HTML WG f2f -- 03 Nov 2011

Intro

paulc: will be doing things un-conference style
... I propose we meet 9 to 5 both days

[discussing room logistics]

paulc: MikeSmith and others will need to be at the AC meeting from 10:45 to 11:15
... propose we have the joint meeting with the Web and TV IG after lunch today
... several people have traveled to TPAC for the Web and TV meeting
... we will plan 1 hour for that session

<pimpbot> changes: sam: Correct link to issue 174 change proposal <11http://lists.w3.org/Archives/Public/public-html-diffs/2011Nov/0016.html>

paulc: we are not planning to have a phone bridge
... can skype people in

[doing self-intros around the room]

the <time> element

tantek: let's get started
... recently this was drop'd from the space
... spec
... Potential to add back a more powerful time element
... summary add back + more

<myakura> http://www.w3.org/wiki/User:Tantekelik/time_element

<pimpbot> Title: User:Tantekelik/time element - W3C Wiki (at www.w3.org)

tantek: for year dates, month days...etc...
... v-card4 expanded to include these uses
... addition to these obvious ones..two more exists

time zone and duration

scribe: lots' of useage...
... apps can use this to deal this the lack of timezone context

<pimpbot> Title: HTML WG f2f -- 03 Nov 2011 (at www.w3.org)

tanket: Ask group - any use cases missed?

mjs: can we see use cases of each?

tanket: lets' take year only...
... YYYY

tantek: wikipedia and copyright uses a 'year only', e.g. 1999
... year month is used alot on bloging...

Time element doesn't match HTML5 input types, which is a mismatch

scribe: another example is credit card dates

tantek: Month Day - e.g 12-25 (x-mas) or a birthday
... duration: use case is it's not possible to differential from a time
... for example 2:30 is this 2:30pm or a movie that lasts two and a half hours
... it's implemented by a few others and it seems to work...
... since people are using this...
... generally do we want more specific or more generic elements?
... div is very generic and could be a span or header, group, etc..

though we have gone down the more specific and added many more specific elements that can be created by a div

anne: not clear why we need this..

tantek: people incorrectly parse times and get the wrong information

james: you can still get this wrong...
... for example a birthday on a page using a time element, could be considered a publishing date

<karl> performance on querying the DOM? parsing all data elements instead of only time?

tanket: this is not true...

james: the time element is not going to solve this use case

mjs: I'd like to jump in and comment

tantek: the search engine example is a little vague

mjs: another use case is to search for a date, what happend on 11/3/2011?
... it's possible that search engines could use a time element for this use case
... not sure if this is a primary use case but, it is a use case

tantek: I am drawing from example of real use cases for search engines

james: we don't have a location element for example...why not?

<pimpbot> planet: Tantek: #HTML5 time element discussion <time>11:30</time> <time datetime=2011-11-03>today</time>: http://www.w3.org/html/wg/wiki/TPAC2011#Day_1 <11http://tantek.com/2011/307/t5/html5-time-element-discussion-today>

tantek: we do in geolocation and I encourage you to take a peek at these WGs

james: we're not adding a location element though...

<pimpbot> Title: HTML WG f2f -- 03 Nov 2011 (at www.w3.org)

tantek: we may...

mjs: we do have evidence that search engines are doing this heuroistically

<karl> we born with a date, a location and a name

mjs: which they get wrong, so by adding it s search engines
... calrification on james...

<pimpbot> bugmail: [Bug 14526] WF2: When adding filenames to the data set, should there be normalization of decomposed forms? <11http://lists.w3.org/Archives/Public/public-html-bugzilla/2011Nov/0245.html>

mjs: question...

james: I'm claiming that time is less useful than something claiming it's a heading

alex: how do we add new elements into html that has good use case for searching..
... microformats has prove the use and if html doesn't add new elements it won't have a future

tantek: We have highly generalized neeeds and data shows that it's has a need
... for the time element
... we not asking to add new elements for tags that are highly specialized (volcano)
... we are not adding alot of elements to html, so it's not a slippery slope

mjs: asking for vote to not having time element vs having time element

1 vote to not have it...larger number of people (7) would like to have time element

mjs: asking for data element vs not having data element
... it seems that it's clear (not full working group) favor having both elements <time> <data>

hixie: most of these use case are irrelevant
... you don't need markup for this..
... all of this can be trivial found since they are ISO standards and are easily parsed

<karl> hixie: you don't need markup for publishing data, you need markup for consuming data

hixie: search engines do this today pretty well...except year and length
... the ambiguity exists for other items...
... for example SF - all search engines know this is san fran

tantek: we can use what people publish to determine what we need
... time is being used and the pattern is that others will

hixie: they build it and it will come is not a good strategy

tantek: I disagree it's worked well for Microdata

<tcelik> krisk s/microdata/microformats

hixie: longdesc is a good example gone wrong

tantek: why make a special case for year? If lots of people are using this then we should promote this to an element
... If people are intrested in time, then I'd like to hear their specific needs

kimberly: TV has a use for duration

mjs: some of these particular specified date don't enable a date() object to be created
... we have a problem since an api exists to provide a date() object from this tag

tantek: It's a perfectly reasonable request to expect the api to do the right thing
... doesn't seem like a difficult problem and should be easy to reach consensus

mjs: any other questions?

alex: what about pre-gregoria and gregoria?

tantek: I have not seen data to justify adding this support

alex: I have seen this in history examples

<anne> http://en.wikipedia.org/wiki/Proleptic_Gregorian_calendar

tantek: a workaround exists...

