Charlie Wiecha, IBM
John Boyer, IBM (chair)
Leigh Klotz, Xerox (minutes)
Nick van den Bleeken, Inventive Designers
Keith Wells, IBM
Erik Bruchez, Orbeon
http://lists.w3.org/Archives/Public/public-forms/2008Sep/0037.html
Charlie Wiecha: We're going to talk
about modularity and streamlining, then give examples of
composition in SMIL, Voice XML, and ODF with Forms, using a data
model for composition. Then we'll discuss MVC patterns for wrapping
calendar pickers and color pickers in YUI, showing how it's a
general techique.
John Boyer: The MVC-Connector
architecture as well.
Leigh Klotz: Is there something you're
going to publish?
Charlie Wiecha: I've got a draft.
We're beefing up the backplane wiki.
John Boyer: In the past, the Tech
Plenary talks have been put up in the W3C space as well.
Leigh Klotz: Nope. Did you get my
submission headers changes in yet?
John Boyer: Not yet.
Nick van: We just released a new
version of our product using XForms.
John Boyer: Can you provide a news
link?
John Boyer: The XML Security
meeting is 11am-12:30. We're not on with CDF for mustUnderstand as
they're in flux (there's no chair). The WebApps WG refused our
meeting request; I tried several times to appeal the refusal.
They've said that they don't want to talk about anything that's not
in the next few months, and that we should join their WG and solve
their problems.
Charlie Wiecha: And HTML?
John Boyer: I've asked on Thursday and
Chris Wilson said yes. The chair didn't ask the membership, and
only four said yes.
Charlie Wiecha: Monday at 4:00 as
Backplane with MMI, and then Protcols and Formats Tuesday after
lunch.
John Boyer: And forms?
Charlie Wiecha: I've asked for it as
Backplane; I'm not sure there's much overlap with MMI. We're going
to go through Ubiquity, the voice example as a composition pattern,
streamlined behavior and MVC.
John Boyer: The 4:00 slot will have
more teleconference members so we should do Forms business, but the
morning A11Y meeting might be a better joint meeting.
Charlie Wiecha: That makes sense.
John Boyer: I've changed the name based on last week's discussion. The three thrusts are
Declarative AJAX
John Boyer: There are a lot of
examples, and all are HTML forms with attribute decorations.
Leigh Klotz: I think we need a new
holistic title; this isn't about forms, and XForms hasn't been
about forms for seven or eight years now.
Charlie Wiecha: Instead of Rich, Data
Rich.
Leigh Klotz: I'm thinking of more Web
on Rails or Form Rails.
John Boyer: We can't really get rid of
Forms, because we're the Forms WG.
Leigh Klotz: ...
Erik Bruchez: I think it's fine to
play on the same level as HTML WG and promote alternatives, but
what we are doing is useful separately. IBM's products, John's for
example, don't require a browser. XForms addresses more complex
problems than what Web Forms 2.0 is trying to do. By positioning
this as a competitor we're just going to lose. There's a huge
difference in the complexity of what you can do. For a form with
300 fields and calculation, you need a data model.
John Boyer: Yes, it apparently is a
mistake to think that we're in a separate space. When our WG was re
chartered, the mandate from W3C was that we've gotten traction in
the "big forms" space but try to drive broader adoption in the
larger web community. So, how do we apply principles of software
engineering to the large monolithic system we have. There's a lot
of human investment in understanding how HTML forms work, so we
need an onramp. This document addresses the ten-field form better
than Web Forms 2.0, but it has an underlying data model. Now that
you have attribute decorations, you can begin to adopt XForms
element modules one by one as well.
Erik Bruchez: If it works, it's
great, but what's the next step? I think it's a good idea. But once
it's out there, who's going to making the next step to support more
elements?
John Boyer: The complaint is that
XForms is too difficult and too different. So we need to get around
the complaint that "it's not available in my web browser." We had
some plugins for Firefox, but they didn't feel it was native enough
because it didn't ship by default. Others complained because it was
only one browser. So instead of taking a direct approach, take the
approach of YUI and Scriptaculous. I've talked about this at ACM
and other places. We'll get implementations out there; Ubiquity
will be one and we'll need another.
Leigh Klotz: Why? Isn't one
open-source project that works in five major browsers good
enough?
John Boyer: That's a good point; they
are separate implementations for different browsers.
Leigh Klotz: "Forms Plus A: Enhancing Dynamic Web Pages with Layered Attributes"
Charlie Wiecha: When is the
publication deadline?
John Boyer: I think a week before. We
can present it as internal working group content.
Leigh Klotz: I think this document
doesn't address the intended audience. I'd say try to pick out the
examples and put them in a more narrative document like Steven's
"XForms FAQ for HTML Authors."
John Boyer: Our charter says to "draw upon the the Web Forms 2 work and create something that is integrated with the Web Architecture." http://www.w3.org/2007/03/forms-charter.html So we need to draw upon this and take an approach that the HTML group can use.
