W3C Forms teleconference October 24, 2007

* Present

Charlie Wiecha, IBM
John Boyer, IBM (chair)
Leigh Klotz, Xerox (minutes)
Lars Opperman, Sun
Mark Birbeck, x-port.net
Nick van den Bleeken, Inventive Designers
Sebastian Schnitzenbaumer, Dreamlabs
Steven Pemberton, W3C/CWI
Uli Lissé, Dreamlabs
Keith Wells, IBM
Roger Pérez, SATEC
Blake Jones, DAISY/ViewPlus
Erik Bruchez, Orbeon

* Agenda


* Previous Minutes

http://lists.w3.org/Archives/Public/public-forms/2007Oct/0068.html IRC supplement: http://www.w3.org/2007/10/17-forms-minutes.html

* XForms Conference

John Boyer: We might want to record the event. Anyone do that regularly? It might be a good idea to figure out some way to get it recorded.

* Backplane

Charlie Wiecha: We now have a third option for patent policy for backplane, for cross-group collaboration.
John Boyer: Can you send a short message to the group pointing out what the third option is, why it's important, and a link. I'll put it on the F2F list.
Charlie Wiecha: It allows us to adopt the same patent policy for incubator groups that we do for working groups, a concern of Voice WG.
John Boyer: That's great news. Is that announced to W3C?
Charlie Wiecha: It's on a member page, but I'll describe it.

* Call Time

Steven Pemberton: This call will, for one week, be an hour different for Europeans. I will send mail.
Nick van: [irc] one hour later I guess
Steven Pemberton: Then the week after that it will be back to normal because the US changes, but hat is the F2F of course.

* Backplane


John Boyer: Jim Larsen and Debbie Dahl said that Voice and Multimodal were meeting on Tuesday, and MMI was also interested in talking about harmonizing events, so they raised the question about whether all three groups should meet on Tuesday.
Charlie Wiecha: I think there is shorter term progress with Voice on data models, possibly SCXML. Eventing is more problematic, bubble, capture, etc. The alternative is two separate meetings which probably don't have time for.
John Boyer: If we don't do it now, it will be another year.
Charlie Wiecha: Maybe we could take voice issues first, then multimodal issues.
John Boyer: I don't mind just keeping Monday with Voice and then Tuesday with both groups on event harmonizing. Steven?
Steven Pemberton: We could do events the second day.
John Boyer: Everybody happy?
Charlie Wiecha: I can coordinate the voice one if Steven does the events.

* Completion of XForms 1.1 schema updates


Charlie Wiecha: I'm done except for a submission element issue which I sent you.
John Boyer: The Schema will be checked in today, if you want to look over it.
Nick van: [irc] I've read it and it looked good I think

* Completion of actions related to select/select1 (88,89,90)