<anne> ^^ article on workaround

tantek: it we have more data then we can add this support
... if you see it we can evaluate it as we gather more data

mjs: any other comments or time in general?

group agrees no more issues, meeting is over and we'll meet again after lunch at 1:30

<pimpbot> Title: HTML WG f2f -- 03 Nov 2011 (at www.w3.org)

<karl> agenda: http://www.w3.org/html/wg/wiki/TPAC2011

<pimpbot> Title: TPAC2011 - HTML WG Wiki (at www.w3.org)

<pimpbot> Title: HTML WG f2f -- 03 Nov 2011 (at www.w3.org)

<tcelik> FYI re: non-Gregorian proposal/research http://wiki.whatwg.org/wiki/Time_element#Calendar_scale

<pimpbot> Title: Time element - WHATWG Wiki (at wiki.whatwg.org)

<karl> s/Agenda: http://www.w3.org/html/wg/wiki/TPAC_2011_Agenda#Thursday.2C_3_November_2011/Agenda: http://www.w3.org/html/wg/wiki/TPAC2011/

<pimpbot> Title: TPAC 2011 Agenda - HTML WG Wiki (at www.w3.org)

<pimpbot> bugmail: [Bug 14689] New: xml-stylesheet with type=text/xsl needs to be handled explicitly <11http://lists.w3.org/Archives/Public/public-html-bugzilla/2011Nov/0246.html>

<pimpbot> bugmail: [Bug 14688] fhfhgfhgfhgfhgfhgfhgfhgfhg giuhgiuhiuhi hbkhikhiuhbi ihbihbjhbjhb <11http://lists.w3.org/Archives/Public/public-html-bugzilla/2011Nov/0247.html>

<tcelik> greetings

<tcelik> rrsagent tcelik is tantek

Web & TV

<anne> scribe: tcelik

W3C Media Pipeline TF Requirements

presentation

November 1-4, 2011

Clarke: two task forces

media task force

home networking task force

media *pipeline task force

… scope of the MPTF is to support commercial grade video in browsers or anything that supports HTML

… set out to develop requirements in that area

… then identified gaps

… good news is found very few gaps

… most use cases adequately supported in HTML or in current specification

… we have some additional areas with ideas for addressing things, not yet consensus

"R1. Combined main + description audio track"

R1 = requirement one

… each of these slides have use case, what doesn't work, bugs filed, sections in spec, suggested changes

Clarke: first use-case

… in US/Canada different from Europe - playing audio tracks

… US/Canada premixed

… other countries separate

… we need to identify premix

… right now HTML5 appears to only address non-prefix

… Suggestions:

… define new category values

<MikeSmith> bug 13357?

… main+description

<pimpbot> 11http://www.w3.org/Bugs/Public/show_bug.cgi?id=13357 b.lund, P2, NEW, 13<video>: Additional AudioTrack.kind categories are needed to identify tracks where audio descriptions are premixed with main dialogue.

… translation+description

Mark Vickers, Comcast (MV)

MV: you could make the category a list

… such as if you had main video and assigned video

Paul Cotton (PC) for the record this bug is still marked as new

… this = bug 13357

MV: we also reviewed this with the a11y P&F group on Tuesday

… there was some discussion about that with some people that were more expert on this

Janina: we agree that this is information that needs to be avail to the consumer

… how we achieve it is matter of discussion

John Foliot (JF)

… the use case of supporting text descriptions is very real

… we would support additional kinds

Janina: you're going to want to know for historical stuff if it is premixed

… can't unmix it

Frank Olivier: do you think we need these 9 categories only, or should we leave it open ended

… so people can create their own categories

Clarke: we should ask the a11y group

JF: when we looked at the values originally, these slipped past us. In theory I support multiple / extensible kinds

<anne> what is wrong with kind=alternative + label=... ?

… what if we do if down the road we discover a new kind

MV: There's an example on the slides I wondered about.

… there's a type for main video

… and signed video

… could be delivered separately or together as well.

Russell: I was going to suggest ...

Anne: what about using just "alternate/alternative" and a label?

Janina: you need to distinguish two situations

<dbaron> participants in Web and TV IG: https://www.w3.org/2000/09/dbwg/details?group=46300

… a) audio+video premixed, can't undo that

<dbaron> Participants in HTML WG: http://www.w3.org/2000/09/dbwg/details?group=40318&public=1

<pimpbot> Title: Participants in the HTML Working Group - DBWG, the Working Groups Database (at www.w3.org)

… b) European practice, main audio, and then describe-audio in separate channel

… as long as it is clear to user if it is premixed or not

Anne: premix just seems like just an alternate

Maciej: you still need the device that's doing the playback to know the nature of the atlernative

<Kai> help

… if I said I want audio descriptions for everything, then when the format is in a premix format, UA needs to detect that so that it doesn't mix them up

… there might be multiple combined resources in the same track, should be able to express that

JF: Anne, we don't have a labeling mechanism. We only have kind.

<Hixie> when people say @kind, what are they referring to?

<anne> Hixie: http://www.whatwg.org/specs/web-apps/current-work/#dom-TrackList-getKind-categories

<pimpbot> Title: HTML Standard (at www.whatwg.org)

Anne: the preselection mechanism makes sense.

Mark Watson (MW) Netflix

… the labels are not useful because you can't automatically intepret them, i18n etc.

<MikeSmith> bug 12544?

<pimpbot> 11http://www.w3.org/Bugs/Public/show_bug.cgi?id=12544 jfoliot, P3, NEW, 13<video> MEDIA CONTROLLER requires track kind for in-band tracks

… captions and subtitles

… not available in other form, want to be able to signal

<Hixie> anne: well that's confusing. That's an IDL attribute, not a content attribute.

