W3C Forms teleconference November 14, 2007

* Present

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

* Agenda


* Previous minutes

http://lists.w3.org/Archives/Public/public-forms/2007Oct/0152.html IRC supplement: http://www.w3.org/2007/10/31-forms-minutes.html

* Forms Joint Task Force

Steven Pemberton: Did anything happen?
John Boyer: Any news this week?
Mark Birbeck: Did you discuss it at the F2F? We had a call two weeks ago.
Charlie Wiecha: I heard The focus was going to be on general directions, not on formal spec writing.
Mark Birbeck: We talked about what consistent architecture meant; we wanted to make sure an MVC approach wasn't ruled out. The markup itself isn't as much an issue, we had previously agreed. So that's a reasonable success. So we want to recast it using syntax close to HTML5, to give a natural progression from HTML4, to HTML5 extensions, through to XForms, which I think is achievable.
Charlie Wiecha: Do you think scope creep with explicit MVC architectures in HTML5 might be a problem?
Mark Birbeck: I hadn't thought about that. But anything can happen.
Charlie Wiecha: ...
Mark Birbeck: As far as I'm concerned that's the big issue, not the syntax.
Charlie Wiecha: So when you cross from a monolithic design to MVC that's what triggers XForms?
Mark Birbeck: As long as we can create a notional model based on, say, HTML forms...you could easily recast HTML forms as an XForms model; there's no reason why our phases of serialization and collecting data and submitting data, why you can re-write HTML forms with its limited attributes in this architecture. SO imagine you start with HTML and a form element and the limitations, then write it again as if it were XForms. HTML says visit a control, ask if the control is disabled, and pluck out its value. So we can re-write this as a model behind the scene that gives the same results. So if there is a disabled attribute it doesn't matter. We've obviously discussed UI-based attributes as well. So essentially it's an MVC model.
Charlie Wiecha: My concern wouldn't happen for some time anyway. We can get 1.2 WD out in the meantime.
Mark Birbeck: I've just come of the XHTML2 call, and the backplane (progress there) would have the same model.
John Boyer: We'll talk more about 1.2 in upcoming calls.

* Completion of action for issue 106, input mode


John Boyer: I sent a message to Martin Düerst and said that we'd re-labeled this appendix as informative; several people already thought it was informative. Does anyone have an issue with the inputmode material being marked informative. No? I think now that it's informative we can move forward.
Steven Pemberton: Without the new text that he's going to provide.
John Boyer: When he does provide it we can add it.
Steven Pemberton: Yes, after CR. It's an implementation comment: there are other scripts, but it doesn't change the meaning of inputmode. I wish we had agreement for an algorithm to extract the values. That would be the ideal. It only changes the UI in a tiny way, not XForms. We should allow anybody to add values and just say where they come from.
John Boyer: Should we re-stress this point? It might make his job easier.

Action 2007-11-14.1: Steven Pemberton to contact Martin Düerst and reiterate advantages to algorithmic determination of inputmode script.

* Disposition of comments


John Boyer: Here are our results.
John Boyer: Steven, do you have any comments on Issue 69?
Steven Pemberton: Not yet, but I may flip the bit at the end.
John Boyer: I also recorded disagreement with one of Michael Kay's issues and I recorded that. One issue is the use of the random function; he doesn't like functions that have side-effects beyond the DOM. I explained the XForms mental model that XPaths are run in a batch anyway; unfortunately, even while that batch is running, events such as setvalue/insert/delete can mutate the DOM, so a lot of the issues around side effects are not so severe, as we already have this problem.
John Boyer: The other is 147, seconds-from-dateTime; apparently it has a different semantic in XPath 2.0 (same name). We resolved to keep the same function because we've long had it; when we go to XPath 2.0, we will address whatever subtle semantic differences there are. I presented the basic use cases and ask if any of those would be affected by the differences, now that we made the change to say we don't support leap seconds. It's a long-standing function we can't throw away at this point.
John Boyer: The fourth and final disagree we have is from David Landwehr; he doesn't agree that you can set readonly on a calculated node to false. He did indicate he was not raising a formal objection.
Steven Pemberton: So that's a live with?
John Boyer: In the set of possible positions, there's a difference between disagree and formal objection. We have no formal objections.
Steven Pemberton: He's saying, "I hear you, go ahead?"
John Boyer: He doesn't agree with it but he can live with it.
Nick van: [irc] "I accept the resolution because I give up."
Steven Pemberton: It is regrettable that one of our implementers disagrees with the position.
John Boyer: So we should report it as disagreement. It will come up on the call and we'll have to explain it, that there were numerous other implementers who felt he other way.
John Boyer: There are 16 of 17 deferred comments missing, two accept, and one modify and accept.
John Boyer: Ah, Shane says there are 125 issues marked as implemented. They are missing because they are not in the state=implemented (deferred).. I could flip their states to implemented.
Steven Pemberton: Let's do this offline with Shane.
John Boyer: We now have a last-call comments implemented document;
John Boyer: Here is the link to the specification we would like to Transition to CR.
John Boyer: [irc] XForms 1.1 editor's draft of CR-ready specification: http://www.w3.org/MarkUp/Forms/specs/XForms1.1/index-all-20071106.html
Steven Pemberton: http://htmlwg.mn.aptest.com/xforms11-lc-doc-20071114.html
John Boyer: I got a few errors from Keith Wells in the examples; they are not substantive.
Leigh Klotz: I sent a list of typos and sentence problems during the F2F.
John Boyer: Please send it to the list.
Nick van: [irc] we have a lot of No Response issues
Steven Pemberton: [irc] http://www.w3.org/MarkUp/Forms/2007/xforms11-lc-doc-20071114.html is now further up-to-date John, issue 45 is marked "editorial" as is 69
John Boyer: Any objections to transition to CR? Are these IRC comments objections?
Steven Pemberton: No.
Sebastian Schnitzenbaumer: No objections.
John Boyer: Let us record a resolution.

