is that we agree that the events should happen on
initial instance and instance replace. Does anybody disagree?
Leigh Klotz: Which events?
Steven Pemberton: xforms-value-changed
but also xforms-invalid and the other related events. Does anybody
think those events shouldn't happen?
David Landwehr: It could be very
expensive on some devices to fire the events on all controls during
initialization. You can get refresh.
Mark Birbeck: You get refresh anyway.
We say to fire the events after done. The event firing is all in
the refresh phase.
David Landwehr: If that's what
makes sense for other people then I can live with it.
Nick van: I don't think it's clear.
Erik was quite convinced that you can't send the events if you
follow the spec, but Orbeon does send the events.
Steven Pemberton: We should see whether the spec requires it and what to do if we it does not. Joern, would you try to resolve this?
Action 2007-05-23.2: Joern Turner to investigate 'Value changes upon instance replacement' http://lists.w3.org/Archives/Public/www-forms-editor/2007May/0011.html including other MIP change events and report back to group.