TAG Weekly

27 Aug 2009

See also: IRC log


Raman, John_Kemp, Jonathan_Rees, DanC, Masinter, noah, Ashok_Malhotra, TimBL
Noah Mendelsohn


<Zakim> DanC, you wanted to ask for a place in the agenda to talk about John's comments on HTML auth http://lists.w3.org/Archives/Public/public-html/2009Aug/1109.html


Meeting is starting. Tim is expected but not here yet.

noah: Sept 3, Henry scribe, Dan as backup

Approve minutes of prior meeting(s)

noah: table approval until someone has read the minutes

Administrative items (Brief)

masinter: might we talk about versioning today?

noah: I'd like to focus on prep for a productive F2F in September
... The HTML group believes they're very close to last call
... So the big theme for the coming few weeks, leading up to Sept F2F, is how can we maximize our (positive) impact on HTML
... Versioning is in that context, so if we don't get to it today then next week

masinter: Versioning has als come up in the device API working group

noah: Continuing agenda review... recent traffic on URI schemes, e.g. web sockets
... and we need to go over action items again

Brief review of summer progress

noah: sniffing, content-type, IRIs, web addresses, versioning, geopriv,

<DanC> ACTION-295 due next week

<trackbot> ACTION-295 Monitor geolocation response to IETF GEOPRIV comments on last call and report to the TAG due date now next week

noah: what is a resource (in http) (please give noah input on that one offline)

Discuss TAG goals through September F2F

noah: 3 main themes: html, web application architecture, metadata
... let's build on those as we move to sept f2f
... pay attention to html plan to go to last call

masinter: Noah, [question about last call]?

danc: Separating HTML and web app arch is tricky
... SVG has a lot of the same event & security considerations
... many normative dependencies from WebApps WG specs on HTML5 now e.g. Origin:


noah: note new html chairs

<Zakim> DanC, you wanted to note difficulty in separating HTML from web app architecture http://lists.w3.org/Archives/Public/public-html/2009Aug/0940.html

<DanC> I noted the WONTFIX response in http://lists.w3.org/Archives/Public/public-html/2009Aug/1204.html

<DanC> +1 defer URL

noah: HTML issue about "URL" was closed without action

masinter: The way you understand an IRI is you convert it to a URI and then parse the URI
... but HTML says that first you parse the IRI - this is a big technical change to the spec
... what I'm hoping for is to change the IRI spec so that HTML can directly reference it

noah: DanC, should we defer?

danc: I'm OK with the discussion we [just] had

noah: We'll schedule discussion again after Larry has updated the IRI draft

<masinter> due monday

<masinter> in HTML working group

<DanC> ACTION-298?

<trackbot> ACTION-298 -- Larry Masinter to notify the TAG of the next IRI draft -- due 2009-09-15 -- OPEN

<trackbot> http://www.w3.org/2001/tag/group/track/actions/298

<masinter> necessary for compatibility with IDNa requirements also

<masinter> want to defer until i have a concrete proposal rather than discussing it in abstract, please

<noah> ok

<noah> I'll take care of it

timbl: But they're used in so many places... [poor sound quality] ... very leery of changing anything

<masinter> also necessary to handle alternate encoding translation for query parameters

<masinter> action-298?

<trackbot> ACTION-298 -- Larry Masinter to notify the TAG of the next IRI draft -- due 2009-09-15 -- OPEN

<trackbot> http://www.w3.org/2001/tag/group/track/actions/298

noah: OK, we'll defer discussion.
... Ian has produced authoring spec drafts, as requested at Mandelieu

<masinter> the versioning goal also has a 'read' on distributed extensibility

<Zakim> johnk, you wanted to ask how this should really work - what kind of feedback is most important for TAG to give?

johnk: I have a massive list of notes on the html draft. What should I do with it - respond individually, or coordinate with rest of TAG?

noah: What progress is everyone making on reading the html draft?

ashok: I've started, was dismayed by presence of algorithms for parsing formats, etc. - will continue

<noah> ok larry

danc: I keep up with the html mailing list, but busy with other things & not reading drafts

noah: Have been reading draft. Have found many things to comment on

<raman> will need to split in 12 mins

noah: I encourage people to flag interesting issues, then go back and ask: Which of these relate to web architecture and/or other specs, that maybe the TAG might want to comment on?
... let's get this feedback in the next couple of weeks
... for the other (non-TAG) interesting issues, leave to discretion of individuals

raman: Last year I divided the issues into 5 or 6 categories. They remain

<DanC> (for reference, http://www.w3.org/2001/tag/2008/09/f2fkc-agenda#htmlAndWebBigPicture )

raman: Maybe these buckets might provide a way to organize

<DanC> 3.4.1 , 3.4.2, etc.

<masinter> ah ok

noah: Can we figure out if our versioning work says anything about possible changes to html draft?

<johnk> that's very useful - thanks!

