W3C Forms teleconference October 8, 2008

* Present

Charlie Wiecha, IBM
John Boyer, IBM (chair)
Leigh Klotz, Xerox (minutes)
Roger Pérez, SATEC
Steven Pemberton, CWI/W3C
Uli Lissé, DreamLabs
Keith Wells, IBM

* Agenda


* Previous minutes

* Next FtF (October 15-16 Virtual, Oct. 20-21 Cannes)

* The Forms Newsletter/News items

Action 2008-10-8.2: John Boyer to publish implementation report from received from Kenneth Sklander in 2008 space.

Leigh Klotz: From Get Addons it doesn't show compatibility.
Keith Wells: Try https://addons.mozilla.org/en-US/firefox/addon/824
Leigh Klotz: That works. It didn't show up as compatible though. You or I can report that.
Keith Wells: https://addons.mozilla.org/en-US/firefox/addons/versions/824 has all versions.

* XForms 1.1 Implementation report progress?

John Boyer: We need implementation reports now. Once I get the submission headers in, we'll need to get the reports and analyze them for features. We may have to yank unimplemented features and have a short last call. We need more implementations.
Keith Wells: Can we do our implementation report on Ubiquity to satisfy that?
John Boyer: We did that for Firefox.
Keith Wells: So we can implement the missing features in Ubiquity.
John Boyer: Yes, we can see if any implementors are willing to implementat the features and amend their reports.
Steven Pemberton: http://www.mozilla.org/projects/xforms/ says XForms 1.1 is a Recommendation and should be fixed. It also doesn't say that 0.86 is released.
Leigh Klotz: I'd say to link to https://addons.mozilla.org/en-US/firefox/addons/versions/824 directly.
Steven Pemberton: Is that the latest version?
Leigh Klotz: That's the unique id for the XForms extension.
Steven Pemberton: And the project page, who owns that?
John Boyer: Aaron Reed
Steven Pemberton: I'll ask him about the text on the page.

John Boyer: So should we contact Alfresco about their use of XForms? It is news, but not as big news as the XForms 1.1 implementation for Firefox.
Steven Pemberton: I will do that.
John Boyer: We can combine the adoption news with the adoption news from Nick.

* Rich Web Application Backplane

Charlie Wiecha: I am preparing the Tech Plenary talk.
John Boyer: I owe you slides.

* Submission Module

John Boyer: Uli, have you started yet?
Uli Lissé: Not yet, sorry.
John Boyer: Take a recent module as a basis and take chapter 11 as a basis and put that in as a file and amend index.xml to include that file and that will get you a first basic module to work on. From there, it's a matter of going through that document and deciding. There's not a ton of work there; the main thing is that you have to factor out relevance pruning and data validation from that module. The submission module will assume the existence of an instance. The validation and relevance modules will add function to the submission module.
Uli Lissé: We talked about this last week. I'm going to try to coin aditional events for validation and pruning.
John Boyer: My comments were about the document process, more low level. I see what you're saying.

* Why are Events attributes on the model element?


John Boyer: This is a question from Erik.
Steven Pemberton: It's because some events get sent to the model, right?
John Boyer: Typically you'd create a child action of the model.
Steven Pemberton: Yes, but you don't have to do it that way.
John Boyer: So you wouldn't make model a handler.
Steven Pemberton: Yes, you would put ev:handler.
John Boyer: You wouldn't put just ev:target. You'd also put ev:handler and probably ev:observer as well, since you it would be a rarity that you would want to listen to the parent of model. Is it a reasonable thing to do?
Leigh Klotz: It's somone else's spec and it's well defined, so why not?
John Boyer: We don't allow them on instance.
Leigh Klotz: Maybe that's not our business.
Steven Pemberton: Why did we not allow them?
Leigh Klotz: I remember why it was; it was to prevent confusion about events targeted to instance elements in the that other DOM which you can't do, but in practice I'm not sure people are confused.
John Boyer: Actually, we allow ev:* anywhere on common because common allows foreign attributes.
Leigh Klotz: We don't define the behavior of XML events, right?
John Boyer: Right.
Leigh Klotz: So why are we putting them in our spec at all? We should just allow them anywhere and call out particular useful examples.
John Boyer: If we get rid of Erik's specific thing that is causing him trouble, not mentioning the events attributes on the model.
Leigh Klotz: You can but we should point out that the common attributes allow ev:*, and that XForms doesn't define XML Events behavior, so the places where ev:* is explicitly listed in the spec are for convenience only.
John Boyer: I can do that. I want to leave the events in the Schema because it calls attention to particularly good places to put them, but they are allowed everywhere common is allowed.

