Steven Pemberton, CWI/W3C (Chair)
Charlie Wiecha, IBM
Erik Bruchez, Orbeon
Leigh Klotz, Xerox (minutes)
Nick van den Bleeken, Inventive Designers
2010-01-13 http://lists.w3.org/Archives/Public/public-forms/2010Jan/0017.html
Steven Pemberton: The call for review hasn't gone out yet. The regular phone call yesterday was cancelled. I presume it's being reviewed right now.
Steven Pemberton: We discussed DOM
3 Events plan to deprecate DOMActivate events; the WAI group has
also raised the issue. They don't like the suggestion that events
are device-dependent. I've now raised the issue with HCG, and I
imagine it will be on the agenda for HCG.
Leigh Klotz: And they want us to use
click?
Steven Pemberton: Yes. I guess you
could always catch click as it comes down the tree on root and
substitute DOMActivate for it. I'll raise it at the HCG. The
argumentation seems to be that "most HTML content uses click so we
should just standardize on click."
Leigh Klotz: This closes my action.
I've split it into XForms11 and XHTML+XForms11. We can replace the
XForms11 now, but the other we should continue to discuss. I don't
think John has made the directories for the unzipped versions
yet.
Steven Pemberton: Not yet.
Leigh Klotz: The first file
xforms11.zip needs to go in http://www.w3.org/MarkUp/Forms/2007/XForms-11-RELAXNG.zip
Action 2010-01-20.1: Steven Pemberton to rename file xforms11.zip from http://lists.w3.org/Archives/Public/public-forms/2010Jan/0019.html and place it in http://lists.w3.org/Archives/Public/public-forms/2010Jan/0019.html
Erik Bruchez: I reviewed DOM 2 and
XBL 2 and had a need to ensure that an event handle was being
targeted only in the target phase (described DOM 3) and wondered
why in XML Events we don't have an easy way to filter on the other
phases. We have capture, and default (activate handler on target or
bubble). It seems it would make sense to recognize all the phases,
and have only capture, bubble, or target in addition to
default.
Steven Pemberton: My opinion is that
if you think there is a use case, all XML events does is capture
syntactically what was in the DOM Events spec.
Erik Bruchez: Even DOM Level 2 has the
three values; in DOM Events you can filter yourself by phase. In
XML Events it's harder if you don't have a way of declaratively
filtering or having context information giving the current phase of
the event. I noticed in XBL, they have declarative event handlers
for XBL bindings, but not as general as XML Events. They have
similar attributes, and extra attributes and a few different
values. The phase attribute has new values, including
filtering on target and bubble.
Steven Pemberton: I don't see it in
DOM 2 Events; it's either capture or target and bubble.
Erik Bruchez: Once you receive the
event, you can test on the current phase; there's a little bit of a
difference. If we want XML Events to match addEventListener
strictly, my suggestion doesn't make much sense, since it doesn't
allow you to restrict. But if we want to provide a convenience to
authors, we can register and filter.
Steven Pemberton: Now I
understand.
Erik Bruchez: We could go the other
way and not provide the attribute values, but could provide context
information to extract the current phase. We do that in
Orbeon.
Erik Bruchez: You could imagine an
event listener on a group that listens to two input controls, that
you want to react to target phase on any descendent controls. For
example, detect xforms-enabled on all those controls. I'll send an
example to the list. It's different from ev:target; with ev:target
you specify one target element only.
Leigh Klotz: Maybe you could CC TV
Raman on the example message; he might have a different way of
looking at this based on his understanding of capture and bubble as
meta-object operations that will provide a good synthesis.
Steven Pemberton: XML Events is on our
new charter. Do we resolve to accept these or accept it as an idea?
I'm happy with adopting these two new values.
Leigh Klotz: I'd like to hear from
Raman first.
Steven Pemberton: Who will contact
Raman?
Leigh Klotz: Can Erik just CC him on
the message.
Action 2010-01-20.2: Erik Bruchez to write example use case for proposal to extend filtering on the event phase http://lists.w3.org/Archives/Public/public-forms/2010Jan/0018.html and send it to public-forms and TV Raman for comment.
Leigh Klotz: Micah Dubinko wants to
bind a select1 to either an element with varying text, or to the
absence of the element. I proposed a couple of ways to do it, but
none are very good. I thought it might be good to look at for
XForms 1.2.
Steven Pemberton: Personally I think
this may be part of a wider consideration for mixed-content. You
know the elements but not the order. It's like a union type.
Leigh Klotz: You may also start with
instance data not having the element at all, and it needs to be
added.
Erik Bruchez: If Micah's question is
whether something has changed, no.
Leigh Klotz: He asks about it for
XForms 1.2.
Erik Bruchez: I think you have to go
through hoops to do this using insert and delete.
Leigh Klotz: The answer is then either
"no" or "not yet."
Steven Pemberton: It may be related to
relevance.
Leigh Klotz: I had two hacks for doing
this: insert and delete with @if, and the hack ascribed to Mark
Birbeck for relevance-during-submit.
Erik Bruchez: It may be related to
copy.
Leigh Klotz: But you can't bind to the
empty spot.
Steven Pemberton: I see this as XForms
2.0, and quite agree.
Leigh Klotz: So you want to put it on
the 2.0 list?
Steven Pemberton: OK
Leigh Klotz: The
relevance-during-submission hack might be something to add
in.
Steven Pemberton: It's ok for when the
data is there, but it doesn't help.
