W3C Forms teleconference August 6, 2008

* Present

Charlie Wiecha, IBM
Nick van den Bleeken, Inventive Designers
Paul Butcher, x-port.net
Keith Wells, IBM
Kenneth Sklander, Picoforms
Erik Bruchez, Orbeon
Leigh Klotz, Xerox (minutes)
Doug Scheppers, W3C
John Boyer, IBM (chair)
Roger Pérez, SATEC
Steven Pemberton, CWI/W3C

* Agenda


* Previous Minutes

* XForms Event


John Boyer: Any info? Steven's not here.
Leigh Klotz: There are some people I've been trying to get involved in this, but not enough notice now I worry.
John Boyer: Perhaps Roland Merrick.

* Headers

Leigh Klotz: I got tasked with writing up the simple informative changes, but those turn out not to be sufficient.
John Boyer: What are the issues?
Leigh Klotz: Serialization mechanism and possible errors for multiple headers with the same name, particularly with HTTP but generically; replace vs. append of header value; HTTP Accept header in case of submission/@replace=all vs submission/@replace=instance and whether the user agent gets to decide or the header element gets to decide.

* Implementation report progress

John Boyer: Are there other implementations?
Leigh Klotz: Orbeon, I think.
John Boyer: Erik?
Erik Bruchez: So far none of our users have hit that problem. It seemed that it was reasonable to explain better what happens with multiple headers with the same name.
John Boyer: It sounds like there was a test that Mozilla was failing? Once that has modified, for that specific test, would it be possible to see if you pass it?
Erik Bruchez: I think that would be doable.

* XForms Event


Steven Pemberton: [joins] At the moment, the dates are at risk.

* XSD 1.1 last call review deadline 12 Sept.


John Boyer: Does anyone feel this might impact our modularization?
Steven Pemberton: I personally feel we should seriously review it.
Kenneth Sklander: [irc] i can review it, as we are implementing xsd 1.1

Action 2008-08-6.1: Kenneth Sklander to review XSD 1.1 last call review deadline 12 Sept http://lists.w3.org/Archives/Public/public-forms/2008Jun/0059.html by next Wednesday

* Forms Task Force

John Boyer: The HTML WG says the charter for the task force has expired without anyone producing anything, but I'm not sure how there's a date on the charter. Does anybody have any more information on why people believe the charter has expired for the task force?
Steven Pemberton: We can extend it if there is a need to produce documents.
John Boyer: We are both working on forms simplification. It seems like the stuff we're doing is very big compared to the size of the group.
Charlie Wiecha: I'm interested in doing implementation of the simplified syntax in Ubiquity, but can't work on the task force.
John Boyer: If all we have is a spec and nothing else to show, then that's not...
Charlie Wiecha: It would be hard to reconnect later on. We're interested in Dojo and YUI and the simplified syntax model in Ubiquity.
John Boyer: So I can push on the question in the HCG about why the charter for the task force is less than December 2009, the groups charters.
Charlie Wiecha: Our interest in implementations is part of our activity.
John Boyer: We have a technical plenary coming up in October; maybe we could have a demo of a portion of it.
Charlie Wiecha: I would think so.
John Boyer: We can get together then and show it working. We might need to have changes to XForms 1.2.

Action 2008-08-6.2: John Boyer to report to HCG that we are continuing to work on the joint task force issues, and in particular are doing implementation work to validate some of the concepts, and are tentatively planning to meet at the tech plenary.

* Review of XPath Functions Module


John Boyer: The first issue is what to call it. It's not "The" XPath functions module. Other XForms modules will add functions to it.
Nick van: [irc] The standard XForms XPath Functions module, The base XForms XPath Functions module
Leigh Klotz: How are they added?
John Boyer: By prose. http://www.w3.org/MarkUp/Forms/specs/XForms1.2/modules/instance/bindingAttributes/index-all.html#expr-li
Leigh Klotz: So it's not added to the XPath functions module, but added to the conglomerate.
John Boyer: Yes. I misspoke. It's added to the function library.
Leigh Klotz: I'm with Nick: core, or base.
Paul Butcher: These are all data, but no instance functions.
Leigh Klotz: There's other stuff like dates.
John Boyer: Yes, Paul, good point. The instance stuff is all removed. The only place instance caused a problem was the current function. I had to merge the two instances into one chunk of XML data.
Leigh Klotz: So channelling Mark, I'd say we can take an extreme position and break this up into a date and time functions spec, a math function spec, move is-card-number() out to where the card type is defined, put if() and choose() in one module, etc.
John Boyer: That's a lot of specs.
Paul Butcher: XPath 2.0 has the date functions.
Leigh Klotz: Another cut is to keep the ones that have XPath 2.0 equivalents out of the same document as ones that don't, so when we move to XPath 2.0 we can simply not update those documents.
Charlie Wiecha: Whatever we do about splitting it up or not, we shouldn't call it The XForms XPath Library.