… when those are available

Janina: Anne, you need to give us 2 examples

… how is it marked when the described audio is burned in, and separate track

… one example is not enough

… need to have two to understand how it works

Anne: I was already convinced earlier by Maciej

<pimpbot> bugmail: [Bug 13986] Update vCard vocabulary to RFC6350 <11http://lists.w3.org/Archives/Public/public-html-bugzilla/2011Nov/0248.html>

Janina: this is the uncontroversial one

PC: as chair, question I ask is, the bug does not include a complete change proposal

… is there anybody in the room that are going to carry this forward and write a change proposal

… I want to know how this is going to work

<anne> Hixie: hmm yeah

… editor handles the bug?

JF: This is obviously supporting an a11y requirement, we could write a change proposal. just requesting 2 values at this time.
... we could explore extensibility later

Sam Ruby: Editor processes it by default

PC: Let's follow process then, have the editor process it

<anne> Hixie: I guess http://www.whatwg.org/specs/web-apps/current-work/#attr-track-kind is the one

<MikeSmith> Hixie, are there any new UA conformance requirements needed for this? any browser implementation needed?

<pimpbot> Title: HTML Standard (at www.whatwg.org)

… I encourage those who have specific proposal to add it to the bug

<anne> Hixie: but that doesn't make sense

Bob from Cable Labs: what's being proposed is two new kind categories

… main+description

… translation+description

… do we need more?

<pimpbot> Title: HTML WG f2f -- 03 Nov 2011 (at www.w3.org)

JF: there may be in the future. at this point these 2 have emerged. a year from now?

… 2 ways to approach. 1. let's propose these for the spec now, 2. is there an extensibility issue?

Bob: Let's move forward with these two kinds

… and parallel with a broader solution

<anne> Hixie: I guess the streams would expose those kind of values

JF: if you want to action the a11y tf go ahead

Giuseppe from Opera

… kind can be main+description, plus something else

JF: that goes back to question of extensible

… sounds like we have extensibility issue here on the table

… it's a worth while discussion

Giuseppe: we need specific requests because we are not familiar

… with the process in this group

PC: because bugs are first processed by the editor
... 1. does the bug reflect the discussion here?

… if it doesn't, add that info to the bug

… because *that* is what will be taken into consideration

Hixie: are we talking about the 'kind' attribute on … ?

JF: The object?

Hixie: these are very different with different behaviors

Anne: it's about the object

Hixie: happy to add to the object

… only limitation is the formats the UA supports

… so far I added everything Ogg supports

… happy to add what H264 supports

… need documentation on that though first

… please add a link to the relevant spec in the bug and I'll add it.

MW: that's a backwards way of thinking about it

… notion of those not have to do with media containers

… backwards to wait for the media container people

… and then copy/paste those into the HTML

… we don't want to go overboard

… where there examples that are clearly in use, even not on web

… we should add it for the web.

Janina: are you Ian waiting for a user agent on the web to support it.

… we know what we want because we have 30 years of history on this in broadcast and broader media distribution

… we can be fairly confident there aren't going to be different things

… question is how to mark when they've been combined into a single track, and how to mark when they're separate tracks

Clarke: it's clear we need to have bug processed by the editor, and we all add any comments to the bug

… next issue

" R3. Handling of In-band Tracks"

<MikeSmith> bug 13358?

<pimpbot> 11http://www.w3.org/Bugs/Public/show_bug.cgi?id=13358 b.lund, P2, NEW, 13<video> also fire a 'change' event at VideoTrackList, AudioTrackList, and TextTrackList objects when their list of tracks changes

<MikeSmith> bug 13359?

<pimpbot> 11http://www.w3.org/Bugs/Public/show_bug.cgi?id=13359 b.lund, P2, NEW, 13A way is needed to identify the type of data in a track element

<MikeSmith> bug 14492?

<pimpbot> 11http://www.w3.org/Bugs/Public/show_bug.cgi?id=14492 jan.lindquist, P2, NEW, 13<video> change event when tracks are removed

… use case: Playing in-band multiplexed media streams (e.g. broadcast tv, live events and recorded movies) with track elements that come and go over time (e.g. secondary audio, subtitles, in different languages, application signaling and content ratings>)

… what doesn't work: application doesn't know type of data tracks or when tracks end.

… submitted bugs: [x] LC1 Bug 13358

…. [ ] LC1 Bug 13359

… [ ] Bug 14492

… sections: 4.8.10.12.2 sourcing in-band text tracks

… suggested changes:

… * mapping of in-band tracks nedds to be done in a standard way

scribe:

FO: It would be good to do enumeration of all the tracks in the file and what their start/end times are

Hixie: the problem is that you don't know if/when they're going to start or end if ever

<JF> s/nedds to be done/needs to be done

FO: but if it is not a live stream...

Clarke: but it is (a live stream)

Hixie: there appear to be two issues here
... 1. multiple inband tracks video and audio tracks that come and go

… I'm trying to better understand why are there so many tracks?

Clarke: multiple secondary audio tracks, multilingual, application signalling

… one way of dealing for example, something similar to a current video channel, you choose to encode it as a continuous stream, every time you get an ad you get new tracks

Hixie: if you go from tv to ad to tv, presumably you have one video, one audio stream, one alternate for the other

Bob: one example is...

… or go from one program to another, and the translation tracks change

… text tracks that come and go

… timed text tracks are useful for interactive TV signaling

… in general over a long period like hours, you would expect to see audio tracks show up and go away

Hixie: for metadata text tracks that makes more sense

Eric from Apple

… when you change format, or change characteristics of the audio, typically done with different tracks. e.g. main program to ad

… diff size or scaled, compressed with different format

Hixie: each ad could be in a new track?

Eric: yeah

Hixie: makes sense. I'll fix that.

Hixie. 2. the other one was the text track. don't fully understand. you want the text tracks to fully say what they are?

Bob: TV applications, e.g. interactive TV app, ad insertion opportunity. if those are inband tracks, no way to identify them as metadata tracks. no way to identify to scripts that one contains one format and another contains another format.