Leigh Klotz: The
relevance-during-submission is the hack that is the dual of the
submission/@relevant attribute.
Steven Pemberton: I see.
Erik Bruchez: With complex
applications, you can generate something with buttons for editing a
complex schema. That's more complex, a whole subtree. I don't know
if that's something else to be considered. Is there really
something else to be done. And use trigger and all the details. If
you write generic tools to create from schemas you can do it that
way, but if you want to do it once, it's difficult.
Steven Pemberton: I think this ought
to be easier than it is, but I'm not sure it's for 1.2.
Erik Bruchez: It depends on whether
there are any good suggestions for how to improve it. If Micah has
a good suggestion, I'd like to hear it.
Leigh Klotz: I don't think we need to
answer specifically, but we can put it on the 2.0 list.
Action 2010-01-20.3: Leigh Klotz to add Binding to optional element http://sourceforge.net/mailarchive/forum.php?thread_name=C006B451-4939-41D9-B4B4-92373D9D21AA%40marklogic.com&forum_name=xsltforms-support to XForms 2.0 requirements.
Leigh Klotz: This was the issue Uli
raised about whether we use one event with context, or two
events?
Steven Pemberton: Any pointers?
Leigh Klotz: Let's raise this again
when John comes back and we can capture the state.
Erik Bruchez: I had an action item to
document some of the discussion we had on the wiki. I haven't done
that yet.
Charlie Wiecha: I pinged Kenneth
off-list. I'd like to implement this in Ubiquity. A normative
definition of the mapping would be useful.
Steven Pemberton: Getting JSON into
XML could be easy.
Charlie Wiecha: It's underdefined;
attributes or elements. People need to count on it.
Steven Pemberton: We just have to
specify how it should be done.
Leigh Klotz: Or use a QName for naming
the mapping with a few as we did. That decouples it from the actual
mapping, so you can still benefit from Kenneth's experience with
PicoForms but you can start now on Ubiquity.
Steven Pemberton: I'll contact
Kenneth.
Action 2010-01-20.4: Steven Pemberton to contact Kenneth Sklander about JSON to XML mapping.
Steven Pemberton: The other
direction is harder.
Charlie Wiecha: The other direction
isn't always tractable. It may be possibly but lossy, for
collecting elements with the same name and putting them into an
array.
Leigh Klotz: Naming them serialization
scheme would then let you know what type of fish you're
getting.
Charlie Wiecha: Or we could hook in
and let people do other stuff.
Steven Pemberton: So the response is
that we can't do it until we have some more details.
Leigh Klotz: We can look at the
submission attributes.
Charlie Wiecha: Kenneth's would be
part of a family; Leigh suggests extensibility.
Steven Pemberton: http://code.google.com/p/xml2json-xslt/
is an XSLT stylesheet.
Leigh Klotz: The lossless ones are
like DOM APIs coded in JSON syntax.
Nick van: http://www.json.org/javadoc/org/json/XML.html
Steven Pemberton: I see.
Leigh Klotz: Is submission the only
place? How about instance/@src.
Steven Pemberton: I see it as a
simplified version of submission.
Charlie Wiecha: And an inline version,
a script element child of the instance element. And ideally some
way to do insert and delete.
Leigh Klotz: Or just an id on a script
tag. You run into use-mention problems if you want to put a script
inside an instance.
Charlie Wiecha: There's a broader set
issues about submission.
Leigh Klotz: For the simple case,
accepting JSON, we need that on submission and where else?
Steven Pemberton: It would be maybe
with @replace.
Leigh Klotz: deserialize or
parse?
Steven Pemberton: instance/@src and
submit/@replace
Charlie Wiecha: Those look like the
two main places.
Nick van: That doesn't document it,
just the implementation. http://www.json.org/java/org/json/XML.java
Leigh Klotz: I think there are
multiple ones.
Charlie Wiecha: JSON.org has different
ones with different policies for replicated data.
Erik Bruchez: http://www.bramstein.com/projects/xsltjson/
Erik Bruchez: This describes
conventions and names them: badgerfish
Charlie Wiecha: Yes, these are like
the QNames Leigh is talking about.
Action 2010-01-20.5: Charlie Wiecha to explore one or more JSON serializations from http://www.bramstein.com/projects/xsltjson/ with an extensible @fish attribute.
Nick van: Where is it?
Charlie Wiecha: I am confident I can
provide a space at IBM in Cambridge.
Steven Pemberton: I had an action to
contact W3C. There is currently space at MIT, but we might not get
the same room all day. The only question would be about sponsorship
for catering.
Leigh Klotz: If it's lunch we can eat
out; is it coffee?
Charlie Wiecha: It's not that big a
deal.
Leigh Klotz: Where is the IBM site?
Not Tech Square anymore?
Charlie Wiecha: It's on Roger's
Street
Leigh Klotz: http://www.watson.ibm.com/general_info_cam.shtml
Steven Pemberton: I'll reply to MIT
then.
Charlie Wiecha: Can we confirm the F2F
before the charter is done, as Nick asks.
Steven Pemberton: I don't anticipate a
problem. We've got an in-principle extension. It will surely extend
until the charter. But we had hoped to have the meeting under the
new charter.
Charlie Wiecha: So we could slip the
date.
Steven Pemberton: But then we lose the
combined travel.
Nick van: And I need two months notice
to get flights.
Steven Pemberton: I'll try to get some
idea for the schedule for rechartering.
Action 2010-01-20.6: Steven Pemberton to check on rechartering schedule and F2F.