TAG Weekly Teleconference

5 Feb 2007


Raman, Stuart, Norm, noah, Rhys, Vincent, DaveO, DanC, TimBL, Henry


<scribe> ScribeNick: dorchard

<scribe> Scribe: DaveO

<Stuart> Date: 05-Feb-2007

<Vincent> zakim + is Vincent


approved minutes of last week [09:08]

<DanC> RESOLVED: to approve http://www.w3.org/2007/01/30-tagmem-minutes.html as a true record

RESOLUTION: approved minutes of Jan 30 approved

dan: openID and SAML are interesting topics for future agenda [09:09]

noah: paper submitted and accepted for enterprise of web services workshop [09:10]
... invited to present so I'll prep slides

stuart: regrets for Feb 12th [09:11]

next week scribe: Noah


stuart: transition of TAG, things to think about [09:13]

vincent: working on some of my open action items [09:14]

<noah> +1 to what Dan said. [09:15]

dan: if stuart has any questions, I suggest bringing them up.
... but I'm against going through the whole list

<noah> I might be a bit more willing to have group skim, but would be opposed to detailed review of all issues as a use of our shared group time.

Norm: namespace-8, let's finish

stuart: versioning

dan: keep versioning high priority [09:18]

<DanC> (I took an action re urns and registries too; apologies, no progress)

dave: also URNsRegistries [09:19]

raman: how long is transition period?

stuart: while we have them around, we keep flogging them. [09:20]

passwords in the clear

<DanC> (mez's text http://lists.w3.org/Archives/Public/www-tag/2007Jan/att-0038/passwordsInTheClear-52Mez.htm ) [09:21]

<timbl> .me notes today was booked out for me many weeks ago.

noah: if this is going to go on for a long time, give somebody else some ownership

tagsoupintegration [09:23]

timbl: validation seen as a repository of truth [09:30]
... w3c has controlled specs via validation
... validator needs to be revisited
... validation could be what w3c, or even TAG, of what is acceptable or recommended [09:31]

<DanC> (I agree the validator is a big part of the story; I'm coming up with more and more justification for budgeting validator work against the TAG... yeah... what tim's saying.hmm.)

timbl: do you say "no quotes, die, reload", vs advisory [09:32]

stuart: what does TAG vs WG do? [09:33]

timbl: two philosophies: 1) stick things in a some stuff may become stds; 2) stick with using namespaces
... how to cleanly develop these things, probs in either case

<noah> I think CDF adds an extensibility story to XHTML, no?

<raman> interesting ... the only thing the tagsoup and xhtml folks would agree on is probably "we dont need RDF:-)" [09:34]

<DanC> maybe, noah, but if they [CDF] have, I haven't seen it.

timbl: TAG crosses many different groups, how do you take all these things and evolve them.

<Zakim> DanC, you wanted to think out loud about encouraging validator dev

dan: validator is close to TAG's work. [09:35]
... don't want html and css validators, want just a validator.
... shorten feedback loop between authors and validator developers [09:36]
... unicorn project..

<Zakim> noah, you wanted to talk about role of WG

dan: maybe we just need some detailed time on this..

noah: help them to fish rather than do the fishing.
... TAG should set forth principles, ie self-description
... TAG should help w3c get WG's going. [09:39]

<Zakim> Norm, you wanted to point out that I don't see how the validator helps by itself. I need to be able to read something that tells me the right answers,

norm: answers need to be written down.

raman: tag shouldn't maintain validator
... today people point their browser at something and see if it works [09:40]

<noah> I think rather than getting credibility by having a fraction of users go to our validators, it's far more effective to get the sort of credibility that will cause the browser vendors to implement the "right" answers.

raman: biggest success would be to have browser do validation.

<noah> That means the answers we advocate have to be practical.

<Zakim> timbl, you wanted to say that the browser could fix it in View Source

timbl: actually, I think that people when they do the, they also do view source
... whenever you do a save as or view source, it offers the cleaned up version

<noah> Suggest one would need to separate: view source from "view corrected source". I think there are a bunch of drawbacks to making it hard to see the "real" source when that's what you want.

<noah> Tim's answer of highlighting diffs is fine, as long as you can reconstruct the original too. [09:43]

<noah> I do view source :-)

timbl: imagine if you did view source and it colour coded. whenever you copied the source, you'll get the cleaned up

raman: people don't do view source anymore because it's too complicated

timbl: my son does it, and I think a lot of people still do this

<noah> My son too.

<Vincent> my sons too

timbl: and if it's generated, then it's even less excusable that it doesn't validate

dan: stick cases such as the forms and table nesting "documented hack" in the test suite

raman: people do open table, insert data, close table [09:45]

<Stuart> ack vincent [09:46]

timbl: how about forms have id tags, and always refer to forms via ids rather than the tags

vincent: many pages "near" him are out of his control and not valid.
... many of the content management tools, like blogs and wikis, etc. produce just garbage [09:47]

<Norm> If only we had either a carrot or a stick...

raman: chicken and egg
... some times content produces for old bugs and not necessary but sometimes the browser won't fix it. [09:48]

<Stuart> ack dave [09:49]

<DanC> DO: I hear TV pointing out problems with lots of suggestions, and wonder if TV has any suggestions going forward. [09:51]

raman: you run into the browser issue right away, then you run into legacy issue. [09:52]

<Stuart> ack DanC [09:53]

<Zakim> DanC, you wanted to observe that TV clearly has more details swapped in than I do, and to ask if he has suggestions, and to re-raise the "free to a good home" possibility

dave: was hoping to hear ideas for what we could do..

danc: thinks he hears TV say I don't hear the answer, but TAG should do more work. [09:54]

raman: architecturally the TAG needs to answer the question: should tagsoupintegration exist on the web, and then how does it coexist with structured markup.