… one way would be identifying them with a mimetype

… or other ways

… idea, expose program description

… javascript would have enough to look at the tracks and figure out which are which

Hixie: can you reuse label?

Bob: that's non-standard

Hixie: we could make it

Bob: that would be the kind

Hixie: my concern is that this usecase seems specific to one domain, very concerned about adding a feature just for this one domain

Bob: not sure if domain specific

Hixie: if you have a normal web page, it has some out of band text tracks that are kind equals metadata

… you're doing it from script

… so you don't need it

Bob: this is an inband track problem
... they're not limited to...

Hixie: that track doesn't have label right?

Bob: depends on media format

… Ogg or Webm has an attribute where you can specify a human readable string for the channel which would logically go into the label

… some tracks have labels, some done

Hixie: I'll have to look at that more closely. That doesn't help me understand the problem.

<pimpbot> Title: HTML WG f2f -- 03 Nov 2011 (at www.w3.org)

PC: Hixie, any other questions?

Hixie: two issue I asked about are enough

MV: it would be useful to thru the mapping spec

Hixie: A link to that would be very useful in the bug

MV: this is the mapping spec

"Mapping from MPEG-2 Transport to HTML5"

<pimpbot> changes: hixie: Attempt to update to the latest vCard spec (whatwg r6816) <11http://lists.w3.org/Archives/Public/public-html-diffs/2011Nov/0018.html>

Hixie: I don't think I've seen that.

Bob: it's not on a wiki

PC: can we post it now

Hixie: you can attach it to the bug

… that would be great

Clarke: are there any opinions about whether this mapping specification you want to refer to separately or W3C would want to maintain?

Adrian of Microsoft: short answer yes

… I suggest sending it to the list

… maybe posting it somewhere and sending a pointer

… have people on the mailing list discuss it

Clarke: ok let's move to the next one

"R7. Additional Media Paramters"

… this is the more controversial one potentially

<MikeSmith> issue-179

… hopefully less controversial with how we're approaching it

<MikeSmith> issue-179?

<trackbot> ISSUE-179 -- {audio,video} require param child (or equivalent) -- open

<trackbot> http://www.w3.org/html/wg/tracker/issues/179

<pimpbot> Title: ISSUE-179: {audio,video} require param child (or equivalent) - HTML Weekly Tracker (at www.w3.org)

… Use case: Playing adaptive rate video via video element.

<MikeSmith> bug 13333?

<pimpbot> 11http://www.w3.org/Bugs/Public/show_bug.cgi?id=13333 glenn, P2, RESOLVED WONTFIX, 13audio, video (and source) elements require param children or equivalent

… Currently deployed object element adaptive rate video players allow application control of adaptive play-out.

<MikeSmith> bug 13625?

<pimpbot> 11http://www.w3.org/Bugs/Public/show_bug.cgi?id=13625 b.lund, P2, NEW, 13There is no way to pass audio and video content metadata to the user agent that is required in some cases for playback.

… Common parameters or other media should also be considered.

… what doesn't work: HTML5 spec has no API to control adaptive video

Clarke: one idea suggested in #htmlav club meeting yesterday was need to come up with specific parameters to be passed

… we don't think there is a need for new functionality to HTML, but perhaps additional errors for more specificity, or other events

… for communication to media elements

MV: on this and the next few bugs, issue 179 is mentioned

… 179 proposed a general parm mechanism

… discussion said to propose specific parameters

… so next few are proposing specific parameters

… but no one is against a generic param mechanism

Anne: could you clarify what you mean by passing data to the media elements.

… adaptive streaming sounds more like a protocol issue more than what media element should handle

Clarke: there are certain parameters that would be useful to media general problems

… general idea of passing params is helpful e.g. provide an experience where people watching get best experience, to know the bandwidth

Anne: where do you pass params to, who gets them, uses them

scribe:

Anne: I've heard two proposals so far

… 1 is to let protocol solve it

… UA implements protocol, server does too, they negotiate quality/bitrate

… 2 is where page has open with server, and gets passed bytes over that connection where they implement the protocol at the application level and pass blobs to the video element that would allow this sort of thing

… at this stage it is inactionable

… need more concrete proposal

<pimpbot> bugmail: [Bug 13359] A way is needed to identify the type of data in a track element <11http://lists.w3.org/Archives/Public/public-html-bugzilla/2011Nov/0253.html> 4** [Bug 14690] New: <a> should be allowed as a child of <dl> Current browsers allow this already. A semantic use case is file downlads <dl><a href="filename1.pdf"><dt>Filename</dt><dd>description of file</dd><dd>filetype</dd></a>etc...</dl>. <11http://lists.w3.org/Archives/Public/public-html

Adrian: What Mark said was, the idea of that request everybody now understands is a good one. There's some work to do to figure out that concrete suggestion.

… should happen somewhere, think it should happen in HTML Working Group

PC: current status of issue 179 that asks for the generic PARAM parameter

… we have a counter change proposal to not do that

PC: we haven't had the original change proposal withdrawn

… still time for counter proposals

PC: I want to know what are we going to know with 179.

Adrian: I'll withdraw my overgeneralization of everybody.

… maybe almost everybody agrees now

… I suspect that at end of issue 179 there will be consensus for not making a change

… because we have a better understanding how to move forward.

MV: re 179, I spoke with Glenn Adams. he is still interested in pursuing that. he is not able to be here.
... the interest group does not have unified support for that bug

… we are coming in with Sylvia alternative proposal

… to discuss specific parameters

… the whatwg link in this slide goes to a survey/analysis of adaptive parameters across a wide variety of systems

PC: I believe we have 2 change proposals.

… we haven't yet called for counters.