Leigh Klotz: To appeal to that
group, you'll have to start on the glass.
John Boyer: I agree; the example
content is the driver for the talk.
Charlie Wiecha: Mark has an example
of using Dave Ragget's Slidey with embedded Ubiquity code. You
might be able to embed executable WebForms-A code.
John Boyer: Yes, I realized it might
be possible through Ubiquity right now, to show the form, data
model, and corresponding forms, and have the behavior and have the
spec become an interactive document.
Charlie Wiecha: So in the FAQ version
you might be able to do that.
John Boyer: It would be a different
kind of spec from any W3C spec.
Charlie Wiecha: Don Knuth used to talk
about that kind of stuff, but this is on the web.
John Boyer: Maybe we can discuss some of the technical details.
John Boyer: There are three names:
name, type, check. I changed the prefixes to wfa-. It's not clear
we need that but it can make it easier. What do you think?
Charlie Wiecha: Who would use
these?
John Boyer: The examples are after
initialization:
http://www.w3.org/MarkUp/Forms/specs/XForms1.2/modules/streamlined/index-all.html#default-data-instance-initialvalues
John Boyer: It looks like HTML;
it's not even well-formed XML. We're layering semantics on top of
that.
Charlie Wiecha: Who uses the
hyphenated versions?
John Boyer: I want to get rid of them;
they seem like a fallback plan to avoid any conflicts.
Erik Bruchez: ...
John Boyer: In section 1, I talked
about local and global attributes. I mention that a host language
may adopt them without the prefix; maybe we should instead say they
can adopt them with the prefix. Since there are all these
variations, in the spec I'll use the namespace-qualified version in
the spec and the non-qualified version in the examples. But it
seems like it's a good idea to get rid of the wfa- business.
Charlie Wiecha: It clutters up the
text.
John Boyer: I'll do that.
John Boyer: Back to the tables,
section 2 is about indicating forms and it just has "form." Section
3 talks about the default data instance. You could put
'serialization="application/xml"' and that one
attribute gives you 'replace="all"' XML submission
from XForms. With a couple of more attributes, you could submit
only part of a document. To graduate up from that into doing
instance replacements, moving a couple of attribute decorations to
the form tag to the submission button does this. That plays on
Nick's work with submission attributes on the XForms submit
control. They could easily be placed on an input of input of type
submit.
John Boyer: In section 3.2, http://www.w3.org/MarkUp/Forms/specs/XForms1.2/modules/streamlined/index-all.html#default-data-instance-initialvalues it shows you the produced instance data. Next is a fieldset.
John Boyer: Later down we show
duplicate names with radio buttons and checkboxes. I don't know if
there are other occurrences of duplicate names but we can get that
during the review.
Leigh Klotz: Reusing the name is one
way to do a repeat.
John Boyer: We have a startsize
attribute in section 3.3. Our underlying model lets us solve
repeats without this problem. I understand that the repeat module
has been removed from Web Forms 2.0; that's interesting considering
that it can be solved with a single attribute.
Keith Wells: In the second example,
shouldn't checked have a value?
John Boyer: It fleshes itself out in
the submission. My understanding of x-www-form-urlencoded is so
XForms can be used in lieu of HTML Forms. It does need to be looed
at further.
Keith Wells: I can.
John Boyer: In section 3.3, see
that the startsize attribute indicates that the tr is the repeating
element (not its content as it says). The runtime produces multiple
copies of the content. During instance creation, the startsize
shows how many are to be created.
John Boyer: Next, we start with an
xforms group whose ref is order; it has the same tables, header,
and thead. Rather than having a tr, I have an XForms repeat around
the tr. I think that's allowed.
Nick van: I think it's not
allowed.
John Boyer: Not by the HTML DTD or
XHTML Schema but what are people actually doing? I understand there
are only shallow implementations of attribute repeats. Keith?
Keith Wells: I've seen it done with
styling to be table rows and cells. I've seen repeat with table
rows under them like you have, but mostly people try to do it
through CSS styling. You style the repeat item as if it were a
table-row item.
John Boyer: How does this look
written?
Leigh Klotz: http://en.wikibooks.org/wiki/XForms/CSS_tables
Nick van: Does it work in IE?
Leigh Klotz: That's more a Ubiquity
question; John's question was about how it works in Firefox XForms.
IE quirks aren't the question here.
John Boyer: It has the loan form
here as markup. My hope is the have the xpath expression context
and variables ($, no $) done.
Leigh Klotz: In your talk, I'd start
with the examples from the end and show maybe a couple of the
areas.
John Boyer: This is part of Charlie's
talk so there's only one example.
Leigh Klotz: So can you consider
WebForms+A instead of Webforms-minus-a?
John Boyer: Good point; I was
considering it a hyphen.