Resolution 2007-11-14.1: We request transition CR of XForms 1.1 with http://www.w3.org/MarkUp/Forms/2007/xforms11-lc-doc-20071114.html

* Responses

Nick van: [irc] Are we going to clean up the no-response states?
John Boyer: A no-response is a green light, but WG members can clean up their responses. Anybody want to peruse their issues and make sure they are all at an appropriate state other than no response.
Steven Pemberton: [irc] I will.
John Boyer: Erik has. I have. Nick?
Nick van: I don't have issues. I checked the Michael Kay ones.
John Boyer: The ones he didn't respond to we assume it didn't bother him. The table looks normal to me.
Steven Pemberton: It's getting that way. There might be a remark on the number of no responses, but it's the red things they look for.

* Implementation Report

John Boyer: The other major thing we need for transition here is the expected implementations. Chris Lilley said that our status section said there were no preliminary implementation reports, and suggested we put up information on our Wiki with the implementers we list and a paragraph listing the things already implemented from XForms 1.1. It doesn't have to be a test suite. The web services with SOAP, the modifications for insert and delete, etc. That would be more indicative of why W3C should advance the recommendation.
Steven Pemberton: Would that be Mark?
John Boyer: Any implementers with a brief paragraph.
Sebastian Schnitzenbaumer: We are trying it; it's the top of my agenda. I cannot promise.
Leigh Klotz: You just need to have a paragraph saying what's done now, not a full implementation.
Sebastian Schnitzenbaumer: Great!
Nick van: [irc] I've done all the new XPath Functions
Sebastian Schnitzenbaumer: Yes, Nick can do that.
Keith Wells: [irc] I can get a list of Mozilla xforms 1.1 supported features (or ones in development)
John Boyer: Can we get this tomorrow? That would help me put the paragraph together and put it in the status section of the document?
Mark Birbeck: Kind of hard to do that by tomorrow. What kind of issues? if and while?
John Boyer: Yes, no details.
Mark Birbeck: I can do that.
John Boyer: We'll have a strong presentation.
Nick van: Is just a list of features good?
John Boyer: Yes.
Nick van: I can do it this evening.
Rafael Benito: [irc] you can count on some features in DataMovil 3.0

* Duration of Call Period

John Boyer: Any concerns about the transition period? No.

* Next F2F

Steven Pemberton: The date we have pencilled in Amsterdam is not possible. I need to move it a little later in May.
John Boyer: When?
Steven Pemberton: It depends on the week; we've cut loose from XHTML2. May 12th is the last day that's a holiday and that's a Monday. So May 14-16 at the earliest, possibly the 13th. We're currently pencilled in for May 5-7. It would have to be at least a week later. Otherwise I don't mind, though we have to watch out for the Web conference.
John Boyer: I think it's very early this year.
Charlie Wiecha: I think that works better. Tue-Thu.
Steven Pemberton: That would be the 13-15 May.
Mark Birbeck: It's better to have a Saturday; starting Monday is better.
John Boyer: So May 14-16. Return on Saturday.
Mark Birbeck: That doesn't get you an overnight Saturday.
John Boyer: So it would be better May 19.
Mark Birbeck: For me, a European flight, doesn't make a difference.
Charlie Wiecha: I hate burning the weekend for travel.
John Boyer: Then May 13-15 would be better.
Charlie Wiecha: I thought Saturday was less important now.
Mark Birbeck: From here to Canada it was.
John Boyer: We'll pick it up again next week. Let us know if May 13-15 is a problem.

* Participation

John Boyer: How do we get more people to come? Why did XHTML2 go to three meetings?
Steven Pemberton: To make it cheaper.
John Boyer: More telecons?
Steven Pemberton: No.
John Boyer: Would more telecons substitute for a F2F?
Steven Pemberton: SVG has an day-long IRC to discuss things and do action items. There's a timezone issue but there's also travel. I'd rather sit up all night at home than around the world. If other people would commit, then video plus teleconferences or IRC, late in the night, could work.
Mark Birbeck: That implies that there's still a physical F2F. If we were to do that on the bridge, that quality of conversation is better than what we have for remote conferences. I just work from home then.
Sebastian Schnitzenbaumer: I'd rather meet somewhere other than the US because it takes two hours through immigration; it's not a money issue.
Mark Birbeck: Or two F2F, one in Europe and one in America, at the same time, video linked.
Steven Pemberton: That's an interesting approach. Then the video link is easier. The video link with you on the screen worked, Mark.
Erik Bruchez: Two meetings is hard in different time zones, Europe half the night.

* End of scribe minutes

Leigh Klotz: I have to leave.

* IRC Supplement


* Meeting Ends