… normally that runs 30 days

… ended on Oct 28

<pimpbot> Title: Error (at wiki.whatwg.org)

… next automatic step would be to evaluate whether we call for counter proposals

… would be good to hear from as many WG members where they stand on this.

(silence)

Hixie: When Henry rejected the bug, he rejected the generic concept

… specific issues would be considered.

PC: new related bugs would be treated as subtypes of this bug and treated as last call

… e.g. we have cases where we as we refine the question in a bug, the chairs are willing to enter into a consideration that subtype bugs are then treated as ontime last call bugs.

… reach out to us and ask if you have any questions about that and have a dialog with us.

Clarke: R7 was about passing params

… R8 is about getting feedback

… R10 and R11 we consider parallel issues

… specifically related to content protection

… "R111. Content Protection Feedback and Errors"

… we the people who know the specific events, errors, params need to be passed and come to a consensus of a subset of these that we would propose as changes.

MW: currently the errors on the media element are rather limited

… I'd like to learn more about the rational for that

… reason for more error codes where we have customer service reps

… trying to fix things

…. we want web page be able to provide more specific errors

… to send back

Hixie: reason limited is baby steps. reason object instead of number is to be able to add things later

… should be slow to add things because they we get too far ahead of the implementations

<arnpro> what do you think about Dart? (google's new)

… errors have some CORS aspects

?? need 10

Hixie: that's a lot, please prioritize

Clarke: if there is advice for how to get this into HTML quickly that would be great.

Clarke on R7

… exposing information on available bit rates, maybe set maximum

<pimpbot> Title: HTML WG f2f -- 03 Nov 2011 (at www.w3.org)

… exposing and setting parameters for fragment selection

… playlists want to prefetch segments to play things seamlessly

… in R8

… common media error

… additional events/info

… DNS failure, TCP failure, TLS failure

… stats such as packet drop rate

… changes in the stream of events

MV: this is where the referenced WHatWG wiki is good to look at

Clarke: R150

R10

… getting DRM system information

… exchange DRM related messages

… ref: OIPF DAE specification

… R11. Content Protection Feedback and Errors

Tantek: are the slides in the respective bugs?

Clarke no but they could be

ACTION Clarke add contents from the slides of the presentation to each specific bug

<trackbot> Sorry, couldn't find user - Clarke

MW: Different conclusions though, might be components exposed through the web interface, pass through for parameters is different than webarch that hides platform/device specific things

… re: content protection

… some devices have content protection, some dont'

… my question to the group is: is it acceptable to introduce this capability to the web platform (pass thru), or some other strategy?

Anne: the last time this was discussed on WHATWG list

… least controversial suggestion was API feed bits into the video stream

… DRM would be implemented at the application level

MW: you can't implement the DRM in the JS

… needs to be implemented in trusted execution environment

<karl> there is nothing secure in life. useless pun.

MW: objective of DRM is not to make it impossible but difficult, for some well defined values of difficult

… need these capabilities in the trusted computing layers

… below the video element

Anne: we are already at the place where you have H264 decoding in JS

… I was just stating the last thing that was discussed on this topic that had some traction beyond we don't want to go there.

John Simmons of Microsoft: Echoing what Mark said.

… key takeaway

… DRM systems are going to be used on video delivered to devices

… people who develop DRM understand those constraints

… e.g. can't implement in JS

… key question is whether HTML will be able to accomodate DRM protected content

… or continue like silos today

… which prevent large scale growth of internet television that we'd like to see

MV: related to 179

… there are these two proposals

… 1. for parm

… 2. not use it

… data- and x- are supposed to be prohibited from use

… but no change proposal says use those

… didn't make sense

… details are in the issue 179 two alternate proposals

… can anyone comment?

PC: project change proposal?
... original change proposal

MV: just above the blue box

… furthermore the data-* mechanism etc.

… can't be used

… but then in Sylvia's counter proposals says to use them.

… does the quoted text need clarification

… or is it just a misunderstanding?

Hixie: the data-* attributes are intended to only be used by scripts in the page, not by the user agent

… if you have something the user agents would use, then either we would add it to the language

… or if it was experimental we would use x-* syntax

… e.g. webkit has x-webkit-*

Adrien: the misunderstanding was how x- attribute might be used on the way to standardization

… e.g. like it's done with vendor prefixing in CSS.

… in order that those experimental implementations don't drive actual content, we use a vendor prefix

… the proposed standard doesn't have a vendor prefix

… the vendor prefixes are a temporary measure to experiment while standard is being agreed.

MV: concrete case is UPnP

… available spec

… but no support for it

… slightly different model UPnP video vs. regulary HTML video access

… a separate group UPnP could publish use x-upnp-* params, and encourage others to

Hixie: if they're writing an actual standard they wouldn't use x-

… just upnp-

MV: so ok for them to just use upnp- ?

Hixie: I wouldn't recommend it, would prefer to come to the group

MV: I agree primary idea is bring these in

bug 13625 (13635?) also tracks separate attribute

PC: I want to point people to the conformance clause in HTML5
... It says you can conform by pointing to HTML5 and other set of additional specifications

… a vertical industry could say you conform to HTML5 and this other specification

… too many people believe you want to force everything into the HTML5 spec.

… I want to support Ian's view that if something is generic we want that discussion and get that in

… or at least know about it

… so that we can detect when something becomes emergently generic

Hixie: I agree

… the conformance section really it says what's been true of any spec

MW: I wouldn't encourage people to write extensions. we have an interest in strongly discouraging separate extension. we don't want fragmentation.