<DanC> do, raman is re-iterating 3 options from HT's summary http://lists.w3.org/Archives/Public/www-tag/2006Oct/0062.html [09:55]

raman: henry said it well, tagsoup exists and come up with strategy for migrating toward structured markup 2) structured good; 3) tagsoup /structured co-exists
... lean towards co-existence and a strategy for moving towards structured [09:56]

timbl: there have been many extremes, step function for moving
... meet halfway on each one.
... for example people that want to omit quotes, and that's in xml. [09:58]
... we're looking at each thing separately, and looking at the benefits of moving in the direction we want
... question of which battles. eg, namespaces and microformats. [09:59]

<noah> I have mixed feelings about calling a a format in which attributes need not be quoted "XML", because the major selling point of the "XML Brand" is that the expectations for interop with deployed parsers are very high. Don't want to break that. I do think having a sort of "nearly XML" with known, stable transform to and from is a fine idea. I just don't want to quite brand it as XML.

timbl: namespaces very important

imbl: microformats can't cope with scale of evolution that we should have [10:00]

<noah> +1 The importance of namespaces for supporting self-description and as a basis for distributed innovation is the sort of principle the TAG itself should be driving.

timbl: the battle for namespaces is different level than quoted attributes

<DanC> TV: yes, (1) consistent tree, and (2) namespaces are my main things. quotes are in the noise. [10:01]

<Zakim> noah, you wanted to say success is bringing the browsers along

raman: problem is coupling these things, can't do namespaces unless you do xml and you can't do xml unless you have attributes

noah: there was a tone to the discussion that "oh yeah there's the browsers"
... I think it should be the other way around [10:02]
... success is fostering discussion that happens in the browsers


noah: path forward has to include browsers [10:03]

<Zakim> Norm, you wanted to ask if we really believe that fixing quotes on attributes would really help

norm: every time this issue comes up, quotes on attributes
... I wonder that this level of syntactic fixup is really a big part of this issue. [10:04]
... we could write 10,15 rules on this [10:05]

<DanC> (TV has given examples of how it's deeper than that, indeed.)

<raman> Agree with Norm that fixing quotes on attrs is in the noise; I blieve myself and timbl were saying the same

<noah> If Norm is right that the problems are really deeper browser dependencies, then I wonder how much mileage we get out of Tim's proposal for a self-correcting View Source? Seems like that would deal with the minor issues more successfully than the deep ones. [10:06]

<noah> I could be wrong about that.

<DanC> (the 10 or 15 rules evidently cost 1747 lines of python. 1/2 ;-)

ht: I don't think that the xml syntax issues aren't the big deal.
... I think it's more the unclosed tags, if anything

<noah> I suspect there's also a perceived convenience/DWIM issues. Not sure how big a deal that is.

ht: in the examples that hixsie gave, what happens when you pollute your xml with bad html and plug that into good xhtml
... spent some time on formal languages in that area.
... there is some work out there, but there was nothing that covered all the cases we came up with

<DanC> (http://code.google.com/p/html5lib/ )

ht: john cowan's approach was the last thing standing that might get to some consensus [10:08]
... if you cannot live with "here is a large collection of semi-formal english", then i want to go back and look at john's tagsoup

raman: at least it's in a ruleset for program use [10:09]

<DanC> (I started looking at cowan's tables in discussion with #whatwg folks, but only scratched the surface. I should update http://esw.w3.org/topic/HTMLAsSheAreSpoke )

raman: raggett's tidy used to have c code, which got migrated to java

danc: talk to the folks who worked on Access [10:10]
... maybe at f2f

noah: politics to who we invite.. [10:11]

(some discussion about individuals) [10:12]

stuart: any comm team in japan? [10:15]

<noah> Maybe the right thing would be for someone (Rhys) to approach Tommy (name?) informally and ask what would be the most effective way to pursue such contacts.

<noah> I would welcome some more formal contact with HTML 5. Not all of us take the trouble to participate directly with WHAT WG, etc.

<Rhys> The person from Access is Tommy Kamada who is Access CTO and their AC Rep [10:18]

<noah> I think it's appropriate to make the liaison with HTML 5 community at least more semi-formal, perhaps formal.

stuart: we should have some metings with the various groups.
... and how about the chartering process? [10:19]

ht: question is really, what is the architectural space the 1,2,* WGs are working in, and how can help [10:20]

<noah> I think the TAG can help in commenting on the W3C WG structure insofar as our technical insight contributes to an artful decision as to whether this work seems to admit working in separate groups or one, for example.

<ht> HST gives apologies for next week (sorry if that tips the balance)

elaborated infoset [10:21]

ht: I suggest adjourning this item for 2 weeks, as timbl sent regrets [10:22]

meeting adjourned

<noah> http://dev.w3.org/cvsweb/~checkout~/2002/scribe/scribedoc.htm [10:26]

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.127 (CVS log)
$Date: 2007/02/09 16:23:34 $