W3C Forms teleconference November 2, 2011

* Present

Nick van den Bleeken, Inventive Designers
Philip Fennell, MarkLogic
Alain Couthures, AgenceXML
Steven Pemberton, CWI (chair)
Leigh Klotz, Xerox (minutes)
Uli Lissé, DreamLabs
John Boyer, IBM

* Agenda


* XML Prague F2F


Alain Couthures: I will be able to check after TPAC.

* XForms Test Suite 1.1 - Test case 2.4.a


Leigh Klotz: Seems like the schema is wrong; ask John Boyer. My belief is that it is just a Schema conformance issue.
John Boyer: We use a live XML schema engine that isn't tweaked.
Leigh Klotz: Because we left off the elementFormDefault="qualified" the elements are in the wrong namespace. Before we agree, we'd like to make sure it doesn't break any implementations.
John Boyer: I think it won't break my implementation.

Nick van: I will update the test suite. Do we need to create a new revision of the test suite?
Leigh Klotz: This is entirely a schema conformance issue, no XForms.
Nick van: If we publish it do we need an updated implementation report.
Steven Pemberton: I'm not sure the test suite is live, but if the test reports refer to an earlier version, maybe we should start a new version. We have that for the next version of the spec.
Nick van: So change the test suite for 2.0, not 1.1.
Leigh Klotz: That's fine with me.
Steven Pemberton: Hmmm...
Nick van: Did we ever publish errata for XForms 1.1? Then we can put the new test suite there.
John Boyer: I'm not sure the test suite is normative. I don't think we need to version the test suite.
Nick van: Then I'll update it in CVS.
Leigh Klotz: Isn't there a ZIP file as well?
Nick van: I'll make the zip file as well.
John Boyer: I'm not sure this blog article is correct...checking XML Schema Part 0.
Nick van: The default is unqualified.
John Boyer: When you say qualified you can have namespace prefixes or not, but unqualified you are forced to have no namespace prefixes. Sounds like he is getting a real error.
Steven Pemberton: I think that is the case.
John Boyer: It looks like putting elementFormDefault=qualified is a safe thing to do; it's not going to change the behavior and so we should do it. My recollection is that qualified can have a namespace prefix or in the target namespace.
Leigh Klotz: When you say elementFormDefault=qualified and name="ccnumber" that ccnumber is in the targetNamespace?
John Boyer: It is in the name="ccnumber" in both cases. The name is not interpreted according the defaultnamespace.
Leigh Klotz: As I recall, the name is always in the targetNamespace because it makes no sense to do anything otherwise. Does elementFormDefault=qualified refer to type and other attributes in the schema?
John Boyer: Making the change does not hurt anything, but it is correct either way.
Nick van: Then it's strange to make the change.
John Boyer: It wouldn't be the first time a blog is wrong.
Leigh Klotz: We should ask him to validate the instance as a standalone file with the schema pulled out. This has nothing to do with XForms.
John Boyer: No, he's right, we've done it wrong. If you look carefully, the PO example doesn't declare a default namespace. All are in empty namespace.
Nick van: I just did that in Eclipse and elementFormDefault=qualified is needed. http://www.w3.org/TR/xmlschema-0/#QualLocals http://www.w3.org/TR/xmlschema-0/#QualLocals "All such attributes' values may each be set to unqualified or qualified, to indicate whether or not locally declared elements and attributes must be unqualified. ... To specify that all locally declared elements in a schema must be qualified, we set the value of http://www.w3.org/TR/xmlschema-0/#QualLocals to qualified."
Steven Pemberton: We should say thanks and we'll correct it.
John Boyer: We're not that surprised that others have passed because we don't emphasize structural validation; we allow structurally-invalid instances with relevance and switch and so on, until submission.

Resolution 2011-11-2.1: We change test case XForms Test Suite 1.1 - Test case 2.4.a to add elementFormDefault="qualified" http://lists.w3.org/Archives/Public/www-forms/2011Oct/0000.html

Action 2011-11-2.1: Steven Pemberton to reply to http://lists.w3.org/Archives/Public/www-forms/2011Oct/0000.html saying he is right.

ACTION-1841 Nick Van Den Bleeken to update test 2.4.a according to http://lists.w3.org/Archives/Public/www-forms/2011Oct/0000.html

* A processing-instruction to activate XForms


Steven Pemberton: I found this in October 2003: http://lists.w3.org/Archives/Member/w3c-forms/2003OctDec/0132.html
Steven Pemberton: He called it xml-edit instead of xml-form. He did give it a type. We did have this which we were planning to produce at some day, as a REC-track document. I don't recall why we didn't. I'm quite happy doing a spec, though I'd prefer it as a Note instead of Rec-track document.
John Boyer: How does the browser know about this processor instruction?
Leigh Klotz: You have to have an XForms engine in your processor.
Alain Couthures: There is already a PI for the XForms processor.
John Boyer: Is the browser detecting the processing instruction?
Alain Couthures: The processor.
Leigh Klotz: There's two PI's: one says do XForms and the second says do this.
John Boyer: The browser knows about xml-stylesheet and that turns on XSLTForms. So something has to tell it to launch XForms.
Leigh Klotz: Or XForms looks for this. We have done similar things in server-side processing, where PIs tell you to put in XForms, etc.
Steven Pemberton: There is a difference; Mikko's system knows how to put use the correct editor.
John Boyer: So the browser knows to run this PI based on xhtml+xml?
Nick van: It's just XML. You have to have some kind of native extension in the browser.
John Boyer: So it says run me when you get this content type? The PI provides information? The type= isn't really activating anything.
Nick van: In the example, you can do XSL-FO or use different forms. It's a bit like CSS stylesheets.
Steven Pemberton: Yes, he modeled it on that and implemented it in X-Smiles.
John Boyer: What does this processing instruction give us that isn't already in there with content type?
Nick van: You can put this in your XML document. But you can edit only one instance at a time.
Steven Pemberton: I'm not sure the semantics are the same for Alain.
Alain Couthures: I think they are the same.
Leigh Klotz: Go to http://xformstest.org/klotz/2011/05/MVC/buy-test4.xml and then press F1 and Profile and that uses the PI.
Alain Couthures: It's applied internally. It's possible to do this without editing the form, putting the PI in place.
Steven Pemberton: Which bit does your PI do?
Alain Couthures: Sometimes people want to save instance data.
Alain Couthures: The profiler is a separate document.
Leigh Klotz: Here's how it's loaded:

xsltforms.js:1651:              s += <code><?xml-form type="application/xhtml+xml" href="</code> + Core.ROOT + <code>xsltforms_profiler.xhtml" instance="profile"?></code>;

Steven Pemberton: And it loads that in to profile it?
Leigh Klotz: yes.

Steven Pemberton: What do we do now? It's useful but I don't recall why we didn't go forward.
Alain Couthures: I could take an action to edit the document.
Steven Pemberton: I'll get the latest version and maybe the best thing would be to convert that into wiki form.

ACTION-1843 Steven Pemberton to Convert Mikko's PI document to Wiki form

* IRC Minutes


* Meeting Ends