Nick van den Bleeken, Inventive Designers
Erik Bruchez, Orbeon [left early]
Leigh Klotz, Xerox (minutes)
John Boyer, IBM (chair)
Uli Lissé, DreamLabs
visible from
dialoghttp://lists.w3.org/Archives/Public/public-forms/2009Jul/0045.html
John Boyer: I've got heard back from Chris Lilly and I hope the call will be this week.
John Boyer: In the XForms 1.1 document, I made 3 changes:
publication date and review period end date
John Boyer: For the important
changes list, one important change was the Combine attribute, and I
wanted to tweak the wording.
http://www.w3.org/MarkUp/Forms/specs/XForms1.1/index-diff.html#submit-header
John Boyer: Leigh, most of this is
your wording but I added the word "implementation."
Leigh Klotz: Sounds fine.
John Boyer: Aaron Reed needed to
configure this, but in JavaScript it may not be possible.
John Boyer: I also added "if supported
by the protocol implementation" under the combine attribute.
Leigh Klotz: What did we decide about
not setting it?
John Boyer: I recall discussion about
this but we had not added the text. If implementations don't allow
it, they don't. This seems to be easy to implement. It's done in
Ubiquity, though we don't have test suites for it. The merging with
the user agent we can't do. All we can do is provide headers to the
submission protcol implementation. We can't specify the combine
value then, but we process combine for the "custom" headers.
Leigh Klotz: Why not split it into two
sentences then?
John Boyer: And use the word "may" on
the second. I'll do that.
Nick van: [irc] that is also what I
did for Chiba
John Boyer: We have this for Ubiquity
and Chiba now. We don't have tests because append, prepend, and
replace are easy to understand. We also have the Firefox
implementation, which proposed the change.
Leigh Klotz: Aaron implemented
this?
John Boyer: No, but he implemented a
specific setting for combination=append in the Firefox internal
API. His API may be subject ot limitations. I asked if he would
consider an upgraded implementation.
Action 2009-08-5.1: John Boyer to split combine attribute optional part into separate sentence in http://www.w3.org/MarkUp/Forms/specs/XForms1.1/index-diff.html#submit-header
John Boyer: I thought Erik had one
completed action but the minutes are different.
Erik Bruchez: Yes, I said I had done
it.
John Boyer: Please send a note to Nick
saying it is done.
Erik Bruchez: I went through the
DOM2 attributes and proposed them for every event dispatched, and
types for DOM2 IDL and XPath 2.0 types. I talked with Nick about
the types for the properties. The time doesn't specify the epoch so
we use a number (integer or long). For XPath 1.0 types it matters
less. Also there were questions on naming and I used camelCase as
in DOM2. I added an id using camelCase. It's not spec-ready.
John Boyer: Event phase? Is that
common to be a number? We have names "capture" and "bubble" in XML
Events. For number, you sometimes get non-integer integers.
Erik Bruchez: I think 1, 2, and 3 are
OK in floating point. It is a mapping to the DOM2 value. We could
use a string.
John Boyer: The claim was that XML
Events had capture and bubble and you couldn't get AT_TARGET under
DOM2 events, and yet here it is.
Erik Bruchez: Yes, here it is in
DOM2.
John Boyer: We can get it here then,
if XML Events doesn't have it.
Erik Bruchez: Is there talk about us
talking on XML Events 2?
John Boyer: After 2009.
John Boyer: We can use the longer
names for now and put in as an editor's note that they might change
in XML Events 2.
Nick van: or use capture, atTarget,
and bubble
John Boyer: Or force the discussion
about the ugly names fix in XML Events 2. We should wind up with
better names.
Nick van:
http://www.w3.org/TR/DOM-Level-3-Core/core.html#Core-DOMTimeStamp
typedef unsigned long long DOMTimeStamp;
Erik Bruchez: XML Schemas has int as
well. Let me find the list of types.
Erik Bruchez: http://www.w3.org/TR/xmlschema-2/#built-in-datatypes
Erik Bruchez: They are a
hierarchy.
Nick van: int is good then.
John Boyer: Are there places where we
have non-integer numbers?
Nick van: ...
Nick van: We should use a general rule
instead of specifying each one by one in our spec.
John Boyer: Probably so. XPath 2.0
type for a timestamp should be xs:long.
Nick van: No, xs:integer holds more
than long.
John Boyer: So no action there.
Leigh Klotz: What will Erik change for
the capture and bubble strings?
Action 2009-08-5.2: Erik Bruchez to change http://www.w3.org/MarkUp/Forms/wiki/EventContextInfo event phase property to use capture, bubble, at_target to string with names from DOM2, but an editors note to coordinate with XML Events 2.
John Boyer: No change on id. camelCase looks good.
visible from dialogJohn Boyer: I propose just getting
rid of it the attribute.
John Boyer: And we don't have a reset
action.
Nick van: We don't have reset for
switch.
John Boyer: ...
Erik Bruchez: [leaves]
Nick van: Will it reset repeat and
switch as well?
John Boyer: I suppose it will now
because it is implicit instance data.
John Boyer: "visible" is a bad name.
It looks static. For switch, "selected" was a good name
originally.
Nick van: Maybe we can discuss
draggable as well. It's UI centric.
John Boyer: Anybody object to removing
visible? We have the option of doing show.
Leigh Klotz: What does Erik
think?
Nick van: He's left already.
John Boyer: It could be put in under a
different name.
Leigh Klotz: Why do this now?
John Boyer: If we remove those we can
do a thin-spec FPWD now.
Leigh Klotz: OK
John Boyer: OK, no objectios.
Resolution 2009-08-5.1: We remove @visible and @draggable from dialog for FPWD for XForms 1.2.
Leigh Klotz: I am getting ready to
test with RNC with jing and with nXML, both from James Clark.
John Boyer: What are the advantages to
you?
Leigh Klotz: Tools integration for
authoring, more readable Schema, and Sam Ruby's point about
validator.nu which requires RNG.