W3C Forms teleconference October 3, 2007

* Present

Keith Wells, IBM
Sebastian Schnitzenbaumer, DreamLabs
John Boyer, IBM (chair)
Charlie Wiecha, IBM
Mark Seaborne, PicoForms
Nick van den Bleeken, Inventive Designers
Steven Pemberton, CWI/W3C
Leigh Klotz, Xerox (minutes)
Rafael Benito, SATEC
Blake Jones, ViewPlus Technologies/DAISY

* Agenda

* Previous minutes:

http://lists.w3.org/Archives/Public/public-forms/2007Sep/0094.html IRC supplement: http://www.w3.org/2007/09/26-forms-minutes.html

* Progress on assembling XForms @ XML 2007 announcement?

John Boyer: Steven, can we get the abstract page ready for me to post in the database that Dave Megginson sent us.
Steven Pemberton: Leigh?
Leigh Klotz: Yes, I have them in http://del.icio.us/leighklotz/xforms+paper+submission
John Boyer: I changed my abstract not to say "proposed".
Leigh Klotz: Should we do that?
John Boyer: I think so. Also contact and ask for photos.

Action 2007-10-3.1: Leigh Klotz to edit abstracts and ask authors for photos for XForms @ XML 2007.

Action 2007-10-3.2: Steven Pemberton to post final abstracts for XForms Day to XML 2007 database that Dave Megginson will provide John Boyer.

* Fix for dateTime math


John Boyer: Two fixes, to add optional date-time parameters to deal with calculating "two hours from now" in the same way we calculate "two days from now." http://www.w3.org/MarkUp/Forms/specs/XForms1.1/index-diff.html#fn-seconds-from-dateTime The boolean defaults to true (as in XForms 1.0) using UTC time, but false makes the seconds-from-dateTime calculation in the current time zone.

Nick van: Adding the boolean makes seconds-from-dateTime incompatible with XPath 2.0.
John Boyer: What about seconds-to-dateTime?
Nick van: That's not in XPath 2.0.
John Boyer: That's a good point. The critical one is seconds-to-dateTime. http://www.w3.org/MarkUp/Forms/specs/XForms1.1/index-diff.html#fn-seconds-to-dateTime
Nick van: They have another way to do that.
John Boyer: They have natural date math functionality. I found I couldn't calculate "two hours from now" in a single calculate. If you look at the examples, you can see I don't need the parameter on seconds-from-dateTime.
Steven Pemberton: [irc] Ugh
John Boyer: Yes, Steven, XPath 2.0 will be better. But if you look in the email, you'll see that without this parameter, you have to have four binds and two or three local instance date variables to calculate "two hours from now." At a minimum, if we can keep the optional boolean parameter, days-to-date(days-from-date(local-date()) + 31) will add 31 days.
Leigh Klotz: So the issue is that now returns UTC?
John Boyer: The issue is seconds-to-dateTime, which is not capable of returning a local date. http://www.w3.org/MarkUp/Forms/specs/XForms1.1/index-diff.html#fn-seconds-to-dateTime You will see that seconds-to-dateTime produces a UTC dateTime.
Steven Pemberton: You can say that again.
John Boyer: [repeats]
Steven Pemberton: No, I mean I agree with you entirely.
Nick van: I agree that we need something to convert to local timezones, but boolean is a bit...it is not such expected behavior. It feels a bit of a hack. In fact what you really want is conversions between timezones, as in XPath 2.0. Maybe it's a bit too much to ask. You could have a convert timezone function to localtime.
Leigh Klotz: It seems to me a function that would convert from UTC to localtime would do the same thing but be discoverable.
Nick van: Those are all in XPath 2.0.
John Boyer: I agree that XPath 2.0 would be nicer and I could go for seconds-to-localDateTime. If we do nothing, we have to do all this...
Leigh Klotz: I think a separate function is better. It doesn't have to be totally generic if XPath 2.0's, is; we can just do convert to localtime.
John Boyer: Yes, and in part because modifying seconds-from-dateTime isn't a good idea.
Nick van: [irc] http://www.w3.org/TR/xquery-operators/#func-adjust-dateTime-to-timezone
John Boyer: So we could just have adjust-dateTime-to-localtime.
Nick van: I wanted to do this, but we are trying to add not too much stuff that isn't in XPath 2.0.
John Boyer: We're about to go to a call for implementations and this looks easy to implement.
Nick van: I can live with it.
John Boyer: So the proposal is to add a adjust-dateTime-to-localtime function.
Leigh Klotz: The XPath 2.0 function has lots of cases to consider, empty, no timezone etc.
John Boyer: Yes, we'll have to have similar language.
John Boyer: [irc] Proposal: add a convert-to-local-dateTime; Similar language about edge cases as well as main case of converting from UTC to local (if timezone info known by impl)
Leigh Klotz: I'd say that it should handle all timezones in the date, just as in seconds-from-dateTime.
John Boyer: What if it isn't specified?
Leigh Klotz: It says just append the timezone if the timezone is unspecified.
Nick van: I would say we can just use the XPath function without the optional timezone arg.
Leigh Klotz: It's called adjust-dateTime-to-timezone and they have it defined without the timezone parameter as well.
John Boyer: That would probably make Erik happy. Anybody unhappy with that? So resolved.

Resolution 2007-10-3.1: We adopt adjust-dateTime-to-timezone from http://www.w3.org/TR/xquery-operators/#func-adjust-dateTime-to-timezone in the single-parameter version only.

Action 2007-10-3.3: John Boyer to adopt adjust-dateTime-to-timezone from http://www.w3.org/TR/xquery-operators/#func-adjust-dateTime-to-timezone in the single-parameter version only for XForms 1.1.