<masinter> things have matured enough that new buckets are probably useful

danc: Where we are on the buckets [Raman's]:
... something about html and xml: We should watch the mime type registration issue. HTML5 says it replaces all previous registrations.

timbl: [poor audio quality] (about serving XML as text/html)

<masinter> there used to be (or was intended to be) an overlap between content that could be served as text/html and things that could be served as application/xhtml+xml such that they would have the same meaning, and this overlap is either not required or is empty

danc: You can still do what TimBL wants, but the spec isn't advocating it

noah: It explicitly talks about parsing XML and parsing HTML... that's part of the parser state & affects what it does

<raman> need to drop off ...

masinter: "You must serve it this way or that way" is the answer to the wrong question

noah: Tim raised the suspicion that if you served XML as text/html, maybe the client will do the wrong thing

masinter: Whether something is XML is the wrong question, since something can be both XML and HTML

<DanC> (so my opening comment, and still my summary, is that re the "3.4.1 HTML And Web The Big Picture" bucket, the/an issue to watch is ISSUE-53 mediatypereg http://www.w3.org/html/wg/tracker/issues/53 )

timbl: [audio quality[

danc: I don't think this would be a fruitful discussion (with htmlwg)... the spec can be read in both ways so everyone will get along
... worth watching the media type issue

<DanC> "3.4.2 Issue: HTML5 Should Be Modularized?"

danc: URL issue - text has been moved out but term "URL" remains
... web app arch - window object - origin - this all remains
... canvas spec

noah: the javascript URIs move out to the new spec

danc: So there has been modularization.

<DanC> "3.4.3 URL Parsing In HTML5"

masinter: The boundary has moved a little bit. There's a step going from the HTTP protocol stack to the HTML-consuming agent

<DanC> (http://www.w3.org/html/wg/tracker/issues/28 http-mime-override is closed; the subtlety that Larry is explaining isn't obvious in the WG decision.)

masinter: consumer is not required by protocol to perform any particular action

<Zakim> DanC, you wanted to offer to reflect on the KC buckets

<Zakim> johnk, you wanted to note my issue about 2617 interaction

johnk: Informative text in HTML seems to be in conflict with RFC 2617

<Zakim> noah, you wanted to say the compromise quoted by Larry seems dangerous for interop

noah: Allowing user agent to ignore the text/plain directive seems wrong - a threat to interoperability

masinter: Look at minutes from the Stockholm meeting

<noah> Can we publish poorly formed XML (I.e. first bytes appear to be XML) as text/plain and have it work. The described compromise implies "no", which is really bad.

<noah> So, maybe what I'm saying is: even if http-bis allows sniffing, having browsers in particular do it without prejudice is very bad.

masinter: The compromise was to say it's not the business of HTTP to forbid nonconforming interpretations of text/plain ... responsibility is moved elsewhere (e.g. to W3C)

danc: Getting this fixed is very hard... like changing the price of oil

<noah> I agree, but one needs to do the best possible to minimize damage.

timbl: [audio quality]

danc: This issue is closed in the HTTP working group

noah: Any hope?
... Maybe update the text/plain media type registration?

masinter: content type sniffing only happens in http, not in email

<DanC> (wonder who has the ball on http://www.w3.org/2001/tag/group/track/issues/24 contentTypeOverride-24 ...)

<DanC> ACTION-257?

<trackbot> ACTION-257 -- Noah Mendelsohn to invite Mark Not or Lisa D to revisit progress in IETF/HTML liaison on content sniffing -- due 2009-09-04 -- OPEN

<trackbot> http://www.w3.org/2001/tag/group/track/actions/257

<noah> Yeah, I've been very remiss on 257. May need some guidance on what makes sense at this point wrt Mark and Lisa

danc/masinter: I'd still like to talk with them

<DanC> "3.4.4 HTTP And HTML"

danc: ping attribute

<noah> Dan says there's a move not to implement and to pull it.

danc: this has open html issues
... htmlwg issue about embedding svg & mathml into html ... almost closed

noah: they're not examples of extensibility, just wired in

danc: Editor's position is centralized extensibility

<Zakim> DanC, you wanted to get back to John's comments on HTML auth http://lists.w3.org/Archives/Public/public-html/2009Aug/1109.html

<DanC> "Handling HTTP status 401 responses / User Agent Authentication Forms" http://www.w3.org/html/wg/tracker/issues/13

johnk: 401 unauthorized -> list of authentication mechanisms
... would like to see 2617 relaxed

<johnk> the issue is that HTTP server sends back 401 Unauthorized

scribe did not get what the 401 discussion was about. hoping for info from johnk

<johnk> and specifies one or more mechanisms

<DanC> (here's the proposal to close the urls-webarch issue http://lists.w3.org/Archives/Public/public-html/2009Aug/1068.html )

<johnk> and if the recipient understands NONE of the mechanisms suggested, it may display the entity-body in the response

<DanC> (it seems 1068 was sent before the decision process was refined to ping issue raisers)

<johnk> HTML has informative text that is somewhat conflicting

<masinter> depends on http://www.w3.org/html/wg/tracker/actions/137

winding up

noah: Do your actions
... Let's wait for Larry's next draft (5 days), then get 2 reviewers


Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.133 (CVS log)
$Date: 2009/09/11 13:02:40 $