… if you look at OIPF and HDTV you have 100s of 1000s of pages of stuff

… a uniform platform, that should be our goal

Hixie: in practice as people do start writing specs that start extending HTML in that way, it probably means we failed, but in reality it might happen anyway

PC: time check

… have gone through all the material

… several suggestions made

… including Tantek's that all this slide material be added to the associated bugs

<pimpbot> bugmail: [Bug 13357] <video>: Additional AudioTrack.kind categories are needed to identify tracks where audio descriptions are premixed with main dialogue. <11http://lists.w3.org/Archives/Public/public-html-bugzilla/2011Nov/0262.html> 4** [Bug 13239] Add support for in-page dialogs <11http://lists.w3.org/Archives/Public/public-html-bugzilla/2011Nov/0261.html> 4** [Bug 12409] WF3: Automatic capitalization in input fields <11http://lists.w3.org/Archi

… have we covered everything that the interest group wanted to cover?

… process discussion can be taken offline over coffee

… chairs are open to taking that stuff orally or over email

MV: two things

… 1. one more residue, that we're not bringing. as FYI: need for a service discovery API

… to allow discover of things thru like zeroconf and UPnP

… being discussed at DAP group

… we didn't bring it to this group because it is a whole new thingl

… 2. we studied dozens and dozens of use-cases

… this is the residue that came out

… testament to the spec, many applications worked just fine with the spec

…. other than these few areas

… it's all right on track

… a lot of verification

… there was a huge amount of positive work that we're not showing you because it all worked.

PC: any last comments?

… thanks to the chairs of the IG for approaching the chairs of the HTML WG so early

… gave us a chance to coordinate chairs

… WG meeting is now recessed until 15:30-0700

<anne> wait

<anne> is the value "coffee/"

<anne> or "coffee"

anne - ask the parser ;)

<pimpbot> Title: HTML WG f2f -- 03 Nov 2011 (at www.w3.org)

<pimpbot> bugmail: [Bug 13359] A way is needed to identify the type of data in a track element <11http://lists.w3.org/Archives/Public/public-html-bugzilla/2011Nov/0264.html> 4** [Bug 13550] type table in 4.10.7 should indicate control type changes based on list attribute <11http://lists.w3.org/Archives/Public/public-html-bugzilla/2011Nov/0263.html> 4** [Bug 14691] <a> should be allowed as a child of <dl> Current browsers allow this already. A semantic use cas

<hober> ScribeNick: hober

rubys: this is the last session of the day

HTML Decision Policy report

paulc: these slides are from the plenary day session
... I'll go through the slides and replay some of the Q&A that we had yesterday and this morning in the AC meeting
... we have bugs & issues; editor responsible for bugs, escalation procedure for others that produces issues
... chairs find consensus through the path described in the decision policy
... v2 of the policy endorsed by the wg this summer
... chairs look at the reactions that change proposals get
... [description of Enhanced Change Control]
... e.g. the <time> element discussion earlier, the chairs have requested a revert; issue will be turned into multiple proposed changes
... [paul displays & describes diagram of the decision process algorithm]
... feedback from ac meeting: decision policy isn't approachable for new people
... chairs have informally discussed, working up a simple faq would help

<pimpbot> Title: HTML WG f2f -- 03 Nov 2011 (at www.w3.org)

paulc: the process is detailed because the chairs don't want there to be any confusion
... issues can sometimes be reopened
... when writing up a decision the chairs describe the sorts of triggers that could get the issue reopened
... interested in new information that wasn't part of the original discussion
... Formal Objections are the escape valve of the w3c process
... we have a webpage that lists our formal objections thus far
... e.g. the polyglot document claims to be normative, there is an FO on this
... received a comment that, given the importance of the decision policy, changing it is akin to rechartering
... there is a bugilla component on the policy; we welcome bugs
... maciej processed as many bugs as possible between v1 and v2 of the policy
... decision policy isn't an edict from the chairs; we welcome feedback on it
... another comment: some find the wg to be a hostile place, and that the way editors process bugs is unusual
... the chairs work very hard to make this a level playing field in which people feel that they can present arguments
... one person complained that they couldn't unsubscribe from the mailing list
... dsinger & others asked how is someone supposed to be able to watch all of the many changes
... paul stepped through the <time> element changes, showing the granularity of spec change; the editors are really good about matching changes to specific bugs
... somewhat difficult to see what changed in the w3c document because the svn commits are on the whatwg side

<pimpbot> bugmail: [Bug 13570] why does input type=color support autocomplete? <11http://lists.w3.org/Archives/Public/public-html-bugzilla/2011Nov/0271.html> 4** [Bug 13550] type table in 4.10.7 should indicate control type changes based on list attribute <11http://lists.w3.org/Archives/Public/public-html-bugzilla/2011Nov/0270.html> 4** [Bug 13558] input type=email should support friendly names <11http://lists.w3.org/Archives/Public/public-html-bugzilla/2011

paulc: not many contentious changes slip through
... [discusses <time> element as an example]
... had questions over coffee about the relationship with the whatwg, about the wg size, etc

test suite status report

The slides: http://www.w3.org/2011/Talks/TPAC/HTMLWG-DP/

<pimpbot> Title: HTML WG Decision Policy (at www.w3.org)

krisk: lots of growth over last year in terms of submitted tests
... we're pretty far ahead relative to other groups
... [runs through slide describing the various sorts of tests in the suite already]

<paulc> Decision Policy slides: http://www.w3.org/2011/Talks/TPAC/HTMLWG-DP/#%281%29

<pimpbot> Title: HTML WG Decision Policy (at www.w3.org)

krisk: over 106K submitted tests
... participation in tf is up since last tpac
... public-html-testsuite@w3.org is the list; much less traffic than the main list
... please send jgraham feedback on the test harness
... [thanks to contributors]
... tf meets biweekly on irc

Change Proposal status

stevef: there are a number of issues that are waiting for the chairs to evaluate the CPs
... when will the evaulations occur?

janina: at least one issue from pf

<anne> http://intertwingly.net/tmp/wgstatus.html

paulc: [describes how issue status page is organized]

