See also: IRC log
Stuart: Some actions will be closed without discussion if no one asks to discuss them.
Dan: I'd like to discuss some of them (scribe missed exactly which ones)
Proposed: Accept http://www.w3.org/2007/08/13-tagmem-minutes.html as a true record?
<DanC> sounds like we're approving the version of 2007/08/13 18:10:12 without "DRAFT" at the top
Stuart: I propose 10 Sep
... 3 Sep is US Labor Day. 27 Aug is UK public holiday.
Dave gives regrets for 10 Sep
<Rhys> Rhys does give regrets for 27 August
Stuart: Raman can you scribe 10 Sep?
Accepted, next meeting is 10 Sep 2007.
Stuart: I put a wbs form up,
... Two slots work: this one and 12-1:30p ET Thursday
Henry: That slot isn't available
... I think this slot has significantly impacted our productivity.
... I'd like to look a little harder.
Raman: I'm willing to accept an 8:30PT slot, I'll just have to be a little late
Henry: Tim would miss the team lunch, but Amy thought he would consider it.
Norm: What are you proposing?
Henry: I'm proposing 11:30a-1:00p ET on Tuesdays
Norm: I'm afraid I have a conflict on Tuesday's now.
<DanC> ok by me
Norm: I can see about getting it moved.
Henry: The only other slot that has any obvious hope is 1:00-2:30p ET on Thursdays.
Stuart: Rhys has a "no" on that slot.
Rhys: I could be persuaded to live with it.
Henry: I propose we try for 11:30-1:00 ET on Tuesdays pending Norm moving his meeting his meeting and TimBL agreeing. Failing that we come back to consider 1:00-2:30 ET on Thursdays.
Stuart: If we can shift to another day this week, I'd like to reclaim the next two weeks. Meeting, on Tuesday next, for example.
DanC: Call both and then cancel one.
Stuart: I propose that we next meet on Tuesday, 28 Aug 2007 at 11:30 ET.
Stuart: And Thursday, 30 Aug 2007
at 1:00p ET.
... I'll cancel one or both depending on the survey outcome.
Stuart: DanC, did you get approval for the URI?
DanC: I didn't really make any progress on either action.
Stuart: Any idea when you might make progress?
DanC: Its straightforward for me to get a response after 27 Aug.
Stuart: Ok, we'll wait until its straightforward.
Stuart: I've put up a logistics page
Stuart: I've put out a call for agenda items, we could talk about that now.
Dave: I'd like to see XML Versioning on the agenda.
Stuart: You gave regrets.
Dave: Yes, but I'm willing to call in; so maybe we could do it late in the day UK time.
Stuart: I believe Rhys would like
to talk about httpRange-14.
... Distributed extensibility, microformats, HTML 5 looks like a possibility
Stuart: The other thing I think we said was that we'd discuss the confluence of Web 2.0/Web Architecture, but we don't really have material to discuss.
DanC: We have the thread on fragment identifiers.
Raman: I'm not going to try to call in, but I don't mind if you talk about it in my absence.
Stuart: I'd like to see some email discussion to help shape the agenda.
Some discussion of HTTP Redirection, etc.; Rhys is working on related materials.
DanC: I'll try to prepare for the
distributed versioning discussion.
... Where's Semantic Web Architecture?
Norm: Stalled with TimBL and I. Probably not productive unless Tim and I can get together and write more words.
Stuart: Does that look pretty good? Are there other things people would like to see?
DanC: How far are we from our f2f?
We're three calendar weeks away
DanC: The agenda should be announced at T-minus two weeks.
(Per W3C process)
Dave: So can we talk more about
... I brought up a topic, versioning, and we picked a time slot, but the obvious next thing is, what will we do about versioning?
... Seems to me that we should take another pass at the documents. That means reviewers.
Stuart: I'm on the hook to review
the terminology document.
... Do we have reviewers for any of the other documents.
Norm: I'll review the XML part again if you point me to the most recent version.
Stuart: So we still need a reviewer for the strategies document.
Dave: Yes, 2.2.2 is where we kind of finished at the last f2f. I did a bunch of work on that, that would be one logical place to pick up from again.
Stuart: This is on the agenda to discuss Norm's action.
<dorchard> The latest versions of versioning finding are listed in the email at http://lists.w3.org/Archives/Public/www-tag/2007Jul/0004.html
Norm reviews his action, suggests that Dave plonk it in somewhere.
Stuart: Is it likely that you can do that before the f2f, Dave?
