Steven Pemberton, CWI/W3C
John Boyer, IBM (chair)
Uli Lissé, DreamLabs
Charlie Wiecha, IBM
Leigh Klotz, Xerox (miutes)
Erik Bruchez, Orbeon
Nick van den Bleeken, Inventive Designers
http://lists.w3.org/Archives/Public/public-forms/2009Jun/0030.html
Steven Pemberton: I have annotated
the changes and would like an extra check from you.
John Boyer: I'll have a look at that
and get back to you.
Steven Pemberton: Please make sure you
agree with the contents as your name is at the bottom of it as
well. As soon as you say it's OK I'll send it off.
John Boyer: I'll double check on the
"should" and "must", submission header. I'll do another scan
through the diff marks.
John Boyer: At the end, we thought
we'd send a comment to XHTML2 but I'm not sure that comment got
produced. I think it was about not being able to set context
information. We need someone to do this.
John Boyer: There exists context
information for events; while you can dispatch events you can't set
context information.
John Boyer: And, when handling the
event, there must be a way to access the information (perhaps the
spec already has this).
Steven Pemberton: If I recall, it does
mention the context information.
John Boyer: How does a handler get
access?
Action 2009-06-17.1: Uli Lissé to send comment to XHTML2 to ask for XML Events 2 dispatch-event to be able to set context information, and also access context information in a handler if that's not currently possible.
John Boyer: Can't we just say to
Aaron Reed that events go away from the queue as the runtime
elements are destroyed, such as form shutdown or repeat changes, or
switch changes? So then if the event queue is empty, why do I care
when I get rid of it? Any problems?
Leigh Klotz: Maybe this is really a
memory management question.
John Boyer: Or what if there are
events on the delayed event queue and the queue exists after the
model-destruct event is still dispatched? Maybe he wants to know if
the runtime elements still exist, which is a good but separate
question.
John Boyer: If model-destruct happens
and form controls exist and a delayed event happens, what is the
harm in it running? It may not be able to get at anything, or do
all its actions.
Leigh Klotz: Is model-destruct an
information-only event?
John Boyer: It's notification; you can
still do stuff during model-destruct. We have a test for
that.
Leigh Klotz: That may be the answer;
it's OK to run actions during model-destruct.
John Boyer: xform actions can still be
executed (load, submission, model-based ones) when during capture
or bubble of model-destruct. We don't test UI ones as the UI may
already be gone in some implementations. We haven't been strict
about when the UI disappears relative to model-destruct. I'm not
sure we need to be more specific about the delayed event
queue.
John Boyer: I think we didn't define
it but it's reasonable to assume that no further events happen
after model-destruct, but we didn't specify it. Does anyone think
that's not a good enough answer? So our answer is we don't specify,
but it's safe to assume that after model-destruct event dispatch is
finished, no further delayed events occur.
Charlie Wiecha: [irc] +1
Action 2009-06-17.2: Leigh Klotz to respond to Aaron Reed http://lists.w3.org/Archives/Public/www-forms/2009Jun/0006.html saying we didn't define it but it's reasonable to assume that no further events happen after model-destruct. It's safe to assume that after model-destruct event dispatch is finished, no further delayed events occur.
Leigh Klotz: Markus Gylling has
updated the XForms 1.1 schema for RNG. I'd like to publish it as
RNC. We can publish RNG as well. Then I'd like to merge it with my
integration of James Clark's nXML for XHTML 1.0 and use it to test
our test suite. Markus has delivered an integration also with
XHTML, but I'm not sure version.
John Boyer: Please check to make sure
the switch/case local-global change has been made.
Leigh Klotz: OK.
John Boyer: We had backplane
concepts, markup on the glass, aggregate applications, XForms for
other contexts (ODF, non-XML web pages), extensibility (custom
things, JSON), architecture/language features (incremental changes,
larger patterns, dialog). Can we formulate a new charter?
Charlie Wiecha: I find something
missing: a version of XForms that could become a module of HTML6?
Perhaps that's the non-XHTML goal?
John Boyer: Yes, under the broader
contexts.
Charlie Wiecha: Can we add XForms for
HTML* there? An MVC version, different from XForms for HTML.
John Boyer: I'll add that now.
Charlie Wiecha: A variant with non-XML
instance, non-XPath, etc. Modularity around expression language and
instance.
Steven Pemberton: We've talked about
instance and I don't see any problem that at all. Changing the
selector language makes content no longer compatible.
Charlie Wiecha: Are there islands of
compatibility with strong interest in islands but not across? We
found CSS selectors in the Ample SDK. It would be a shame to lose
the pattern over a subdialect of selector languages. People keep
re-inventing XForms.
Nick van: XPath 1.0 and XPath 2.0
could be a start.
Charlie Wiecha: Yes, maybe it's a
pluggability theme, as with Schema languages. An abstraction of
basic pattern vs. particular technologies.
John Boyer: How closely do we need
to connect these to XForms 1.2 and XForms 2.0? The "transitional"
was from the last charter.
Charlie Wiecha: You mean map themes to
deliverables?
John Boyer: The set selected for
XForms 1.2 and 2.0 and then describe the selection criteria.
Charlie Wiecha: What's the time
frame?
Steven Pemberton: Our charter ends at
the end of this year so in the fall we should have something
ready.
John Boyer: We need to write a draft
charter.
Leigh Klotz: Does Steven do
that?
Steven Pemberton: Perhaps.
John Boyer: Steven and I can bounce a
draft around.
Action 2009-06-17.3: Steven Pemberton and John Boyer to work on draft Forms WG charter.
Erik Bruchez: Simplification for
authors. We had confusion about XForms for HTML and Simplification
for XHTML authors, and we agree simplification was worth doing.
That's important to me. Syntactic simplification that makes it
easier for authors to get started, including MIPs on controls. Is
that markup on the glass?
John Boyer: That's integration of
components and clients; I should move that elsewhere. Charlie can
you fill out more?
Charlie Wiecha: Maybe lifecycle of nodeset bindings. We can discuss it at the Backplane.
Action 2009-06-17.4: Charlie Wiecha to lead Backplane XG discussion of potential additions for Forms WG Rechartering.
Charlie Wiecha: Is the HTML(k)
modularity interesting? Snapping in modules unobtrusively of
interest?
Leigh Klotz: I'm interested if it's
going toward implementation.
Charlie Wiecha: Is it too far
afield?
John Boyer: I don't think so. Data
dependencies, for example, are challenging for a new expression
language. Whether we use namespaces or not, XHTML2 has adopted our
elements without our namespace. The syntactic means may require
some spec work but won't be very difficult.
John Boyer: Back to Erik's MIPs
point; is it its own theme?
Erik Bruchez: Regardless, it's worth
mentioning.
John Boyer: How about defining
XForms abstractly? Applying XForms to broader contexts?
Charlie Wiecha: Yes, or a separate
theme. Or architectural advancement.
Leigh Klotz: Yes, it's
architectural.
John Boyer: Model driven switch, other technical topics.