W3C Forms teleconference July 25, 2007

* Present

Erik Bruchez, Orbeon
John Boyer, IBM (chair)
Leigh Klotz, Xerox (minutes)
Lars Opperman, Sun
Mark Seaborne, PicoForms
Nick van den Bleeken, Inventive Designers
Uli Lisse, Dreamlabs
Blake Jones, ViewPlus Technologies/DAISY Consortium

* Agenda


* Previous Minutes

http://lists.w3.org/Archives/Public/public-forms/2007Jul/0087.html IRC supplement: http://www.w3.org/2007/07/18-forms-minutes.html

* XForms 1.0 Third Edition


John Boyer: It wasn't announced on teh W3C site but our PR got published today; the call for review should be going out shortly. It will be announced then. http://www.w3.org/TR/2007/PER-xforms-20070725/ The review period is through August 31st? The AC reps will be asked to vote for it, so please remind them.

* XForms Conference at XML Conference

http://lists.w3.org/Archives/Public/public-forms/2007Jul/0055.html http://lists.w3.org/Archives/Public/public-forms/2007Jul/0036.html http://lists.w3.org/Archives/Public/public-forms/2007Jul/0024.html

John Boyer: We ended up with seven proposals, and we have six slots. We are working on logistics for the keynote. Or we may run all seven and no keynote.

* Need respondent for this www-forms question


John Boyer: We need to answer this.
Erik Bruchez: He's asking to re-use what's in the Schema already. One thing we do is to load the Schema as an instance.
Leigh Klotz: How do you get the labels out of that?
Erik Bruchez: For example, US states, where the labels and values are the same. Maybe annotations.
Leigh Klotz: I saw that Chiba's schema2xforms project is looking for a new maintainer.
John Boyer: Can you handle it?
Erik Bruchez: Yes.
Leigh Klotz: Then I'll answer, or Mark, and point out this other project, but it doesn't belong in the main answer.

Action 2007-07-25.1: Erik Bruchez to answer http://lists.w3.org/Archives/Public/www-forms/2007Jul/0014.html

* Removing binding property from xforms-insert and xforms-delete


John Boyer: Erik said he'd agree and Leigh said he wanted it removed, and I could live with it. Does anybody object to removing it?
Lars Opperman: In Open Office we use the binding as the only means to hook into the model; we use our own controls and they don't have a ref attribute, just a binding. If an application continued this way it would be nice if the binding could stay with the universal notion, though there are problems in other aspects, so we can't maintain that anyway so I would be OK with that going away.
John Boyer: The way this particular property has been defined, it's easy to get a similar way. You can ask for local name of the other elements. Leigh, was your objection about not being able to resolve it bind or ref?
Leigh Klotz: Yes, there's no way to find out the value space.
John Boyer: Or even find out what the bind was.
Leigh Klotz: Unless you load the form itself as an instance ;-)
John Boyer: OK. So we remove it.

Resolution 2007-07-25.1: For http://lists.w3.org/Archives/Public/public-forms/2007Jul/0089.html, we remove binding property from xforms-insert and xforms-delete

Action 2007-07-25.2: John Boyer to remove binding property from xforms-insert and xforms-delete

* Conclusion of email discussion on Issue 103 (submission verb vs. method)

http://lists.w3.org/Archives/Public/public-forms/2007Jul/0067.html http://lists.w3.org/Archives/Public/public-forms/2007Jul/0060.html http://lists.w3.org/Archives/Public/public-forms/2007Jul/0059.html http://lists.w3.org/Archives/Public/public-forms/2007Jul/0047.html http://htmlwg.mn.aptest.com/cgi-bin/xforms-issues/Submission?id=103

John Boyer: I've put the results of this in the editor's draft. http://www.w3.org/MarkUp/Forms/specs/XForms1.1/index-diff.html
John Boyer: I've done other edits to the document; here is a new table in Section 11.1: http://www.w3.org/MarkUp/Forms/specs/XForms1.1/index-diff.html#submit providing a copy of Chapter 3 table. This is to resolve one of Steven's issues of Chapter 11. I realize we don't like to copy stuff but it's useful to have it here. There are other errata addressed here. We now link to NCName throughout the document. So we lose an attribute by this change.
John Boyer: The second change support is the serialization attribute, which turns out to be nice. There is no MIME type for no content; we decided not to use "none" so we used empty string.
Erik Bruchez: I understand the rationale behind it; I wouldn't advocate "none" given that it's not a MIME type. The empty string is a little funny, but "none" isn't it.
John Boyer: I didn't find "none" all that bad when I tried to imagine it being there, because it's not a MIME type. It's not so bad...maybe we can give it some more thought. For the specification at least, there is a bit of parallelism with the empty string.

Lars Opperman: It's a bit funny because some serialization go into the body (POST) and some go into the URL, and there is no mediatype.
John Boyer: That's true; that's already the case in XForms. It describes the fact that we serialize the data but not where in the protocol we put the data.
Lars Opperman: Do we have a way of using the serialization you would use in GET but putting it into an entity?
John Boyer: Yes, the urlencoded-post method keyword.
Lars Opperman: Sure.

