W3C Forms teleconference August 27, 2008

* Present

Charlie Wiecha, IBM (co-chair)
Erik Bruchez, Orbeon
John Boyer (co-chair, phone problems)
Kenneth Sklander, Picoforms
Leigh Klotz, Xerox (minutes)
Nick van den Bleeken, Inventive Designers
Paul Butcher, x-port.nte
Rafael Benito, SATEC
Steven Pemberton, CWI/W3C
Uli Lissé, DreamLabs
Keith Wells, IBM

* Agenda


* Previous minutes

* Upcoming telecons

* Rich Web Application Backplane

Charlie Wiecha: When we kicked things off, we rallied around the Ubiquity code library and evangelize its use and show its potential for integration into AJAX libraries such as Dojo. We looked at the submission and instance modules with just XML data, for example querying flickr and picasa. So we have a REST call for querying the data and then wrapping the Dojo data source API with XForms markup, and then wrapping up existing widgets such as the YUI keystroke box. The general idea was to approach people familiar with those frameworks and show a higher level of abstraction, and show how a small change could re-purpose from flickr to picasa, for example. We do have an XForms UI layer, and have been working on that, using xf:output as a YUI calendar. Mark had a pseudoelement under the XForms control, to expand to render. The xforms control handles the data binding. It's a clean way to expand XForms for styling. That's all working nicely.
Charlie Wiecha: The latest theme is ARIA enablement for those calendar widgets to show to the Protocols and Formats group and A11Y. [Someone] in the group has picked up the example form we have and has been decorating it with ARIA properties and working with the grid so you can navigate it in an accessible way. We have feelers going out to multimodal, ubiweb (Dave Raggett), webapps, and eventually xhtml2.
Steven Pemberton: Ubiweb is the old device independence group.

Charlie Wiecha: Also, on a related topic, Thomas Ling in my team is doing the first cut of our simplified syntax. He has some initial postings coming up. He re-wrote the simple loan form. He has questions about labels which he will cross-post to the WG list and the Ubiquity project.

Steven Pemberton: Google has grabbed the word "ubiquity."
Charlie Wiecha: Actually, Mozilla. But there are quite a few other uses of ubiquity.
Steven Pemberton: It's a branding issue. It's not quite different enough and may confuse people.

Steven Pemberton: Also, I've been asked to co-chair the technical plenary day committee. That's both long presentations and lightning presentations, and demos. The theme of the day will be "getting technologies to go together."

Charlie Wiecha: I'll chair until John gets his phone fixed.

* Test Suite License


Charlie Wiecha: Does it affect us?
Keith Wells: At first glance, I think they just want us to add the copyright notice to the test cases.
Nick van den Bleeken: it's a dual license BSD and W3C test suite license
Charlie Wiecha: Can you report back if there is any concern?

Action 2008-08-27.1: Keith Wells to investigate http://lists.w3.org/Archives/Public/public-forms/2008Aug/0070.html licenses and report back.

Nick van den Bleeken: ...
Charlie Wiecha: I want to ask if BSD is appropriate terms or if we need Apache. We just need to take a crack at it.

* Draft Instance Data Module

http://lists.w3.org/Archives/Public/public-forms/2008Aug/att-0055/index-all.html http://lists.w3.org/Archives/Public/public-forms/2008Jul/0001.html http://lists.w3.org/Archives/Public/public-forms/2008Jul/0008.html

