Steven Pemberton, CWI/W3C (chair)
Uli Lissé, DreamLabs
Leigh Klotz, Xerox (minutes)
Erik Bruchez, Orbeon
John Boyer, IBM
Philip Fennell, MarkLogic
Nick van den Bleeken, Inventive Designers [late]
Steven Pemberton: Please register before the 22nd. One other person is arriving; Joern Turner is suggesting Lars Windauer.
Leigh Klotz: Dr. Ray has announced this forum software is available under BSD-like license. It's almost all XForms but some XBL1.
Steven Pemberton: Erik, please
suggest topics.
Erik Bruchez: I would like to see what
the group's position is for next steps. Some features are proposed
and have significant content, such as Nick's XPath 2.0 page. So
what is the process?
Steven Pemberton: I need time, but we
said we would use the wiki for the specs. I produced a wiki version
of the spec (except for internal links). The next steps are
re-factoring the 1.1 spec to start editing it, and getting the
software to convert wiki to convert the wiki to HTML.
Erik Bruchez: How would refactoring
1.1 work? I thought we had proposed ...
Steven Pemberton: I mean re-factoring
the spec for 1.2, splitting it out into independent pieces.
Submission, controls, events, and so on.
Erik Bruchez: Do we need to do that
absolutely.
Leigh Klotz: I thought Erik was asking
about 2.0?
Erik Bruchez: All the future features
are XForms 1.2. It seems like that's what we're doing. So my
question is about immediate next steps.
Steven Pemberton: We're doing the text
in wiki is until we get the spec to start editing. We could decide
not to break the spec up into chunks, but I think with many editors
it will make life harder. Editing the whole spec is enormous.
Erik Bruchez: For the XPath 2.0
sub-spec that Nick did, I'm not sure what I think about mixing
things back into the big spec. There was an appealing notion we
could say how we'd do it in XForms, and how we amend the 1.1 spec.
Is that something we're still thinking about? The text on the wiki
sounds great, but is merging still the final step?
Steven Pemberton: You mean a thin
spec, describing the changes?
Erik Bruchez: I thought that was
discussed. There's a lot on the plate. There's agreeing on things,
which is time-consuming. There's editorial work.
Steven Pemberton: It seems to me
breaking it up and adding to it is less work.
Erik Bruchez: So a wiki version of the
1.2 spec?
Steven Pemberton: It's already
converted to wiki format. The struggle is with internal references.
The next step is to produce the final spec; we have some software
that another group used but I haven't tried running that yet.
http://www.w3.org/MarkUp/Forms/wiki/XForms_1.1_in_Wiki
Erik Bruchez: So for now we just work
on the individual sub-pages for items.
Steven Pemberton: Yes
Erik Bruchez: So we need people now
working on individual features. I know a few were started but none
are completed yet. We have some big ones with no conclusion:
improved UI events, a lot of time, but we haven't agreed on basic
things like what relevance means. It seems like more discussion is
needed. case function is fairly simple. Some need finishing, such
as XPath 2.0. I have a few myself. Are we clear; are people owning
features?
Steven Pemberton: In the absence of
one single editor we'll have to split it. We need champions for
each feature, responsible for that section.
Erik Bruchez: I'm particularly
interested in variable, xpath 2.0, input ui events, custom xpath
functions. We might have some orphan features.
Leigh Klotz: Charlie was doing JSON
submission.
Erik Bruchez: How are we going to put
a cutoff point?
Steven Pemberton: We have a high wish
list. If they get written, we include them. Strictly speaking, we
should have our first WD now. I'm hoping that at the end of the F2F
we'll have solidified what is in 1.2 and have the path.
Steven Pemberton: The important thing
we should do a definitive triage for 1.2 at the F2F, and have a
list. And the introductory paragraph of differences with XForms
1.1, plus owners for each part, a draft within a month.
Erik Bruchez: Is there a F2F
agenda?
Steven Pemberton: We need that in the
next couple of weeks.