John Boyer: You completed these?
Erik Bruchez: Yes. I can go through the email. The sections for select and select1, I noticed they don't tackle xforms:copy. It is described separately, and mostly is about xforms:value. The changes I made retain that direction. If we agree, it wouldn't be too hard. There was a message similar to message 89 about this.
Erik Bruchez: Also, a change for xf:select1 which wasn't in the message. I harmonized the wording of select1 with select: I took a more detailed approach in select, when the control gets a value, when the user selects, and deselects and propose doing the same in select1.
John Boyer: The harmonizing certainly sounds reasonable. On the first issue you raised, should value and copy behave in an analogous way? You're matching strings in value; with copy you're matching subtrees.
Erik Bruchez: In fact, the spec doesn't talk about this at all. You could conceivably either match a deep copy, or just match on the qualified name, or even on the prefix. We don't specify this at all.
John Boyer: We don't say how to match them exactly, but I think the word match is fairly clear. If you only compare the root elements, you could have a schema in which the root element has a choice so it's not a subtree match. When copy sets the value it's a deep copy; it's reasonable to conclude a deep comparison.
Erik Bruchez: Yes. I think that's fairly reasonable but we don't say it right now. The "algorithm" is described in xf:copy but it's complete. The "algorithm" for xf:value is described in xf:select. It's inconsistent. One thing to do would be to harmonize.
John Boyer: That makes the most amount of sense.
Erik Bruchez: ..
John Boyer: I've been taking the approach that if there's something in the neighborhood of a last-call comment we should fix it. We have a comment about value but not about copy, but we should fix it.
Erik Bruchez: If the WG agrees I'll fix copy as well. But I want to be reasonable sure that it's the right direction first. I heard from Nick but nobody else.
John Boyer: OK. I'll go over that today as well. Anyone else?
Steven Pemberton: Yes, eventually. We're reviewing it at the F2F?
John Boyer: Good question. Are we then?
Steven Pemberton: What are the options?
John Boyer: If we've reached resolutions, I hadn't thought we'd spend more F2F time on CR issues.
Steven Pemberton: I have to warn you that I saw the word substantive change in a message header and that rang some ugly bells for me. That means a second last call, I'm afraid to say. If we ask for a transition to CR and we report substantive changes since last call, then we'll have to go to last call again.
John Boyer: We can talk about that at the F2F.
Steven Pemberton: If the changes aren't substantive, then it's OK. It would probably be only short, three weeks to check just certain places.
John Boyer: This particular issue doesn't call for a substantive change.
Steven Pemberton: No, but it's about whether we need to discuss CR issues. Just look for "substantive change" in the header.
John Boyer: I see; the question is what does "substantive change" mean. I was writing some examples, but as I was writing them, I noticed that the mediatype attribute suggested that form authors should make sure the mediatypes they assign are consistent with application/xml; this isn't true in XForms 1.1 because we have a mechanism to submit plain text. It's only a substantive change in that it was a change in not in example material.
Steven Pemberton: Then that's probably safe.
John Boyer: If you look at all the changes for last-call comments, we've changed the verb element to method element. Is that a substantive change or just addressing someones last-call issue? It's clearly the same concept, still implementable. Otherwise it would be impossible to avoid a second last call for anyone. Should we look through and see if there is a need for a second last-call?
Steven Pemberton: Yes.
John Boyer: How long did this take for XForms 1.0?
Leigh Klotz: We had more than one last call.
Steven Pemberton: If someone says there is a substantive change we have to discuss it. It might be good to have that in the minutes of the meeting that we've done the evaluation, for the transition call.
John Boyer: So a couple of hours at the F2F?
Steven Pemberton: No, a half hour or so.
Sebastian Schnitzenbaumer: Just look at the final version and see if there are big implementation changes.

John Boyer: I'll put the last-call issue analysis for substantive changes on F2F agenda.
Leigh Klotz: Should John put the select material from Erik in the spec now?
John Boyer: Steven wants to review it; I'll review it adding it in. Steven, are you OK with putting that text in?
Steven Pemberton: Sure.

Action 2007-10-24.1: John Boyer to include material from http://lists.w3.org/Archives/Public/public-forms/2007Oct/0084.html in XForms 1.1.

John Boyer: And Erik, you'll look at the xf:copy harmonization?

Action 2007-10-24.2: Erik Bruchez to provide spec-ready text for xf:copy harmonization as an extension of http://lists.w3.org/Archives/Public/public-forms/2007Oct/0084.html

* Completion of action for issue 106, input mode


John Boyer: Any comments from Martin Düerst?
Steven Pemberton: We asked him if it would be more efficient to have the list by reference. He said it made some sense, but said that it wasn't clear that future versions would be the same. Because there are changes in spelling, it would be more useful to have the whole list.
John Boyer: [irc] http://htmlwg.mn.aptest.com/cgi-bin/xforms-issues/Appendices?id=106
Steven Pemberton: So his official answer is no.
John Boyer: So we have to update the list?
Steven Pemberton: Does anybody understand the list item?
Leigh Klotz: Martin gave us this list the first time as he was on the WG. Maybe we can ask him to do it the first time.
Steven Pemberton: Then just give me the action.