Charlie Wiecha: The main thing we've done is to add some IDL to make it useful outside XForms. The lifecycle is implicit in XForms. That's not the case in a data module outside XForms. We don't say anything about extensions to it because we decided that other modules can add anyway. We did add the common attributes.
Charlie Wiecha: We haven't changed the name of the 1.2.3 xforms-link-exception to xforms-link-error so it can be caught by a higher-level module. Here it will be a notification event. In 1.2 we have additional events. It will be targeted at the instance element, not the model.
Charlie Wiecha: load and instance ready have been added. revalidate triggers the default action and can be caught. In the IDL in section 1.3 you see the method for loading instance data. It does the load and then terminates with instance-ready. It's patterned after refresh and revalidate.
Charlie Wiecha: What's the sense of whether we want to support this IDL load? And also there's getInstanceDocument, unlike the XForms one today where you required the ID. We did say last time there was some way to get at the runtime document. load opens up the possibility for the application to reload asynchronously.
Leigh Klotz: So load operates asynchronously and doesn't operate.
Charlie Wiecha: Yes.
Leigh Klotz: How does it get the data?
Charlie Wiecha: Through the resource-uri.
Leigh Klotz: When we hook this up to submission with the next module, how does the submission module write to the instance?
Charlie Wiecha: We need a set in addition to the getInstance document, from serialized form or DOM.
Leigh Klotz: So load doesn't handle all the lifecycle.
Nick van den Bleeken: And reset.
Charlie Wiecha: For inline data we will need a reset.
Leigh Klotz: For inline data, we don't need reset if we have setInstance. It doesn't handle the reset-from-src-or-submission as Nick says.
Paul Butcher: You could have a resettable attribute.
Leigh Klotz: I don't see the value, just use setInstanceDocument. I think it's a rarely used cost.
Nick van den Bleeken: Can we have another module examine the host document and see if it contains a reset?
Leigh Klotz: Examining the host document from an XForms module makes me queasy; what if the reset is called from JavaScript.
Charlie Wiecha: OK, looks like we have consensus on setinstnace and context-ready event. It's not clear how we need to do reset as a higher-level concept yet.
Leigh Klotz: It's clear we need to be able to do reset, but we don't know how yet.
Charlie Wiecha: Then I'll work on his this interacts with submission.
Leigh Klotz: It might also be good to take a look to make sure that namespace inheritance from inline and from setInstanceDocuement works properly when you describe the interactions with other modules.
Charlie Wiecha: OK.

* Review of Actions module

http://www.w3.org/MarkUp/Forms/specs/XForms1.2/modules/instance/actions/index-all.html http://lists.w3.org/Archives/Public/public-forms/2008Aug/0029.html

Nick van den Bleeken: When John was on vacation we decided to base the actions module on XML Events 2, and then we decided that everything it needed was in XML Events 2, but John thought we couldn't go to a major new version of XML Events in a minor release of XForms. So what the action module is is XML Events 1.0 plus some things equal to XML Events 2. I don't see why it's a problem to go to XML Events 2 then.
John Boyer: [back] The if and while attributes would have to be namespace-qualified now. Such as does it allow the context node setting the way we do?
Nick van den Bleeken: Yes, it says the host language.
John Boyer: And what about the namespace?
Nick van den Bleeken: I don't know.
Charlie Wiecha: This seems to be a bridging module not useful afterwards.
Nick van den Bleeken: Does XML Events 2 support chameleon namespacing?
John Boyer: That's the question.
Steven Pemberton: I don't know either.
Nick van den Bleeken: "This draft alludes to supporting chameleon use of event elements and attributes. It is not clear that this is easily supportable by implementors. It may be that these elements and attributes must be in their own namespace" from http://www.w3.org/TR/xml-events2/
Nick van den Bleeken: "When XML Events are included in a host language, all of the facilities required in this specification MUST be included in the host language. In addition, the mandatory elements and attributes defined in this specification MUST be included in the content model of the host language."
Charlie Wiecha: Is there going to be a 2008 update?

Steven Pemberton: http://www.w3.org/MarkUp/Drafts/#xml-events2
John Boyer: We should put the "Editor's Draft" in our 1.1 and module specs.
John Boyer: It still has the confusing picture showing the capture phase going down to the target.
Charlie Wiecha: [irc] the diagram in the underlying DOM API is very nice; hash dashed lines down to the target
Steven Pemberton: I'll send myself a message.
John Boyer: It doesn't say Chameleon.
Nick van den Bleeken: http://www.w3.org/MarkUp/2008/ED-xml-events-20080625/#hostconf
John Boyer: What if there are multiple host documents?
Leigh Klotz: Attributes aren't usually namespace-qualified.
John Boyer: The host element decides, which seems reasonable. Is it OK to have xforms content in an xhtml document and both having unqualified attributes? To me that seems OK.
Leigh Klotz: If the xforms processor and the xhtml processor can't cooperate on DOM events they can't cooperate on XML events.
John Boyer: My processor isn't XML, but I use XML events in the XForms section.
Leigh Klotz: Then you've implemented two things: XML events and DOM Events.
John Boyer: What about generated content?
Leigh Klotz: We initially through repeat would be a problem for mobile devices, but it's not. Then we allowed output/@mediatype=text/html. So the host processor and the xforms processor need to communicate on DOM modification.
John Boyer: yes.
Leigh Klotz: Is your concern that an IE plugin might not be able to process unnamespaced attributes on host markup?
John Boyer: Yes.

