HTML/XML Task Force

Meeting 4, 18 Jan 2011


See also: IRC log


Noah, Robin, Henri, Yves, Norm, Michael Champion, Anne, Michael Kay, John


Accept this agenda?



Accept minutes from the previous meeting?

-> http://www.w3.org/2010/html-xml/2011/01/11-minutes


Next meeting: 25 Jan 2011

Noah is at risk for 25 Jan.

Use cases/goals

Norm: The only new use case I saw in the last week was XForms, which seemed controversial.

MikeK: I think there needs to be a story as there are a community of people that want to use it with HTML

MikeC: Is it a separate use case because there are mixed names, or is it just another XML vocabulary?

MikeK: It's a use case in the sense that doing it mixes HTML and XML. It has particular problems, which is why we consider more than one.

Henri: I have a bit of a problem with calling using XForms a use case; as opposed to creating an interface to allow users to submit values to a server.
... I think XForms presupposes a solution to the problem. It's a use case in terms of vocabulary mixing.
... But I refer back to my email; there was a task for explicitly for XForms+HTML and it petered out. I think that's evidence that there isn't as much interest here as it might seem. They've missed their opportunity to have it discussed.

<Zakim> noah, you wanted to answer Norm on is it a use case (after Henri)

Noah: I guess my feeling is a bit in the middle. With respect, I think Henri goes a little further than I would with taking this off the table. I think the TF was about putting XForms in HTML.
... Here, were talking about the broader issue of low hanging fruit or high benefit to cost ratio wrt lining up HTML and XML.
... The use case might be interactive form input. That's like saying to someone who asks for a C language interface, "you don't want a C language interface..."
... Just saying XForms is listing a technology. I think we could say that one or more communities have vocabularies that be mixed.
... I also think it's in the nature of what we're doing that we want to describe general facilities. They might not be fully justified by any particular use case.
... Some of the email suggested an ambiguity in defining the use case; I think we should resolve that.
... Some folks refer to XForms as an XML island; others seem to be thinking of HTML in an XForms kind of template.

John: What's a use case and what's a solution depends on the level we take. In one sense HTML isn't a use case, it's a particular solution to the problem of dispaying information on the web.
... At our level, I think it's ok to describe technologies as a use case. We want to consider the question of how I can use technology X.
... At some level, the question is how do we keep the machines turning and ourselves employed.

<Zakim> darobin, you wanted to it's not about XForms, it's about the integration problems that XForms might demonstrate

Robin: I don't much care about XForms as a solution is a use case or not. The part that's interesting here is, if there are specific problems that XForms has integrating into HTML. Those are problems that might surface elsewhere.
... I'd like to look at the problems that come up, even if XForms as a solution aren't within our remit.

Anne: The problem is that not everyone accepts that XForms is a technology that we need. Therefore looking at the underlying use case might be better.
... In some sense we started the WHATWG because we didn't like XForms.

John: But some people do.

Henri: I think Robin has a point. XForms demonstrates a different kind of integration. MathML and SVG are pretty much self contained islands. There are certain well defined exit points.
... XForms intermixes more with its host language. In that sense it's a different kind of vocabulary. The problem that I have with running too far in that direction is that if it happens that the vocabularies that will get integrated are all like MathML and SVG, then preparing for integrating other types may be architectural astronautics.
... We might be doing work for a case that we'll never actually need. It might be that integrating the vocabularies that intermingle more is in the "you don't need it" department.

Norm: Does anyone have other use cases in mind?

Noah: This could just be me, I'm having trouble answering that question. We've done a good if somewhat scattered job of a first pass. But I don't think I could write down a net of all this discussion.
... Do you have a process in mind, Norm?

Norm: No, not really. I thought we'd get a lot more use cases and I'm not sure what the next steps are for just five cases.

Henri: I'd like to propose some next steps.
... Writing down the list of use cases we've discussed. I think there were one or two that were added as we went.