John Boyer: Note also we've changed method to verb. It's just spelled differently. Note that the resource attribute should be used in place of the action attribute. This is due to a last-call comment from Steven, and we've gone with resource. We are more proactively indicating that the action is deprecated in favor of resource.
Leigh Klotz: Do we say what happens if you have both?
John Boyer: Yes, the resource takes precedence, and includes the resource element.

John Boyer: For method I realized that it wasn't true that it was the protocol but actually the 'protocol operation.' verb has been deleted, and minor changes to validate and relevant for the default coming from serialization. And serialization changes from a boolean to a string. This is the major change. The default is based on the setting of something else and the other section describes the defaults.
Leigh Klotz: Is there a default for version, the next attribute?
John Boyer: It is optional. What should the default be?
Erik Bruchez: This applies only to application/xml. It seems it should be 1.0.
John Boyer: This is another last-call comment from Steven about attributes. Someone ends up with the action item to put these in. I'd be happy to put them in as long as I know what to write.

Action 2007-07-25.3: John Boyer to put 1.0 as default for submission/@version.

John Boyer: Also indent. It sounds like a boolean. False?
Leigh Klotz: Is that right?
John Boyer: What does XSLT do?
Leigh Klotz: Even if true it's not guaranteed to do anything.
Erik Bruchez: I'm not sure I see a default.
John Boyer: Anyone opposed to a default of no? I think that's what implementations are doing?

Resolution 2007-07-25.2: We default submission/@indent to false.

Action 2007-07-25.4: John Boyer to put false as default for submission/@indent.

John Boyer: Encoding? UTF-8? So we should set the defaults in accordance with XSLT where they are parallel.

Action 2007-07-25.5: John Boyer to put UTF-8 as default for submission/encoding.

Leigh Klotz: Why do we still have mediatype and serialization?
John Boyer: Mediatype controls the header, but serialization is the serialization used.
Leigh Klotz: So mediatype could be text/xml or application/xhml+xml.
John Boyer: right.
Leigh Klotz: Sounds like we need a note.
John Boyer: I can do that.

Action 2007-07-25.6: John Boyer to write note in chapter 11 up front clarifying when to use mediatype and when to use serialization.

John Boyer: We now define in http://www.w3.org/MarkUp/Forms/specs/XForms1.1/index-diff.html#submit-method the concept "submission method." This section is changed: method instead of verb, and the first sentence is a term definition.
Leigh Klotz: Is the "immediately after" a result of schema 1.0 constraints about ordered and unordered and optional?
John Boyer: I hadn't realized that was a schema constraint. It asks for resource, method, and headers. There is a last-call comment from Erik about this already so we'll deal with it then.
John Boyer: Neither the method attribute nor element is required by itself but you do have to specify one of them. That sentence may have been omitted. We say that for submission resource but not for method.

Action 2007-07-25.7: John Boyer to fix 11.7 to add that one of method element or attribute required.

John Boyer: Here is the submission method element: http://www.w3.org/MarkUp/Forms/specs/XForms1.1/index-diff.html#submit-options

John Boyer: In http://www.w3.org/MarkUp/Forms/specs/XForms1.1/index-diff.html#submit-options we cleaned it up and dealt with last-call comments by using the linked submission method. See teh table with the new default serialization. The new row is the one that has any other NCName as application/xml and the submission method is as given; this is old verb attribute.

John Boyer: That's the spec-ready changes.
Erik Bruchez: I think it's great. I'm glad to see verb disappear.
John Boyer: Does this look like the way forward? Does anybody have any objection to using this in lieu of the verb we used to have?

Resolution 2007-07-25.3: We accept the verb changes provided by John.

Action 2007-07-25.8: John Boyer to implement verb changes provided in editor's draft.

* Last Call Issues Needing Approval of Proposed Resolution:

John Boyer: We'll address these as time allows.

* 11.1 The submission Element


John Boyer: We've told Steven here to be using resource in lieu of action. I talked to Steven and left some notes about why we can't use the resource attribute to provide an XPath expression; it's inconsistent with resource being xsd:anyURI as it is elsewhere, and he was satisfied with that. He included a few things in this issue; he wanted the allowable values for @method to be listed. I would argue that that's now done. My comment is that they were listed in chapter 3, so copying it to chapter 11 seemed to do the job there, and he indicated he was satisfied with that. And verb. There's one final issue, noting that the validate and serialize attributes they appear to be verbs, and relevant is a adjective. Submission has 20 attributes, and serialize turned into serialization. I don't think there's an expectation that these things have to be a particular part of speech. Does that seem fair enough as a response?

* 11.6 The resource Element and Attribute


John Boyer: Steven indicated he was happy with these. Does anyone find the language to be a problem?

* Next Meeting

John Boyer: We'll have a meeting next week.

* IRC Minutes


* Meeting Ends