Media Fragments Working Group Teleconference

<trackbot> Date: 24 November 2010

<scribe> Scribe: Raphael

<scribe> scribenick: raphael

1. Admin

<tomayac> +1

Propose to accept the minutes of last telecon: http://www.w3.org/2010/11/17-mediafrag-minutes.html

+1

<Yves> +1

minutes accepted

<silvia> +1

ACTION-183?

<trackbot> ACTION-183 -- Raphaël Troncy to send reminders to all relevant groups -- due 2010-09-15 -- OPEN

<trackbot> http://www.w3.org/2008/WebVideo/Fragments/tracker/actions/183

close ACTION-183

<trackbot> ACTION-183 Send reminders to all relevant groups closed

Raphael: I did send an email to whatwg and html5
... but the mail to the html5 mailing list didn't get through
... subscription problem

Silvia, could you please forward my email to the HTML5 mailing list?

<silvia> oh, did it not go through?

<silvia> will do

scribe: I have also contacted Chris Double, Frank Olivier and Eric Carlson to get more reviews
... and the CSS working group regarding the bug http://www.w3.org/Bugs/Public/show_bug.cgi?id=10723

No, silvia, my mail went through the whatwg mailing list but NOT the html5 one

<silvia> ok

Discuss HTML5 bug

See: http://www.w3.org/Bugs/Public/show_bug.cgi?id=10723

Question is what we should do since this bug has been closed by Ian

<foolip> If he's wrong, we should explain why.

<silvia> I think we need to propose an actual spec change

<foolip> That will also do fine

<silvia> then there is something to discuss

Yves: for images, if the cropping is the default rendering behavior, it should be specified in our spec
... for video and audio, since there is more controls from the HTML5 spec, if the default behavior is highlighting, then it could be specified in HTML5

<Yves> (for time dimension)

<silvia> text about controls: http://www.whatwg.org/specs/web-apps/current-work/multipage/video.html#attr-media-controls

Yves: silvia, is there in html5, something that says how external controls are displayed for video?

<Yves> media-controls should know how to interact with a fragment

Silvia: there is only recommendations

<silvia> Yes, I agree - there should be recommendations on what to display for a <video> or <audio> element that has a media fragment URI - in particular since we may recommend a UI change

<silvia> but we need to formulate that recommendation

I agree Silvia

<Yves> for everything linked only to the content, the content has to define the behaviour, for external artefacts, like video controls, as they are part (even informally) of html5, somehting need to be in the spec

<Yves> +1 to Silvia

<silvia> and we need to do this differently for temporal to spatial fragments

Problem is also process: bug is closed. Should we escalate it? Or open a new one for temporal dimension since we agree on the space dimension

<tomayac> isnt part of ian's point that there are besides cropping no concrete use cases? only skimmed very rapidly, so might've misread

<silvia> I'm not 100% sure about process (maybe Philip knows better), but I think we may be able to reopen the bug with new information

<tomayac> aryeh's

<Yves> silvia, I think so too, we need clarification of the intent, and proposed text, as currently, Ian is perfectly right in closing this bug

<silvia> I agree

<silvia> we haven't provided any answers to the questions raised in the bug

Yes, Thomas is discussing the non-cropping use case for spatial region

<silvia> the bug was not registered with a particular focus on spatial fragments - it was generic for media fragment uris

<silvia> we need to have changes at least on how the control display should change and also how the scroll-to-fragid should be done

<silvia> http://dev.w3.org/html5/spec/Overview.html#scroll-to-fragid

<scribe> ACTION: silvia to draft the paragraph that the group will propose to HTML5 regarding how the control of media fragment URI should be done [recorded in http://www.w3.org/2010/11/24-mediafrag-minutes.html#action01]

<trackbot> Created ACTION-202 - Draft the paragraph that the group will propose to HTML5 regarding how the control of media fragment URI should be done [on Silvia Pfeiffer - due 2010-12-01].