Action 2007-10-24.3: Steven Pemberton to ask Martin Düerst to provide updated list of scripts for http://htmlwg.mn.aptest.com/cgi-bin/xforms-issues/Appendices?id=106

* Completion of actions for issues 102, 162


John Boyer: Steven, this is the required property, respecting it, showing it, etc.
Steven Pemberton: Let me do that Friday.
John Boyer: We'd agreed to add a description of changes to section 1 of the spec and you'd provided some text; could you find it for me?
Steven Pemberton: Someone took it in Madrid and was going to turn it into XML Spec. Sebastian?
John Boyer: Sebastian wasn't there.
Nick van: I was doing it; I spent quite some time on it but it's hard to convert it to something we can put in the spec. I was trying to add references to relevant sections; now it's just bullets.
Leigh Klotz: Why don't we just publish it as a note then.
Steven Pemberton: Or just list it as bullet points. I think, quite honestly, it would be good to have in the spec.
John Boyer: If we wanted bullet points now and then links later, then that's ok as it's not normative.
Steven Pemberton: You could put in in as an informative appendix.
John Boyer: A larger description, yes. Then later down the road you can make later links.
Nick van: I can send in what I have but it's also not...I tried to use ordered lists and specXML but I'm not sure if I'm using the right ones. Maybe you can fix the exact structures.
John Boyer: I can do that and send it back to you. Do you have an action item on this? I can plug it into XMLSpy for the DTD.
Nick van: Yes, it is valid. I removed formatting and tried to make it look like spec text and that was more work.
John Boyer: Is there a scaled-down version you can do in the next week or so?
Nick van: Yes, I'll send it in.

Action 2007-10-24.4: Nick van den Bleeken to send current version of spec-ready-ish text of changes to list.

* Issues with the use of xforms-binding-exception and

xforms-compute-exception http://htmlwg.mn.aptest.com/cgi-bin/xforms-issues/Events?id=36 http://lists.w3.org/Archives/Public/public-forms/2007Oct/0078.html

John Boyer: I've asked Erik if we can defer this.
Erik Bruchez: It's a multi-dimensional issue. There's the name issue, for example. I think it may take a bit of time to figure out what we want and the implications of changes.
John Boyer: Yes.
Erik Bruchez: It's hard to decide which is which by reading the spec. It's a longer term issue unless we agree right now. The events cannot be stopped, for example. It's not very hard but there's more than we can do now. 1.2 or 2.0.
John Boyer: Then we can classify this as a defer; we acknowledge that we're not too clean on exceptions. They halt processing anyway. We would like a better analysis of when we should not halt processing, and what more information we can give, even for debugging. Everyone OK to defer for now?

Resolution 2007-10-24.1: We defer http://htmlwg.mn.aptest.com/cgi-bin/xforms-issues/Events?id=36

* Detecting more than just syntax errors in XPath


John Boyer: If the XPath expression parser chokes, we generate an error.
Erik Bruchez: They don't use those terms static or dynamic. They do specify variable naming, function mappings...
John Boyer: There is a context. It wouldn't be too hard to say that thinks like variable references are wrong. Or function names not in the function library. The only two other things are the context node position and size and those are under our control.
Erik Bruchez: Those are part of the dynamic context; function library is static context.
John Boyer: ... static context ...
Erik Bruchez: In XSLT you can determine statically what variables are in scope. They may or may not be in the static context in xslt.
John Boyer: That are not statically valid, a variation on Steven's suggestion?
Steven Pemberton: I have my doubts about the word valid; maybe incorrect
Leigh Klotz: Isn't it any error that's independent of the instance data?
Erik Bruchez: Static analysis is when you don't run it.
Leigh Klotz: If it's not statically valid over the domain of all instance data then it's problematic.
Erik Bruchez: As in XPath 2.0, the syntax is correct, the function library has to be correct.
Leigh Klotz: I would put in more stuff.
Erik Bruchez: Just keep statically correct.
John Boyer: The context is pretty well defined: expression string, node, position, variables, function library. So those are the only things that could go wrong.
Leigh Klotz: No matter what the instance data is you shouldn't get static errors.
John Boyer: We say that already all over.
Erik Bruchez: We should do it like XPath 2.0 if we add more text. Instead of saying static context, include in-scope functions, etc. We don't have variables.
John Boyer: We're out of time. We do need a resolution. Erik, can you propose text?
Erik Bruchez: OK, maybe next week sometime.

Action 2007-10-24.5: Erik Bruchez to propose text for http://htmlwg.mn.aptest.com/cgi-bin/xforms-issues/XPath?id=139

* Meeting Ends

* IRC Minutes