John Boyer, IBM (chair)
Blake Jones, Daisy Consortium/ViewPlus Technologies
Charlie Wiecha, IBM
Jan Kratky, IBM
Leigh Klotz, Xerox (minutes)
Rafael Benito, SATEC
Roger Perez, SATEC
Sebastian Schnitzenbaumer, DreamLabs
Susan Borgrink, Progeny
Nick van den Bleeken, Inventive Designers>
Ulrich Nicolas Lissé, DreamLabs
John Boyer: I can't produce the agenda for next week and have talked to Sebastian.
John Boyer: Please register. Remind Steven to update the page; it still says "Joint HTML Forms Group" and fix the "Forms group"
Action 2007-05-9.1: Steven Pemberton (in absentia) to please update pages http://lists.w3.org/Archives/Public/public-forms/2007May/0022.html http://lists.w3.org/Archives/Public/public-forms/2007May/0022.html
Susan Borgrink: Do we have updates
on the member page?
John Boyer: We're transitioning away
from it to the public page. I need to check with Steven before we
remove it.
John Boyer: The status reflects the
new patent policy, and I made a change to the
QName-but-not-NCName.
Ulrich Nicolas: [joins]
John Boyer: Are there people who
would would want to work with this closely?
Charlie Wiecha: I was hoping Mr.
Raggett would be leading that.
John Boyer: He is a chair of another
group now, but remains involved and is still presenting on the
topic.
Sebastian Schnitzenbaumer: It may not
be time yet. It's required in the charter, so it must happen,
though.
John Boyer: Dan Connolly has been
active; there has been a proposal to use HTML5/WHATWG work products
as the "basis for review." By this, they mean checked into CVS and
undergoes changes to make recommendation track; in other words, it
is WF2 unless otherwise argued in their group. The questionnaire
closed and there hasn't been much objection. The most vocal folks
in the group argue that they don't want any editorial involvement
from the Forms WG; their idea of architectural consistency doesn't
include alignment of tagsets, just "they can repeat, we can
repeat." So there's no advantage to standardization.
Sebastian Schnitzenbaumer: We have to
wait and see.
Charlie Wiecha: Should we continue to
evolve XForms Transitional in the meantime?
Sebastian Schnitzenbaumer: I think
Dave should move forward with it. We said the group should go
forward with XForms 1.1 though, as Raman pointed out.
John Boyer: We do need someone to
participate at the spec level.
Sebastian Schnitzenbaumer: Yes, we
should do it. But the options are at the moment limited?
John Boyer: Please note the action
item rotation list, for assigning action items where it appears
nobody is the best qualified. It's like minutes; if you really
can't do it, please let us know. Does that really scare
anybody?
Charlie Wiecha: No more than anything
else...
John Boyer: Blake?
Blake Jones: OK.
Leigh Klotz: I sent a summary on
the list again. I think the "finding leaf nodes" and "eliminating
complex content" got conflated, with the result that empty leaf
nodes don't get submitted.
John Boyer: Anybody opposed to
changing to submit empty leaf nodes?
Resolution 2007-05-9.1: Empty leaf nodes do get submitted.
Action 2007-05-9.2: Blake Jones (assisted by Leigh Klotz) to write update eliminate pruning of empty leaf nodes on url-encoded submissions.
John Boyer: It seemed like there might be an informative note. The additional trouble is related to the fact that we don't have a good definition of form control. It does need to end up becoming an action item to propose additional text. Any volunteers? Although our model says that events for MIPs flow to form controls that express UI bindings, that doesn't mean that those MIP states do things to the controls; readonly on trigger, in particular.
Action 2007-05-9.3: Charlie Wiecha to produce short note for trigger about readonly.
Charlie Wiecha: It may not be only
trigger.
Sebastian Schnitzenbaumer: We have a
problem in Chiba that if a control changes binding, when it changes
back, you may not know. If you have a transcoder and change the
control changes to a date picker, then if you want to change it
back, you may not have the information any more. I'd like to ask
Joern about this.
Charlie Wiecha: That sounds a little
different than what I was talking about; but I'll bounce off you
what I propose to write.
Leigh Klotz: Is this related to the
problem of letting the user know that a submit button isn't ready
to be pushed again yet until the xforms-submit-done is done?
Charlie Wiecha: I don't think so;
that's more than trigger.
John Boyer: Yes, let's just put that
on the future features.
John Boyer: Some places a form
control is atomic, some places repeat, some places group and
switch. Repeat doesn't have a single-node binding so the MIPs might
apply to the items, not the repeat itself. There is a general
problem that needs to be looked at in detail, perhaps at a F2F. We
decided to leave it as is about six months ago, but we can revisit
that.
Sebastian Schnitzenbaumer: The form
control itself, the markup, is a like a macro. You can
programmatically change the behavior of the form control from the
model. In a transcoder, you need to keep track of two info sets,
the actual markup and the programmatic space in the model. The
question is how you deal with that superposition of information,
using a hashtable or a shadow tree.
John Boyer: I thought I had some
similar issues; that's not the same as "form control is not
rigorously defined." I've seen where people are trying to change
things on the client with transcoders on the server, and we have
had to detect those programming efforts and detect them and send
them back to the server.
Sebastian Schnitzenbaumer: Joern
thinks this is a language issue, not just an implementation
issue.
Leigh Klotz: Sebastian, I'd like to
talk to you about it as well.
Charlie Wiecha: This sounds like a
Mark Birbeck-kind of issue.
John Boyer: This sounds like an issue
for the next F2F.
John Boyer: I'd like to bring up
the XForms Basic test suite issues here.
Jan Kratky: Leigh went through the
test suite in the context of XForms Basic and we have a hand column
on the chapter driver pages for the test suite that tells you
whether success is required for XForms Basic. After going through
that list, Leigh sent me some that were marked as not applicable.
It led us down the path that, from our point of view, we'd like to
make sure that we're in sync with the spec. Leigh and I came to
some agreement that the spec language is vague on some of the
points, particularly around Schema processing.
Jan Kratky: It's the text of the
second bullet http://www.w3.org/TR/xforms-basic/#id2606183
Jan Kratky: This could be interpreted
as saying you don't support URI in Schema attributes or the Schema
element. So one possible interpretation is that Schema processing
isn't required. The other interpretation is that we are creating a
profile of XML Schema.
Leigh Klotz: [summarizes how we got
here]. Also, the big question is what PicoForms does, because if it
doesn't implement the Schema profile, then we can't go forward to
PR with that.
Jan Kratky: Yes. The next issue is the
xsi:nillable property; is it defined on instance elements at all?
And it doesn't say that the Schema processor should/shouldn't fail
if there is a full Schema present.
Leigh Klotz: Right, it shouldn't
fail.
John Boyer: Right, several questions
for David and Kenneth about their implementation. My own
recollection is that it would read the schema but only look for
simpleType definitions, but at this point, it's up to David and
Kenneth to tell us what they have. If we have to modify the
wording, we may have to go back to working draft. I wasn't aware
that nillable was a structural thing.
Jan Kratky: It's not on a type, it's
on a declaration.
John Boyer: We're at the top of the
hour here. We didn't get around to optimizing the test suite. Is
there some willingness to do more nature on this?
Jan Kratky: What Mark said is common
sense. I'll be uploading a draft of the 1.1 test suite with some
improvements in this area. Something out of the box, hooked into
Selenium or something else, would be great. I'm not sure I can sign
up for that.