<pimpbot> Title: HTML WG Status (at intertwingly.net)

<anne> ^^ displayed on screen

rubys: we have reviewed one of the CPs of ISSUE-30, there are two others, one is being reviewed

paulc: chairs started to give themselves action items for this review in the last 3 or 4 weeks
... all 3 of the chairs have been very busy
... we have to start to move to make the deadline dates in our timeline

rubys: we're prioritizing ISSUE-30

paulc: we've moved into a mode of reviewing CPs so that the resultant surveys will be better

rubys: e.g. the issue-30 cp we reviewed is much improved after incorporating feedback

janina: pf wants to highlight issue-133

stevef: there should only be one CP for issue-133

paulc: there are two

stevef: one was withdrawn

rubys: we'll issue a call for consensus

<tantek> <aside> Hixie, re: search engines know SF is San Francisco, but they don't however know MV is Mountain View </aside>

paulc: we could issue the CfC tonight
... & maybe take it off of the discussion list tomorrow

stevef: is there an overall timeline for these issues?

paulc: no, but we want to get through them as fast as we can
... ~250 open bugs to be processed by 12/31
... [overview of early 2012 schedule]
... chairs are working to clear the backlog before the end of the year

Schedule status (AC report)

adrianb: there were 2 bugs opened, one against microdata and one against rdfa
... based on TAG feedback that resulted in the html data tf being formed
... is there an expectation that that tf will create feedback for the wg
... if so, when will that happen

<adrianba> TF -> http://www.w3.org/wiki/Html-data-tf

<pimpbot> Title: Html-data-tf - W3C Wiki (at www.w3.org)

<adrianba> bug 13100?

<pimpbot> 11http://www.w3.org/Bugs/Public/show_bug.cgi?id=13100 mike+html-wg-mailbot, P1, RESOLVED WONTFIX, 13TAG issue on HTML+RDFa and Microdata last call drafts

<adrianba> bug 13101?

<pimpbot> 11http://www.w3.org/Bugs/Public/show_bug.cgi?id=13101 mike, P1, RESOLVED NEEDSINFO, 13TAG issue on HTML+RDFa and Microdata last call drafts

paulc: [checks tag agenda for status of that tf]

rubys: we will hold that feedback to the same schedule as other feedback

paulc: the editors are required by our schedule to resolve all bugs by 12/31

<tcelik> why are there two bugs for this?

paulc: the bugs could just be closed if the tf report doesn't come in on time

tcelik: one per component
... the bugs could be turned into issues at any time between now and 1/14

s/adrianb/adrianba/

rubys: we will remind the tag of the schedule

<pimpbot> Title: HTML WG f2f -- 03 Nov 2011 (at www.w3.org)

<scribe> ACTION: paulc to contact the TAG [noah, jeni] to find out the status of the html data tf [recorded in http://www.w3.org/2011/11/03-html-wg-minutes.html#action01]

<trackbot> Created ACTION-205 - Contact the TAG [noah, jeni] to find out the status of the html data tf [on Paul Cotton - due 2011-11-10].

tcelik: the bugs are talking about what the w3c should or shouldn't do, not what this wg should do

rubys: we will require concrete change proposals

tcelik: a possible outcome is the wg could choose to stop publishing one of these specs

rubys: a change proposal could take such a form

tcelik: couldn't these specs be done in a cg?

rubys: i don't know of any proposal to stop any of these specs
... the bugs suggest merging the specs as one possible outcome

tcelik: this is at w3c scope, not wg scope

paulc: noah has asked to come to our meeting tomorrow
... we can ask him for status then

Schedule status (from the chairs' report to the AC)

<pimpbot> bugmail: [Bug 13558] input type=email should support friendly names <11http://lists.w3.org/Archives/Public/public-html-bugzilla/2011Nov/0276.html> 4** [Bug 13558] input type=email should support friendly names <11http://lists.w3.org/Archives/Public/public-html-bugzilla/2011Nov/0274.html> 4** [Bug 13570] why does input type=color support autocomplete? <11http://lists.w3.org/Archives/Public/public-html-bugzilla/2011Nov/0273.html> 4** [Bug 13570]

paulc: we will step through the slides we presented to the AC this morning

MikeSmith: [explains what the AC is]
... for 3 or 4 years we've done reports to the AC about the WG's status
... last 2 presentations that we did focused on what we were doing to get to LC
... we got to LC roughly on schedule
... ended LC round in August
... reminded AC we publish several documents, not just the HTML5 spec

[applause for MikeSmith's hard work publishing our specs]

MikeSmith: [list of our publications]
... told the AC about the author view of the spec (important to the TAG)
... Hixie had to make many edits to the spec to make the author view possible
... [goes through slide describing bug status & stats]

paulc: the wgstatus page gives you stats in fairly fine granularity
... [reads outstanding bug status from wgstatus page]

MikeSmith: [goes through AC report slides re: last call issues, issues awaiting or that have received new information, formal objections, etc.]