Norm: Doesn't worry me if we don't add that before September.
<Zakim> DanC, you wanted to note @role module last call coming, and pick back up on modularity in CDF, XHTML modularization
Dave: I plan no changes to the documents between now and the f2f.
DanC: There's a @role spec coming
DanC: Why DTDs?
Raman: What's missing is an XML Schema that conforms to XHTML modularization.
Henry: If we think that we should say somthing that would encourage them to not use DTDs; my candidate would be to use the modularization framework and an XHTML 1.1 XML Schema.
Some discussion of the status of modularization; apparently it's back at WD.
<DanC> last call http://www.w3.org/TR/2006/WD-xhtml-modularization-20060705/
Stuart: Is this one we should be reviewing?
DanC: Yes, in particular there's
a form of extensibility that's sort of decentralized in a way
that appeals to me.
... There are claims that the subsitution group model they use doesn't work, but I've never been able to get to the bottom of it.
Dave: I wrote something on this too, we should really try to get to the bottom of it.
<scribe> ACTION: ht to review XHTML Modularization (http://www.w3.org/TR/2006/WD-xhtml-modularization-20060705/) [recorded in http://www.w3.org/2001/tag/2007/08/20-minutes.html#action02]
<trackbot-ng> Created ACTION-15 - Review XHTML Modularization (http://www.w3.org/TR/2006/WD-xhtml-modularization-20060705/) [on Henry S Thompson - due 2007-08-27].
Stuart: DanC, did you want to say more about the role document?
DanC: Sort of. One of the ways
that role winds up on my plate is due to accessible web
... One outcome is this role attribute with things like checkbox that help screen readers do better.
... It's been proposed that this role module be stuck in HTML 5. And HTML 5 has completely sworn off DTDs.
... If you're not going to do a DTD for it, what are you going to do?
Raman: One issue is, does this go
in a new attribute? Why not stick it in class.
... The other issue is whether or not it should be namespaced as it is in XHTML 1.1. But firefox doesn't care because it's been hardwired.
... Saying that role has to be in a namespace is bogus.
... I started some of the role work; it's been carefully done as a module so that the WAI folks could decide what the right-hand-sides should be.
... They actually created two attributes in two namespaces, the state values go in one and the role values go in the other.
... That's a three line summary of all the HTML 5 mail.
DanC: One of the design patterns for HTML 5 is "no no no, not namespaces, let's figure out what the authors need and give them a list"
Raman: The problem with role is
that it's cross-cutting so that doesn't really work.
... HTML 5 also talks about following existing practice, so they're trying to go in two different directions.
Stuart: We're talking about this under versioning-41. Is this really versioning, or is ti an HTML 5 topic?
DanC: The versioning topic is about substitution groups in modular design.
Dave: If substitution groups don't work in 1.0, do they also not work in 1.1?
DanC: XML Schemas is about to go to last call.
Henry: There is one change in 1.1, allowing multiple substitution heads. That's been alleged to be problem. I'll be taking a look.
Stuart: Do we agree that Norm's action is closed?
<scribe> ACTION: dorchard to incorporate the NVDL text into the findings. [recorded in http://www.w3.org/2001/tag/2007/08/20-minutes.html#action04]
<trackbot-ng> Created ACTION-16 - Incorporate the NVDL text into the findings. [on David Orchard - due 2007-08-27].
<raman> For the record the collection of properties refered to as state is a collection of attributes -- not a single one.
<DanC> ah, right, raman
Norm: I completed the action I was given ages ago.
DanC: I think it needs to be in a test suite somewhere.
Norm: An XML 1.1 parser should accept the document, an XML 1.0 one would fail.
Henry: There are already tests in the XML test suite for 1.1 features.
<DanC> (yup, norm's action is done to my satisfaction)
DanC: We could just turn this whole ball of wax over to the XML CG.
<DanC> (I'd like ht to take an action)
Stuart: We'll mark that action as done and leave the issue pending.
<scribe> ACTION: ht to check that the XML test suite contains an equivalent test [recorded in http://www.w3.org/2001/tag/2007/08/20-minutes.html#action05]
<trackbot-ng> Created ACTION-17 - Check that the XML test suite contains an equivalent test [on Henry S Thompson - due 2007-08-27].
Stuart: Any discussion?
<dorchard> which article?
DanC: What's novel about Sam's message is that he proposes using namespaces in content that isn't well-formed XML.
Raman: I think he's articulated
something important: it's more than the browser vendors who
have a right to decide what goes into HTML 5. User communities
should be able to add new elements too.