3. Media Fragment Specification

close ACTION-195

<trackbot> ACTION-195 Add a paragraph in the section 7.1 to specify that video, audio, img or any href is all treated similarly (range request issued when facing a media fragment) closed

Raphael: contentious issue, in the case of a media fragment URI in a very particular context, i.e. in the audio or video element, AND that media fragment URI looks like a temporal fragment
... then browser SHOULD?/COULD? issue a range request in a first place

<Yves> MAY

Raphael: i.e. what we called the optimistic processing of media fragments
... if the server ack the fact the resource is a video, then it uses the recipe "Server mapped byte ranges" (section 5.1.2)
... if the server realized that the resource is not a video, then it ignores the Range header

<tomayac> in this case MAY sounds too defensive

<Yves> MAY sounds defensive, but it's the case for all optimisations that are approaching crossing the layers ;)

Raphael: anytime a URI looks like a #t= but only if this is the value of the href attibute of <video>/<audio> or the src attribute of the <source> element

Silvia: well, Apple implements the <video> element so that the value could also be a m3u playlist, not a media element

<silvia> m3u8

<Yves> ok, in that case, time range won't apply and you will get the whole thing

<Yves> hence the "optimistic optimization" (and the MAY)

Raphael: I'm curious what will happen if the video element point to a m3u8 resource in a browser that is not Safari

<silvia> http://developer.apple.com/library/ios/#technotes/tn2010/tn2224.html <- example

foolip, do you have an opinion on this discussion?

ACTION-191?

<trackbot> ACTION-191 -- Yves Lafon to update the production rules of the time dimension with the npt format for making the hours optional -- due 2010-11-08 -- OPEN

<trackbot> http://www.w3.org/2008/WebVideo/Fragments/tracker/actions/191

ACTION-173?

<trackbot> ACTION-173 -- Yves Lafon to produce the code that will check the grammar of both the URI syntax and the Headers syntax -- due 2010-06-22 -- OPEN

<trackbot> http://www.w3.org/2008/WebVideo/Fragments/tracker/actions/173

4. Use Cases and Requirements

<silvia> have you collected all the issues that could be added to the use cases?

Raphael: we have Ericson people using media fragment URI for video chat, Thomas using the spec in SemWebVid, HTTP streaming that might use it too
... I wonder if at some point we should not update the UC note with these initiatives
... silvia, so far I'm only keeping track of the initiatives, and I think we should report them as well as the issues in the document later on

Thomas: currently I'm using simple media fragment to point to temporal sequences of video
... + content negotiation to get either the RDF annotation of the sequence (e.g. closed caption) or the video bits
... the overall vision is that search results could include media fragments URI pointing to sequences that are relevant for a video search
... use case of finding people and highlighting faces with media fragments URI, so something more highlighting than cropping

Silvia: I think you should bring this to the whatwg mailing list

Thomas: concerned about the too large traffic of this list

<tomayac> silvia, maybe you could ping me a link to the thread, and i could jump in

Silvia: I understand, then make sure to include this in our planned reply to the html5 group
... are you aware of popcorn.js that does similar things that what you intend to do

Thomas: there is also a couple of BBC projects that do similar things, twitter streams displayed in parallel of programs
... also people in DERI working on this
... annotating conference media streams

Raphael: so you agree with the principle of reporting all these experiments in our UC note at some point?

5. AOB

none

<silvia> tomayac, you might want to read this thread: http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2010-August/027581.html

<tomayac> thanks, silvia

<tomayac> can i ask an, erm, stupid question: if you say html5, do you say whatwg, or w3c?

<foolip> tomayac, same same, whatever comes up first in your favorite search engine :)

<tomayac> wikipedia ;-)

great answer thomas

scribe: the second one should be W3C

:-)

<foolip> oh, in the context of <audio>, <video>, I think it's possible to send a range request without knowing the MIME type, but not really sane from a purity point of view. In any case I don't see it happening because the only benefit is one less round-trip, and only works with specialized servers

<foolip> I don't think the benefit is tangible enough that browsers or servers would bother implementing it.

<Yves> it will depend on big content provider interested or not in that (to reduce the load)

<foolip> right, should any browser or provider show any kind of interest in it I could of course reconsider

you represent ONE browser, and I'm talking to provider, such as Dailymotion, who could be interested ... so it might be worth considering at some point