Erik Bruchez: I'm trying to get the
ball rolling. We have some unresolved discussion from last year,
especially the discussion on basic XForms notions such as
relevance. It's not really a 1.2 feature; I was hoping for UI event
revamping. I don't know, but postponing it seems risky. The sooner
we agree on core revisions of relevance and visibility the better.
It's hard to discuss on the phone.
Steven Pemberton: A lot of people at
the F2F will be on the phone. Do we want to have 1.2/2.0 split
evenly?
John Boyer: I want longer for
1.2.
Steven Pemberton: Then 3 days for 1.2
and 1 for 2.0? OK. So maybe we can hash out the agenda for the F2F?
Or would you rather talk about some of the issues now.
Steven Pemberton: There's the spec mechanics, but also issues. How about relevance and visible?
Erik Bruchez: You said it wasn't what relevance was about initially. Some non-relevant stuff in my product is visible. We didn't resolve anything. But we need to resolve that so we can revise the UI events. Our product essentially equates relevance with visibility, but we couldn't agree on that.
John Boyer: The only way I can
figure out the spec saying a lot less if we remove the concept of
rebuild altogether and let the implementation figure out when to
rebuild.
Steven Pemberton: So declarative style
and rebuild, and JSON submission aren't in the list.
John Boyer: Is it larger than JSON
submission?
Steven Pemberton: JSON round
trip.
Erik Bruchez: The most frequent use
case was calling a JSON API service.
John Boyer: But XML on the
inside.
Steven Pemberton: I've suggested other
possibilities for that as well. It doesn't just have to be JSON.
Other common data types like VCAL.
John Boyer: Yes, we got pretty
far.
Erik Bruchez: I think I kind of owned
it. I don't know if there was anything controversial. We had a
proposal with a simple solution: a new function as an XPath
expression. http://www.w3.org/MarkUp/Forms/wiki/Custom_XPath_functions
Erik Bruchez: We call it default
but that's a bad name.
Steven Pemberton: It ensures that the
element or value exists?
Erik Bruchez: No, it gives a value
only if it doesn't exist.
Steven Pemberton: If you want to bind
to something that's not there, you have to create it. There ought
to be a declarative way of binding.
John Boyer: You mean a schema with a
choice?
Steven Pemberton: Binding to
attributes that aren't there?
John Boyer: Mini-lazy authoring, that
creates the node.
Leigh Klotz: We have insert.
Steven Pemberton: It's not really
declarative. You wait for the event and insert it if not
there.
Leigh Klotz: If you have four A
elements in a row and have an input bound to an attribute on A,
just using a MIP to declare an attribute into existence doesn't
work. Insert at least has the hooks to let you decide what to do.
We may want some syntactic sugar but insert has the design.
John Boyer: The big brother is the
optional content; you may have cash, credit, COD. Then you insert
them all and use relevance to hide the others. That creates a
disconnect between the live-running instance and the schema-valid
instance. If you reload a saved instance from the form, if you
reload the saved instance you've lost the COD option.
Steven Pemberton: Let's just note this
issue and move on.
Steven Pemberton: model/@src
John Boyer: model enhancement
requirements, src on model, encapsulation, xinclude?
Leigh Klotz: XBL
John Boyer: It's easy to say you
should use XInclude; that's like saying "Use XML Events." Do we ask
processors to support XBL?
Leigh Klotz: I don't think it's as
embedded as XML Events.
John Boyer: Like "MAY"
Leigh Klotz: Like a best practice. I
don't think we need core changes, but I think we need to write
something and publish it. We still have to debate it and decide
what to do.
John Boyer: Dialog.
John Boyer: It removes a significant chunk of the model element. It's a behavioral change; the event and processing will go away.
Steven Pemberton: Anything
else?
Leigh Klotz: I think we should go
through the XForms12 category at the F2F.
Steven Pemberton: Yes, but anything
missing?
John Boyer: For model-based switch,
Michael Sperberg-McQueen says we need the CSS equivalent of
repeat-item for switch cases.
Steven Pemberton: switch-case
styling
Steven Pemberton: Anything
else?
Leigh Klotz: We should add XBL to the
list.
Steven Pemberton: So to add things to the agenda for the F2F add them to Category XForms12.