W3C Forms teleconference March 11, 2009

* Present

John Boyer, IBM (chair)
Nick van den Bleeken, Inventive Designers
Leigh Klotz, Xerox (minutes)
Uli Lissé, DreamLabs
Steven Pemberton, CWI/W3C

* Agenda


* Previous minutes

* Upcoming AC Meeting

John Boyer: The discussion topic appears to be rebranding XHTML2 and XForms as "design languages" and leaving HTML5 as a "browser language."
Steven Pemberton: When the AC voted on the XHTML2 and HTML5 charters, both charters said they would be dealing with XHTML1-family. That was pointed out after the vote, I think by IBM but I'm not sure. They agreed to change it and take it out of the HTML5 charter, but the change got made after the start date of the charter. Someone complained recently and the edit was reverted. So we now have two charters that say they're responsible for the same thing.
John Boyer: What's the goal?
Steven Pemberton: The design as with XForms is to allow a wide spectrum of implementations, server-side, client-side, and in between. So they're saying we should focus on server side.
John Boyer: Sebastian has already commented. ODF also includes the XForms model. Surely that's a client-side interactive document. It shows it works with things outside XHTML. It's also the case that it's more efficient to deliver the XForms to the client side. That's the goal of the Ubiquity project as well, to provide something on the client. For Accessibility, we focus on declarative behaviors, as that's easier for other programs to figure out what's going on.
John Boyer: It's harder to figure out what's going on with these JavaScript things together than declarative languages.
Steven Pemberton: When I used Slidey with Ubiquity, Slidey took all the input characters.
John Boyer: I've mentioned this to our AC rep; others should respond as well. Leigh? Nick?
John Boyer: I think you should spend some time formulating the response.
Leigh Klotz: I haven't seen the AC list.
Steven Pemberton: You can read the member archives.
John Boyer: I can send a link. Or if someone else has it handy.

Steven Pemberton: http://www.w3.org/2002/09/wbs/33280/survey2009/results is there and member only but has no contents.

John Boyer: A further informative link from Philippe Le Hégaret: http://lists.w3.org/Archives/Member/w3c-ac-forum/2009JanMar/0177.html http://www.w3.org/2002/09/wbs/33280/survey2009/

John Boyer: Sam Ruby will be at the AC meeting from IBM.
Steven Pemberton: I saw he received a job offer from Microsoft, on his blog.
Leigh Klotz: There are rumors Microsoft will be moving to either WebKit or a research project they have with a sandbox browser.
Steven Pemberton: I found a library that turns IE5 and above into into IE7 compatible. http://www.w3.org/2002/09/wbs/33280/survey2009/
Nick van: this is a cool library; used it quite some times.

John Boyer: Sam said there was some interest in Henry Sivonen's validator and adding XForms for HTML to it.

Action 2009-03-11.1: Leigh Klotz to provide RNG/RNC for XForms for HTML for Sam Ruby for Henry Sivonen's validator.

* XHTML2 and XForms

http://lists.w3.org/Archives/Public/public-forms/2009Mar/0020.html http://www.w3.org/2009/03/10-xhtml-minutes.html#item02

Steven Pemberton: We discussed the clashes between XML Events 2 and XForms 1.1. I still have to investigate. I'm convinved there's much of a problem.
John Boyer: It was more or less a noop for us to move to XML Events 2.
Nick van: There were one or two attributes named differently.

Action 2009-03-11.2: Nick van den Bleeken to look up issues for XML Events 2 and XForms 1.1 and respond to Steven Pemberton.