... The browser vendors are going to document what they do, but I think it would be a serious loss to the web if that's all we expected to happen.
... I think the important thing is that extensibility is the purview of more than the browser vendors. Henry saw something else when he looked at it.
dorchard: If this is something that would help the HTML 5 WG get closer to what the TAG thinks is the right way to deal with distributed extensibility, then we should support it.
Stuart: Is there a proposal here?
dorchard: We could ask someone to write up a message of support that we could publish as the TAG
<scribe> ACTION: dorchard to review the article and make a proposal. [recorded in http://www.w3.org/2001/tag/2007/08/20-minutes.html#action06]
Stuart: Any other discussion?
DanC: The agenda suggests continuing Ramans action without comment. I'm content that he finished that action.
Stuart: I can't point to that document off the top of my head. It goes back to our Google meeting.
Stuart: OK, with that pointer we'll call the action closed.
DanC: Are we going to maintain this?
Raman: If you send that document to the HTML 5 WG mailing list, I don't think we'll get anything productive back.
dorchard: We need to do our job anyway
Stuart: Do you think it's ready to be sent?
Raman: Yes, if we agree that those are the things we want to say. I just don't think there's value in doing a lot more work in a vacuum.
Some discussion of what actions we could take
Raman: What we've prepared is an outline of items we believe should be covered
Stuart: We should formulate our opinion regardless of what the outcome might be
Raman: I think we should flesh this out at the f2f and then send it.
<scribe> ACTION: Raman to update the document with the latest discussion [recorded in http://www.w3.org/2001/tag/2007/08/20-minutes.html#action07]
<trackbot-ng> Created ACTION-19 - Update the document with the latest discussion [on T.V. Raman - due 2007-08-27].
<scribe> ACTION: Stuart to discuss with DanC what might be most useful and productive wrt providing input to the HTML WG [recorded in http://www.w3.org/2001/tag/2007/08/20-minutes.html#action08]
<trackbot-ng> Created ACTION-20 - Discuss with DanC what might be most useful and productive wrt providing input to the HTML WG [on Stuart Williams - due 2007-08-27].
Henry: We should separate our technical discussions from strategically how to communicate this with the WG
Raman: The more general TAG
discussion I think is that there's an asymmetry between URLs on
the server and URLs on the client.
... On the server, everything after the question mark is sent to the resource as a bunch of arguments.
... Similarly, in the URL you have the # which passes everything after the hash to the client to handle it.
... The community knows how HTML browsers use fragment identifiers; and we know how RDF does it; but there's very little else been said.
... The problem with saying "the client" is the question of which client.
... At this point, I think there are more details that need to be written down if we expect to get interoperability.
Stuart: One of the things that interested me is that no where in that discussion did we talk about the media type.
Raman: I danced around that
carefully. That only really comes up with respect to HTML. It
came up years ago when I was at Adobe with respect to opening
... I don't remember the status of that for the media type application for PDF
... If you write .../?a=1, you'll never expect that to work if you send it to a different server. On the other hand, if it's # then you would expect different clients to act in the same way.
<Zakim> ht, you wanted to ask for clarification
Henry: When you said "which client" you didn't mean that in any given interaction there is more than one client, right?
Raman: Right. Clearly the CNN URL
Henry: It's certainly the case
that there are other media types that specify the
interpretation of fragment identifiers.
... They are careful and they get real value out of them.
Raman: I think Adobe did that, but I don't clearly recall at this time.
Henry: It would be perfectly
coherent for someone to define a media type that said what you
do with fragids that begin with slash is that you issue an
XMLHTTPRequest of the following form...
... But that's not what either text/html or any of the other HTML mime types say. So CNN is flat out in violation of the spec.
Raman: I'm comfortable with that view too, but then you run into the counter-argument that "it works dummy".
Henry: Breaking the rules is bad? I don't want to write a TAG finding that says that.
Raman: It seems to be all about
breaking the rules and making a hack that works these days and
I"m a little worried about that.
... There are two reasonable positions: the MIME type says what the answer is.
... We can either say you broke the rules or we can change the rules. It actually took me a while to figure how it works.
... This is a lot like the "CSS sprite" hack.
Stuart: We're about out of time. I think we'll come back to this.
Henry: If there's a TAG issue
here, it's something about what proper script author
responsibilities are in the context of the standards that
they're operating with.
... It isn't obvious where you would go to find a basis for saying "gee, this doesn't constitute good citizenship"