<MikeSmith> q

??: question for date for when new information is required

rubys: escalation deadline is 1/14

<tcelik> hober, slides: http://www.w3.org/2011/Talks/TPAC/HTML5/

<pimpbot> Title: HTML WG Update (at www.w3.org)

<MikeSmith> s/??/Cynthia_Shelley/

paulc: you want to add new info as soon as possible
... I think the schedule was overly pessimistic
... we might come out of last call after the spring ac meeting
... an AC rep asked about the relationship between html wg and whatwg

<tcelik> relation of HTML5 to the WHATWG? isn't that an FAQ?

paulc: another ac rep asked about what the decision policy would look like in several years
... the question might be better put 'how will work on html.next proceed'
... jim bell from hp said nice things & complimented chairs/wg/etc

<anne> tcelik: http://wiki.whatwg.org/wiki/FAQ#WHATWG_and_the_W3C_HTML_WG

rubys: we did get less flattering comment about the approachability of the process

<pimpbot> Title: FAQ - WHATWG Wiki (at wiki.whatwg.org)

Last Week in HTML5

<tcelik> Thanks Anne - I thought I'd seen that somewhere.

paulc: sent out draft schedule for tomorrow

<myakura> Proposed Friday F2F schedule from Paul Cotton on 2011-11-03 (public-html@w3.org from November 2011) http://lists.w3.org/Archives/Public/public-html/2011Nov/0013.html

<pimpbot> Title: Proposed Friday F2F schedule from Paul Cotton on 2011-11-03 (public-html@w3.org from November 2011) (at lists.w3.org)

karl: a few months ago I suggested to Ian Jacobs that I post weekly summaries of activity in the html wg on the QA blog

<MikeSmith> http://www.w3.org/QA/

<pimpbot> Title: W3C Blog (at www.w3.org)

karl: following html wg, the bug tracker, whatwg, webapps, etc
... not following css or svg
... anne from opera is doing something similar on the whatwg blog, following whatwg & webapps

<anne> http://blog.whatwg.org/

<pimpbot> Title: The WHATWG Blog (at blog.whatwg.org)

karl: thanks to the chairs for well writen decisions with links to relevant bugs etc
... this blog takes about an afternoon each week
... i don't cover everything happening on the open web platform

<MikeSmith> http://www.w3.org/QA/archive/open_web/

<pimpbot> Title: Open Web - W3C Blog - Category Archives (at www.w3.org)

karl: would love to know if this is useful

cynthia: i find it useful

ArtB: I read it & anne's as well, both useful

rubys: lots of cross links between the two

adrianba: proposal: in the telcon agenda, the chairs could point at the most recent post

<scribe> ACTION: rubys to ensure the weekly telcon agendas contain a link to the latest post in karl's series [recorded in http://www.w3.org/2011/11/03-html-wg-minutes.html#action02]

<trackbot> Created ACTION-206 - Ensure the weekly telcon agendas contain a link to the latest post in karl's series [on Sam Ruby - due 2011-11-10].

stevef: please also post the summaries to public-html

karl: ok, will post full text & link

<pimpbot> planet: Tantek: #TPAC2011: iterating on an HTML working group change proposal for an enhanced #HTML5 <time> element: http://www.w3.org/wiki/User:Tantekelik/time_element <11http://tantek.com/2011/307/t7/tpac2011-html-proposal-enhanced-html5-time-element>

<hober2> ScribeNick: hober2

[various requests for subscription options, twitter, rss, etc]

hadley: could it have its own twitter account?

karl: yes

<tcelik> +1 to what James is saying

jgraham: these posts should have some of the "inside baseball"
... developers can't keep up with the email lists

<pimpbot> bugmail: [Bug 13558] input type=email should support friendly names <11http://lists.w3.org/Archives/Public/public-html-bugzilla/2011Nov/0279.html> 4** [Bug 13392] i18n-ISSUE-72: BOM as preferred encoding declaration <11http://lists.w3.org/Archives/Public/public-html-bugzilla/2011Nov/0278.html>

<tcelik> +1 to transparency

rubys: don't want it to be a sleep-inducing sanitized press release

Teleconferences

cynthia: can we have more topical discussions on the telcon?

rubys: yes, we have an "other business" section for this

<karl> ACTION: karl to create a twitter account for the Open Web last week [recorded in http://www.w3.org/2011/11/03-html-wg-minutes.html#action03]

<trackbot> Sorry, couldn't find user - karl

rubys: we've never run over
... so we can use that time to help people get caught up with what's going on

richardschwerdtfe: can you get Hixie to call into the telecons?

rubys: no

<pimpbot> Title: HTML WG f2f -- 03 Nov 2011 (at www.w3.org)

rubys: different people have different work styles [etc.]
... had a useful f2f discussion with Hixie today

MikeSmith: Hixie is available on RIC, you can ping him if you have a question

<karl> s/RIC/IRC/

rubys: we would need to convince him that it would be worth it to join the call

[discussion of timliness of pinging Hixie on IRC]

mjs: there are people in the wg who are more comfortable with different kinds of interactions
... some prefer the mailing list, some irc, some telecons

<tcelik> indeed. this is why I publish my communication preferences: http://tantek.com/w/CommunicationProtocols

<pimpbot> Title: tantek / CommunicationProtocols (at tantek.com)

mjs: the chairs try to bridge between these different kind of people

richardschwerdtfe: asked you for a tech review 6 weeks ago

mjs: it's difficult to schedule a detailed tech review, it needs large block of uninteruppted time

rubys: the schedule has timeouts so that we can proceed even if we don't hear back from people

mjs: the decision policy goes with the cp that draws the weakest objections, and i try to make sure that apple's position is stated in surveys

[bits i missed re: the chairs' role in reviewing change proposals]

rubys: after you submit a proposal, you can still change it

<MikeSmith> scribe: MikeSmith

<hober2> richardschwerdtfe: i'm trying to avoid a 11th hour counter change proposal

richardschwerdtfe: we tried to reach Hixie and did didn't get any feedback
... Charles didn't hear back from him

sam: You got your response today

richardschwerdtfe: but we don't want to have to wait 6 weeks next time

mjs: you shouldn't wait six weeks to submit it

sam: we are ajourned

HTML WG f2f

03 Nov 2011

Contents

Intro

Suggested topics for discussion

the <time> element

Web & TV

HTML Decision Policy report

test suite status report

Change Proposal status

Schedule status (AC report)

Schedule status (from the chairs' report to the AC)

Last Week in HTML5

Teleconferences

Summary of Action Items