Steven Pemberton: And there was the resource attribute; we wish we'd done the thinking for RDFa. It's an attribute we introduced in XForms as a consistency with the child element; it's functionality is the same as src, if I'm not mistaken.
John Boyer: Same a @action
Steven Pemberton: So one option would be for us in XForms 1.1 to XHTML 2 to say that you have to use action, not resource.
John Boyer: What is the meaning of the resource attribute from RDF?
Steven Pemberton: RDFa. It specifies the object of a relation. It's similar to href but it's not the intention that it's clickable. It specifies a URI that's part of a relationship.
Leigh Klotz: Didn't Mark say that resource was what the submission was about anyway? Or are you saying that since @resource is what makes it "clickable" it's not OK?
Steven Pemberton: I see. I thnk it's ok to use @resource then.
John Boyer: We also have it on the load action. It's not exactly UI clickable but it's like submission; you click something that invokes it. For instance you might say it's implicitly clicked.
Steven Pemberton: Since they both have the same datatype, the problems are much less. That takes us to encoding. If we accept the same datatype then similar arguments can be used. In our case, it's a list of character encodings, such as UTF-8. I think in XForms 1.1 it's just a single character encoding, as it can be used for content negotiation.
John Boyer: Ours does like XSLT.
Steven Pemberton: But we can say in this case it has to be a single one and we're done.
John Boyer: So in prose we say it's a single one?
Steven Pemberton: Every element gets a single one.
Leigh Klotz: So we put the string restriction in our Schema and restrict the list length to one, or just put in the base type?
Steven Pemberton: The details can be worked out depending on whether it's done by hand or by modularization.
John Boyer: What is encoding?
Steven Pemberton: It's the encoding for the href, as href can be on everything.
John Boyer: Ours is for the coding to be sent, not the encodign of the thing at the other end of the resource.
Steven Pemberton: That's OK. As long as the type is the same you can say that's what it does.
John Boyer: But we're destroying your ability to do conneg on submission.
Steven Pemberton: There's no @href on submission so there's no need for it anyway. It doesn't do anything anyway.
John Boyer: So it's really just the datatype fixup.
Steven Pemberton: Yes.
John Boyer: So if you have a specific list, is it exhaustive?
Steven Pemberton: You've seen the message from Markus.
John Boyer: What type do they use? We used CDATA so how do we list the encodings?
Leigh Klotz: I suspect they just excluded whitespace so they could have a list.
John Boyer: What type did they use?
Steven Pemberton: http://www.w3.org/TR/xhtml2/abstraction.html#dt_Encodings
John Boyer: I'll look at ours.
Leigh Klotz: It's not an XML Schema list; it's comma-separated and is exactly RFC2616.

Leigh Klotz: I suppose we could change our encoding to be a list. Then we submit using the first encoding that works.
John Boyer: So for XForms 1.1 or XForms 1.2?
Leigh Klotz: For 1.1 you could say in prose it's the first, but for 1.2 you could say you can use the list.
John Boyer: For XForms 1.1 we say we use XSLT's encoding. I think the XSLT encoding attribute says it's an indication of the encoding.
Leigh Klotz: It's encoding for output.
John Boyer: Can someone firm this up?
Steven Pemberton: Let's summarize with Markus first.
John Boyer: If someone put encoding=UTF-8 today it would conform. You want to know if someone says UTF-8, UTF-16. We'd probably just pick UTF-8, but if someone put something odd...
Steven Pemberton: http://www.iana.org/assignments/character-sets
John Boyer: Who should do this?

Action 2009-03-11.3: Steven Pemberton to coordinate with Markus Gylling on XHTML2+XForms attribute clashes.

Steven Pemberton: I think it's broader; I don't think you have to panic. Currently it's rather vague. We'll just tighten it up a bit. I don't think we have to make changes.
John Boyer: So we can adopt the tightened language in XForms 1.2?
Steven Pemberton: Yes.
John Boyer: Good deal.

Nick van: [irc] http://lists.w3.org/Archives/Public/public-forms/2008Sep/0029.html http://lists.w3.org/Archives/Public/public-forms/2008Sep/0029.html didn't we talk about this also on a f2f meeting, can't remember which f2f
Steven Pemberton: Thanks for completing your action item so quickly!

John Boyer: We need a meaning for @target on submission/@replace=all. On @replace=instance it's an expression. He's pointing out the href target.
Steven Pemberton: In XML Events 2 dispatch target is now target-id.
John Boyer: That's probably one of the conflicts in XForms 1.1.

Steven Pemberton: I don't think you can tell a URI from an XPath expression by data restriction anyway.
John Boyer: You can tell from the base.
Steven Pemberton: It's a little odd to say you have to determine based on other attributes. What if it changes dynamically?
Leigh Klotz: Then it changes. It's a co-occurrence constraint.
John Boyer: It's interpreted at the time it's needed.
Leigh Klotz: It's clearly not statically analyzable then, but a lot fo the rest of the stuff isn't either.
John Boyer: Then maybe it's just a string. I think it needs more thought.

* IRC Minutes


* Meeting Ends