Action 2008-10-8.3: John Boyer to get rid of the language in the model element about ev:* attributes as the model element says common and events and should just say common, and add a note in common 3.2.1 that the common attributes allow ev:*, and that XForms doesn't define XML Events behavior, so the places where ev:* is explicitly listed in the spec are for reading convenience only.

* Common event info not exposed in event() function

http://lists.w3.org/Archives/Public/public-forms/2008Aug/0048.html Discussed Aug. 20, 2008. Deferred to 1.2 or later. But can't we get this in XML Events 2? Upgrade to http://www.w3.org/MarkUp/2008/ED-xml-events-20080625/

John Boyer: Useful things to know: currently in cancel or bubble phase, do you capture or bubble, etc. Last night I was going through the agenda items for the F2F and realized we've talked about going to XML Events 2. It doesn't say anything about the properties, as it's a WD. Isn't that the right place to address this problem? Steven?
Steven Pemberton: That sounds like a last call comment.
John Boyer: So it's in last call review. I guess I'm late.
Steven Pemberton: We haven't gone to CR yet. Get as many generic things out of XForms into XML Events 2.

Action 2008-10-8.4: John Boyer to comment to XML Events 2 that bubbleable, cancelable, current phase and event target are valuable metadata for event context.

* Modularization


John Boyer: How do we re-invigorate these modules? Can we write them together?
Steven Pemberton: That's not necessarily unproductive.
John Boyer: We will have observers at the meeting, so discussion material would be preferable.

Charlie Wiecha: I'm going to need more of these.
John Boyer: Evaluation context needs to be part of the instance module because it affects the in-scope evaluation context. Also, I'm passing in the nodeset expression, which is incorrect. It should have a nodeset binding.
Charlie Wiecha: I see where you're going but I am not sure I have all of it.
John Boyer: We can spend some time next week on the call.
Charlie Wiecha: Maybe we could revisit and revalidate our granularity of modules and dependencies. I have a fairly fuzzy picture of how these modules relate to each other and drawing a diagram would help.
John Boyer: An architecture diagram.
Charlie Wiecha: Something to capture the relationships. We have a good idea but we're going slower than we need to. Having a few larger modules in the first cut migth help.
John Boyer: We have two at instance data.
Charlie Wiecha: Then actions and binding.
John Boyer: I agree it's been a useful exercise.

John Boyer: For actions, the claim is that a lot of it disappears with XML Events 2. It has a container for a set of actions? Deferred update maybe goes to the model module. So we could start with a model module.
Charlie Wiecha: We could collapse the instance, instance++, actions, binding attributes. Is that too much to do in the first pass?
John Boyer: Having relevant, readonly, and calculate as separate modules makes it easy to refer to them in streamlined syntax.
Charlie Wiecha: OK.
John Boyer: Can I just plug some modules into hours and decide what we're talking about?
Charlie Wiecha: Seems reasonable.

John Boyer: And next?
Charlie Wiecha: Do we want to talk about Future Features other than modularization?
John Boyer: We have to start to recharter after this F2F or the next, for the end of 2009.
John Boyer: Without more spec editors, we're going to have a devil of a time. I can't do all this work. Maybe the last calls for CR will be shorter documents. A first WD of 2.0 would be astounding. We need these to get rechartered. I didn't understand why our charter went until 2009 and HTML5 went to 2010. But that's where we are. The three-fold strategy of Ubiquity/AJAX, modularity, and streamlined syntax will move us forward.

John Boyer: Where is our next F2F?
Charlie Wiecha: In the UK; I will provide space and Mark will host. I can't book the space until three months before. Hursley is easier, but downtown is what Mark preferred.

John Boyer: Steven, how long does rechartering take?
Steven Pemberton: It often takes too long.
John Boyer: It would be helpful to have the 2.0 documents out then, first public working drafts.

Keith Wells: What are the hours of the virtual day?
John Boyer: 9AM to 4PM eastern. We're also doing extra teleconference time, 30 minuets.
Steven Pemberton: Are we having a call on Wednesday?
John Boyer: That's the virtual day.
Steven Pemberton: Wednesday and Thursday.
John Boyer: See you then.

* IRC Minutes


* Meeting Ends