Nick van den Bleeken: ...
John Boyer: ...
Nick van den Bleeken: We can say that the adds the attributes in the default namespace for all actions.
John Boyer: Technically that allows an author to write an xf:action with event= instead of ev:event=.
Leigh Klotz: What if you have both event and ev:event?
John Boyer: I think it sounds like you're setting up two event listeners.
Leigh Klotz: So you could use event and if on XForms elements but ev:event and ev:if on XHTML elements unless the XHTML processor vendor told you otherwise.
John Boyer: Is everybody happy with that?
Leigh Klotz: With XML Events 2 and importing?
John Boyer: With XML Events 2; importing event and if; ev:event and ev:if work as well; if you use both you get two handlers; some examples.
Paul Butcher: What about xinclude?
John Boyer: That's why we allow both. You can use ev:* and * on an xforms element. The same thing will happen to XHTML2 on the events level as it imports XForms without namespaces.
Nick van den Bleeken: I'll have a closer look at event names in dispatch and list the small differences and see if we're willing to accept those differences. Or we can ask Mark and the XML Events authors.

Action 2008-08-27.2: Nick van den Bleeken to prepare detailed analysis of effects of moving to XML Events 2 for XForms 1.2.

* Action Items

* News item from Steven

Steven Pemberton: I posted it this morning.

* XSD 1.1 review

Kenneth Sklander: Not yet done. 70% through.
Steven Pemberton: Kenneth, I sent a comment about a particular feature in XSD 1.1.
Kenneth Sklander: I see it, yes.
Steven Pemberton: [drops]
John Boyer: Should we review it next week before you send it out?
Kenneth Sklander: Sure.
John Boyer: So please send it out.

Action 2008-08-27.3: Kenneth Sklander to send to list comments on XSD 1.1 for review for next week teleconference.

* HCG and modularization

John Boyer: I explained the modularization work as being for smaller sized upgrades without rip-and-replace, and also for streamlined syntax (attribute decoration and UI element nesting) for expressing an MVC architecture. Under the covers, we take that attribute decoration and map it onto an MVC architecture in order to provide a way of incrementally adopting the other modules we have available as they find need for advanced features, so there's a nice Onramp. The attribute decoration methodology is increasingly well-known to AJAX library users. It seemed to us that this strategy is a good strategy for allowing us to develop and deploy web-centric W3C technologies and make them into recommendations without direct involvement from browser makers. Growth and popularizations of those technologies will then lead them to natural adoption by browser vendors.
Charlie Wiecha: To HTML or WebApps?
John Boyer: Both. I've started making the requests to visit with XML Security and other working groups. I explained I'd like to show them some of the work I've been doing to create an integration between XForms and XML Signatures, and some of the issues that have come up in applying XML Signatures to an interactive document format, which is different from applying a signature to a static, referenceable source. I want to make sure they don't de-emphasize the types of things we will need, such as document-centric XForms in ODF. ... I do expect that they'll want to meet with us as a result of that email.
Charlie Wiecha: Can we open a Wiki page to track this calendar? Monday at 4 for XG to meet with MMI, so we want to avoid that as a conflict.
John Boyer: And the F2F agenda. We also have Wed and Thu virtual days.

* Sentence about XHTML Modularization?

Steven Pemberton: I can't tell from the action what the wording was.
John Boyer: In the bind, data, and functions modules we are trying to do something like XHTML modularization. None of our modules say that though. Does XHTML modules cover extending an XPath function library?
Steven Pemberton: OK.
Steven Pemberton: In modules where we are just adding attributes or extending an element or attribute group, it's a little clear what we're doing.
Steven Pemberton: I think I understand what needs to be said and will send it off tomorrow.

* Action item list updates for test suite actions done?

Keith Wells: I cleaned up some of my actions recently.
Charlie Wiecha: [leaves]
Keith Wells: I'm still working on FF2 and FF3 tests.
John Boyer: You said there were changes not yet in the test suite?
Keith Wells: Only the header ones that are still in discussion.

* XForms 1.1 Basic

Leigh Klotz: I received the report from Kenneth but have been working on Submission header.

* Submission Header

Leigh Klotz: I posted a comment on Philip Fennel's weblog on the O'Reilly XML site and www-forms and haven't gotten a response. I think we'll need combine="append|prepend|replace|error".
John Boyer: Let's pick it back up in email.

* Submission

John Boyer: Uli Do you think you might be able to start the submission module?
Uli Lissé: Not for the next couple of weeks.

* IRC Minutes


* Meeting Ends