<noah> Exactly, Henri, that's what I was proposing. I think we should not only list the 5, but net out in a sentence or two each where they stand in terms of what we learned from the discussion.

Henri: And then going through that list and maybe just copying and pasting the replies from the email and telcons to catalog what solutions have come up so far.
... Perhaps we can conclude that the solutions are all good enough. Or perhaps we'll conclude that more work is necessary.
... Then perhaps we can write this up in a way that will help an XML person learn what the known HTML solutions are for their problems.

<Zakim> noah, you wanted to comment on weighting use cass

Noah: I agree pretty much completely. I suggest we also think about which ones look really really important. We might decide that a couple are really important, some are less so. I'm missing that kind of weighting.
... Either we're going to get really lucky, finding a solution that's all benefit with very little cost, but much more likely we're going to look at choices. We're going to look at changes we can make and be able to say which ones are worth the trouble.

Henri: There's also the chance that we're so lucky that nothing needs to be changed at all.

Noah: Yes, that might happen to.

<hsivonen> maybe a wiki?

Norm: I wonder if we could break up the work.

MikeC: I agree that it's a good idea; since I won't be there to defend myself, you can give me an action item to do something.
... Identify the use cases and divy up the work of going through the mail and minutes to summarize the state of it.
... If there's a template for it, that would be great too.

Norm: Do we want to do it on a wiki?

John: What about people on the mailing list?

Anne: They can send mail and we'll copy it over, we don't have to make this too hard.

Norm: Sounds like we're going to use a wiki. Any opposition?

None heard.

Noah: Time frame?

Norm: I think it would be good to have it by next week, but I suppose the world won't end if it isn't.

<hsivonen> I volunteer to take one

<hsivonen> I could take "I have an XML toolchain and I want to consume HTML5 because I'd like to process HTML5 using XML tools."

1. XML toolchain to consume HTML5 -- Henri

2. HTML5 toolchain to consume XML -- MChampion

3. Islands of HTML5 in XML -- Noah

4. Islands of XML in HTML -- John

5. Making XML easier -- Anne

6. XForms - M. Kay

<anne> (My use case was about easy of authoring of XML.)

<anne> ease*

Norm: I'll take the action of getting an appropriate root page on the wiki and sending out instructions

John: I wanted to mention MicroLark. I have a parser and the implementation of the object model pretty well advanced.

<hsivonen> Did Sam have a seventh use case or was it part of the other ones?

John: I'm running through the test cases, not all of which apply.
... I hope to get something out in about a week.

Henri: Sam sent email about how to generate HTML with legacy tools. So basically, the scenario was that the content generating tools are legacy but you want to generate HTML5 of newer vintage than your tools.

John: You mean HTML tools?

Henri: Yes.

John: How is that relevant?

Henri: Maybe it isn't.

John: Maybe another use case is having XML tools that produce HTML5.

Noah: The text/html serialization?

<hsivonen> for the last point, http://www.w3.org/Bugs/Public/show_bug.cgi?id=11755 might be of interest

John: At least some of it. As long as XHTML isn't fully accepted, people will want to generate the text form. I think we know the shape of the answer, an HTML5 serializer.

Noah: There may be other ways to skin the cat.

Henri: With respect to XHTML, everything except some edge cases like the form feed character, is possible with XHTML that's possible with the text/html serialization.

Noah: A commitment to such bidirectional support in the future can be listed as a possible solution.

Norm: I'll write that one up.

<hsivonen> The "Sample Apps" section at http://about.validator.nu/htmlparser/ is relevant to what was just discussed.

Norm: And after we have the wiki pages up, the community can tell us about the use cases we've overlooked.

Any other business?

<noah> I think we need to point out that this works for the subset of the language that, e.g. does not depend on prefix bindings from xmlns

<noah> right?

<noah> tnx

Norm confirms he'll get the wiki setup.


Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.135 (CVS log)
$Date: 2011/01/20 21:26:19 $