* Need topics of discussion for proposed meeting with Voice group

John Boyer: What topics?
Steven Pemberton: Harmonizing forms; they have their own HTML-like forms. Also, harmonizing events, though possibly that is backplane.
Sebastian Schnitzenbaumer: Is this for the all-group meeting?
John Boyer: It came up at the F2F. We decided to request the meeting.
Sebastian Schnitzenbaumer: What's the reason for the request?
Steven Pemberton: I'd like to discuss the HTML approach vs the XML approach.
John Boyer: Charlie?
Steven Pemberton: What was in the minutes.

* TPAC Deadline

Steven Pemberton: Today is the final deadline for the F2F hotel deadline.
John Boyer: We really need you at the F2F meetings.

* XForms Basic

Steven Pemberton: what is the status?
Leigh Klotz: I'm waiting for the test suite to be updated to PicoForms to test.
Keith Wells: The changes should be in place. The v3 test suite was updated two days ago. Can we get the updated link posted John?
John Boyer: Yes.
Steven Pemberton: Where is the link
Keith Wells: In my email this morning.

Action 2007-10-3.4: John Boyer to update v3 test suite link from home page.

John Boyer: Which test suite are they trying to create an implementation report against?
Steven Pemberton: We are using PicoForms for PR for XForms Basic. Which edition are you using? Third edition?
Mark Seaborne: I think we are using second.
John Boyer: Leigh, have you seen the implementation report?
Leigh Klotz: I've not seen it. I was preparing to ask for one.
Steven Pemberton: I've seen a spreadsheet with reports of basic vs. not.
Leigh Klotz: I don't care where it comes from, but I am not doing the test of PicoForms myself. I just need to produce the document.
Steven Pemberton: Mark Seaborne, can I ask you to contact Leigh and get him what he needs?
John Boyer: Which edition does it reference for PicoForms?
Leigh Klotz: It could refer to any one we want I suppose. Whatever PicoForms implements and tests is the version that we can advance to PR.
John Boyer: Mark, can you let us know what version PicoForms will be testing and get an implementation report to Leigh?
Mark Seaborne: Sure.

Action 2007-10-3.5: Mark Seaborne to let us know what version PicoForms will be testing and get an implementation report for XForms 1.0 Basic Profile to Leigh Klotz?

John Boyer: Keith, can you send mail with links to the messages from David Landwehr (see IRC) and report they are fixed?
Keith Wells: I did that.

* Task Force

Steven Pemberton: What is the status?
Nick van: There is a proposal for the charter from Anne. It's quite vague but it's a start. There is some movement on either a phone call or a tech plenary meeting. It is moving again.
Sebastian Schnitzenbaumer: Yes. I am not showing up but that is a red herring, but I should be there. I introduced myself today. There's a question about which wiki version.
Steven Pemberton: I'll either use our existing wiki or use the new one if it's available.
Sebastian Schnitzenbaumer: I think Anne's just asking for web space for storing documents, not necessarily a wiki; maybe he can just get space from Chris Lilley.
Nick van: Then only he can edit it?
Sebastian Schnitzenbaumer: No, it would be access public. But we need to confirm.
John Boyer: Once the space is set up, let us know and we'll link to it.

* Progress closing action items?

John Boyer: Submission examples from Steven? Will you be able to give us some examples?
Steven Pemberton: Yes, I'll do that Friday.

John Boyer: Uli reported that he is working on the insert/delete examples. Hopefully some of you have had a chance to look at the first four; they're more intelligible than our past ones.
Sebastian Schnitzenbaumer: Did Uli do the past ones?
John Boyer: He did a good job to start, but I asked for a couple of other adjustments on the presentment to make the general pattern obvious at the beginning.
Sebastian Schnitzenbaumer: I'll tell Uli.

Nick van: For readonly examples, do we want them in the same format as for insert? The data before and after will be the same.
John Boyer: How many examples do you have?
Nick van: I have ...
John Boyer: We don't have to go as wild as we did on insert/delete, because we had a last call comment to document common usage patterns, and it feeds XForms 1.2 and turn them into new vocabulary to streamline "append an element," "prepend an element," "duplicate," etc. with an exact ... More like Uli's original format would be sufficient.
Nick van: I did the same as the other examples.
John Boyer: If there's anything to be done other than wiring them in we'll correspond.

John Boyer: The Schema...Nick did a lot but there's one left.
Charlie Wiecha: That's me, actions
Leigh Klotz: What about the unnecessary toplevel elements?
Nick van: New ones are local and old ones are left toplevel.
John Boyer: We can put together a better schema as time goes on. Jan created a version of the Schema that did move some elements away from being toplevel, so maybe we could take a look at that. I think Mark Birbeck is working on a highly-modularized Schema and there was some concern that it was using features that some Schema engines wouldn't support; I'm prepared to believe it but it seems fishy.
Leigh Klotz: I think this is our last chance to say that label can't go just anywhere, or put it in group.
John Boyer: That's why the schema isn't normative any more; the main body of the document doesn't say that. Also we said that the submission children could be in any order.
Sebastian Schnitzenbaumer: The validation battle is lost.
John Boyer: I don't think runtimes are using it, but it's helpful for validation.
Leigh Klotz: Really all we need out of it is the data types.
Sebastian Schnitzenbaumer: Technically we could, but I think we shouldn't.
Leigh Klotz: We could use an RNG Schema.
Sebastian Schnitzenbaumer: Someone should write one.
Leigh Klotz: In fact we have an RNG Schema. Micah wrote one and I'm an editor of it.
Sebastian Schnitzenbaumer: Let's talk about it.
Leigh Klotz: Next time.

* Meeting Ends


* IRC Minutes