John Boyer: Where do we put these? Subdirectories in TR space? Can it be TR/xpath-functions?
Steven Pemberton: We can get directories; the spec is at the top level and we can have any number of directories under that.
John Boyer: So we could have subdirectories for modules.
Charlie Wiecha: Is that a mechanism from the standpoint of the process? Is that one or N; that creates different URLs but does it require different last-call processes?
Leigh Klotz: We used to have the split spec as normative.
Steven Pemberton: The process starts from a short-name which points to the version name and you can have directories under this.
John Boyer: So do we resolve to create several specs out of this?
Leigh Klotz: I don't think we've resolved to do it yet.
Nick van: Are we ready for dozens of specs? I know Mark is pushing it.
John Boyer: Well, this module might be a good test case. Its parts can be used separately.
Leigh Klotz: We're not the first to do this as Mark says.
John Boyer: We're taking the XForms spec and producing 30-40 specs out of it.
Leigh Klotz: Well, the monolithic spec isn't working for us, so I'm with Mark.
John Boyer: OK. Why can't the is-card-number be introduced separately. I guess it's useful. So it seems like there's an action item to propose a breakdown and resolve to do that or not.
Leigh Klotz: If we're going to do that, that's the process.

Action 2008-08-6.3: Leigh Klotz to propose a breakdown of http://www.w3.org/MarkUp/Forms/specs/XForms1.2/modules/instance/xpathFunctions/ into multiple modules today

John Boyer: I didn't find it yet useful to refer to XHTML modularization. Steven?
Steven Pemberton: That's fine by me. Assuming we're planning on using the same techniques we should say we're using the ones defined there.
John Boyer: Does it say anything about XPath functions?
Steven Pemberton: Oh, just for this module. No XML Schema right? Then we don't need to use it here.
John Boyer: Binding and other modules do, though. Could you propose a sentence that we can use?

Action 2008-08-6.4: Steven Pemberton to propose spec-ready sentence to import XHTML modularization for binding and other documents.

Nick van: Do we import these or do we say that the XPath library adds them?
John Boyer: From the last F2F, I got that it's the consuming library that imports the functions. The binding module adds context() but it doesn't mention the others. The binding attributes module has a section http://www.w3.org/MarkUp/Forms/specs/XForms1.2/modules/instance/bindingAttributes/index-all.html#expr-lib that says that other modules of the consuming profile may add more modules to the library. So we will have thirty-odd specs out there for people to adopt our technology; however, if someone wants to do XForms, there is an XForms rec that will compose all of them pulled together.
Nick van: So we are defining XPath functions. Can another module add those functions to script?
John Boyer: That's a separate question. I've been focused on modularizing what we have and producing those. It's got strings, nodesets, etc.
Nick van: You can do that in other languages as well.
John Boyer: We inherit a lot by default, though, such as parameter type conversion. we could look at how hard it would be to write them generically over XPath-ese and JavaScript, but we haven't taken that approach yet.
Leigh Klotz: That's what I was trying to do with the IDL, but we ran into just that problem, that we'd have to write XPath in IDL, which was too hard. Channelling Mark again, he could take our XPath functions recs and implement them in JavaScript without our involvement.
John Boyer: We could say that you're allowed to do that.
Leigh Klotz: I don't see the point; people don't need our permission to recast our specs.

John Boyer: I added editorial notes in each place where we found text that belonged in a to-be-written module. I've seen Nick has done the same thing. In one place you sent email to the WG about idref in the bind module. So, Nick, can you put those editorial notes in?

John Boyer: I removed "instance" and "XForms Markup" where I could. The context function was hard. I kept the example of the setvalue element here: http://www.w3.org/MarkUp/Forms/specs/XForms1.2/modules/instance/bindingAttributes/index-all.html#fn-context I got rid of as much as I could. In the example, I explained setvalue and binding. I explained the relevant bits about the element. In the actions module there are triggers and DOMActivates and things that aren't defined in the module; the shortest path is to say "suppose you had an element called trigger with the following properties."
Nick van: [irc] OK, I will clean it up.

* Review of Binding Attributes Module


John Boyer: Some of these just define attribute groups. Every module that needs these concepts will refer to this spec anyway. This spec is heavy on definitions and knowledge of XPath.
Charlie Wiecha: It doesn't say anything about behavior or events. It's just attributes.
John Boyer: UI bindings, model bindings, and action temporary bindings are all different, but they use ref, nodeset, and context. With context we talk about the nesting of elements. There are some editorial to-do's for stuff that needs to go in other modules. In section 6 there's a glossary of terms. There are more terms than were in XForms 1.1.

John Boyer: And for acknowledgements, what do we do?
Steven Pemberton: Leave it empty with a note saying we'll update it when we get to rec.
John Boyer: OK, we'll comment it out for now and figure out a list when we publish.
Steven Pemberton: Typically we add the people in the WG at time of rec and then acknowledge people who did special work.
John Boyer: Meanwhile we need to ping some people.

* XForms 1.0 Basic testing

John Boyer: Kenneth, is there any chance of moving forward testing on this?
Kenneth Sklander: [irc] We sent an impl. report 18 months ago.
Leigh Klotz: If the chair will provide me the report, I will pair it with the text.
Kenneth Sklander: [irc] do we need additional things? I will forward it to you, leigh

* IRC Minutes


* Meeting Ends