Erik Bruchez, Orbeon
John Boyer, IBM
Leigh Klotz, Xerox (minutes)
Nick van den Bleeken, Inventive Designers
Steven Pemberton, CWI/W3C (chair)
Charlie Wiecha, IBM
Steven Pemberton: We have been
asked to discuss the upcoming TPAC, at Lyon, November 1-5.
Steven Pemberton: I need to fill out
our preferences for joint meetings, dates, overlap, etc. How many
will attend?
Charlie Wiecha: I will probably
attend.
Steven Pemberton: So, eight.
John Boyer: We may have new
members.
Steven Pemberton: Nick would prefer
Monday and Tuesday for flights.
Steven Pemberton: Are we flexible
about the meeting days?
Charlie Wiecha: Yes, as we can stay
afterwards.
Steven Pemberton: Is there any WG we
want to meet up with? No?
Steven Pemberton: Is there any
membership overlap with another group?
Charlie Wiecha: HTML.
Steven Pemberton: And how many would
attend the AC meeting? Just me.
Charlie Wiecha: Let's meet at 1 Rogers Street lobby at 8:30 AM Wednesday.
Leigh Klotz: I think we should
engage in discussion with Tim Bray on explaining how XForms is part
of the XML universe: XQuery, XSLT 2.0, XPath 2.0, etc. What HTML5
does is orthogonal, as XForms presents a path to the sea for XML
data.
Steven Pemberton: Who should
talk?
Leigh Klotz: I would be happy just to
make comments on his blog about this point, which I've made
before.
Steven Pemberton: OK.
Erik Bruchez: The conference seems
to have gone well. There was a talk by Liam Quinn on "The Future of
XML at W3C." I think the proceedings are interesting.
Nick van: The video presentations are
also online from last year; maybe this year will be on soon.
Steven Pemberton: We are in
Interaction Domain with HTML, WebApps, SVG. I wonder if we'd be
better placed in the XML domain. Would it matter?
John Boyer: I think it makes
sense.
Erik Bruchez: We mentioned it in the
charter, the affinity with XQuery, etc.
Charlie Wiecha: Sounds like a good
approach.
Steven Pemberton: I'll discuss it.
Steven Pemberton: There hasn't been
any news as W3C is busy with the AC meeting. It looks good.
Erik Bruchez: Where there formal
objections?
Steven Pemberton: I see, one supports
charter only if changes are adopted. "Alongside OCF, I also propose
ISO-SC34 liaison and for RelaxNG and NVDL schemas. ... XPath 2.0,
XML Schema 1.1"
Nick van: Leigh's working on the RNG
and NVDL is just a wrapper.
Leigh Klotz: Maybe he wants us to use
RNG for XForms instance validation.
Charlie Wiecha: I think it's just
schemas for editors.
Steven Pemberton: Any
objections?
Leigh Klotz: What about XML Schema
1.1? Does he want XML Schema 1.1 for XForms, or XForms support for
XML Schema 1.1 validation of instances?
Erik Bruchez: I'd like to contact him
and make sure. Some we plan to do. There are questions on XML
Schema 1.1. The biggest issue is the ISO liaison.
Leigh Klotz: I think the Mahamoud
might be the liaison, as he is currently the ISO-034 liaison on the
ISO side: http://www.w3.org/2001/11/StdLiaison
Steven Pemberton: All we may have to
do is mention the group. We've had official review from ISO-SC34
before, I believe.
Erik Bruchez: Doesn't that imply we'd
have to do some work with them?
Steven Pemberton: Only if it's called
upon. A liaison is for an external group that has some relation
with our work. It doesn't require meetings.
Leigh Klotz: It seems to me that he's
asking to do the work.
Steven Pemberton: Is everybody OK with
the changes? Please speak now if you have a problem with any of
them.
Resolution 2010-03-17.1: We agree with Mohamed Zergaoui's comments on the Forms WG Rechartering.
Steven Pemberton: Are now going to
invite experts?
John Boyer: We should wait until the
call for participation.
Steven Pemberton: Yes. Kurt Cagle, Dan
MacCreary, Alain Couthuries. Any others?
John Boyer: Draft forms WG charter
is done. Some aren't ready: errata, test suite.
Steven Pemberton: Some of Erik's need
to be removed.
Steven Pemberton: Since Uli's not here, let's do this at the F2F.
Steven Pemberton: The deprecation of DOMActivate has come up under quite active discussion two meetings in a row at the HCG. I don't see the advantage to deprecating it in the DOM Events Recommendation. It shows what to do in both cases and doesn't require DOMActivate.
John Boyer: Apparently we own XML
Events now.
Steven Pemberton: It's only a binding
to DOM Events.
Leigh Klotz: There's a lot to be said
for interoperability of other event users on XML.
Steven Pemberton: If we defined it in
XForms that would make several events, not necessarily
interoperably.
Erik Bruchez: It's a plain
event.
Steven Pemberton: You would have to
listen to multiple ones.
Erik Bruchez: I believe even if we
think it should be kept, we should follow the DOM decision,
whatever it is.
Steven Pemberton: The backward
compatibility issue is a problem.
Erik Bruchez: Our specification could
say something about it: we could mention DOMActivate support.
Steven Pemberton: What about mixed SVG
and XForms? SVG does DOMActivate as well.
Charlie Wiecha: It's a backward step
from getting into other modalities.
Steven Pemberton: We can discuss it at
the F2F.
Erik Bruchez: I think we discussed
this already by mail. It's a common requirement to load more than
one set of data during initialization. We can do it by launching
submissions, because you may reduce latency by calling in parallel.
The trick is that in many cases, we found that you need a way to
wait for submissions to complete. We have an extension action
called join-submissions; it waits for all submissions to terminate
and is a synchronous action.
Erik Bruchez: The code example
http://pastie.org/818536
lets you wait for submission to complete before before control
initialization, by doing the submissions inside
xforms-model-construct-done.
Leigh Klotz: Would a
instance/@submission instead of instance/@src satisfy this use case
without a new join?
John Boyer: We often get these done
server-side and get instance population done separately as an
instance.
Nick van: This is cleaner because you
don't have to thread yourself.
John Boyer: So you crank up the XForms
processor, the async submission, then xforms-model-construct-done?
If all submissions have to be done and joined, what's the point of
doing it within the lifecycle of the XForms processor instead of
server-side?
Erik Bruchez: XForms is an easier tool
to do this.
John Boyer: You can pre-populate a
form.
Nick van: I tend to pre-populate the
instance but everything is in two places.
Steven Pemberton: Hear, here.
John Boyer: It may not be fast.
Leigh Klotz: If they are GET then it's
easy to cache with cache-control and etags.
John Boyer: Then just use @src.
Leigh Klotz: You can't use forms logic
then.
John Boyer: The forms logic isn't
available until xforms-model-construct-done.
Erik Bruchez: Leigh's right about
caching. The submission gives you more power.
John Boyer: It depends on the order of
the initialization sequence.
Nick van: It can be the order of
send.
John Boyer: We don't specify the
order.
Erik Bruchez: The order might be a
problem if you use the attribute version.
Leigh Klotz: I don't agree that that
changes anything because the order of send is not the same as the
order of asynchronous submission response.
Erik Bruchez: OK but...
Steven Pemberton: We're out of time.
See you all next week.