HTML Media Task Force Teleconference -- 04 Dec 2012

<trackbot> Date: 04 December 2012

<scribe> ScribeNick: adrianba

<scribe> Scribe: Adrian Bateman

<chriho> I am

<paulc> Who is <chriho>?

<chriho> can I listen?

<chriho> writing my master thesis about EME

<paulc> This is a MSE meeting.

<chriho> ok

<paulc> Chriho, Are you a HTML WG member?

<chriho> only public

<paulc> Then I suggest you drop off this call and contact me at pcotton@microsoft.com to discuss invited expert status so you can join the next EME call. Okay?

<chriho> Okay, I dropped it

Roll call, introductions and selection of scribe

paulc: done

Previous meeting minutes

paulc: we took most of November off so I pointed to the TPAC F2F minutes

http://www.w3.org/2012/11/01-html-wg-minutes.html

paulc: we might need to pick on peoples' memory for some items

Review of action items

ACTION-6?

<trackbot> ACTION-6 -- Aaron Colwell to give a couple of examples for section 2 -- due 2012-11-01 -- OPEN

<trackbot> http://www.w3.org/html/wg/media/track/actions/6

acolwell: I sent out a message right before TPAC but no one responded

<acolwell> http://lists.w3.org/Archives/Public/public-html-media/2012Oct/0062.html

acolwell: I've already started work on this - i assumed that silence meant nobody objected

close ACTION-6

<trackbot> ACTION-6 Give a couple of examples for section 2 closed

acolwell: i'll land these changes in a couple of stages because some is removing things, some is rewriting, etc

paulc: I had an action item at the WG level
... ACTION-223
... http://www.w3.org/html/wg/tracker/actions/223
... this is related to bug 17002
... we had two discussions - the first implied that we might need a change to IDs for 5.1
... later we picked up 17002 again and decided that there was a solution that removed the dependency on 18960 or any changes in HTML 5.1
... i believe this action item isn't need for MSE?

acolwell: we don't require it any more but we might need to link to something

paulc: the issue from TPAC is do we want to link to HTML 5.1 or HTML 5.0

acolwell: ok

paulc: my plan is to close ACTION-223 since we don't need this

acolwell: we don't have the dependency any more

paulc: let's deal with the detail when we discuss 18960

Baseline documents and Bugzilla information

paulc: aaron, you made an update on 28 nov
... one bug 19531 appears to be done but was not marked as resolved

acolwell: i didn't resolve it because there was a comment in the thread that it was ambiguous if the user agent supported a MIME type that it automatically implied MSE also had to support the MIME type and i'm making some modifications that it is possible for an element to play a MIME type but MSE doesn't need to
... within a day or so i will complete this bug

<paulc> Summary of Nov 28 changes: http://lists.w3.org/Archives/Public/public-html-media/2012Nov/0012.html

paulc: you've mentioned immiment changes - can you summarise which are the changes you have in mind

acolwell: i've been working on 19531 we just discussed, i also started on 18575 (removing sub-sections from section 2)
... concentrating on things i said i would remove and moving 2.4 and 2.5 into section 8

paulc: let's go through the bugs in the order i have them

acolwell: those are the ones i started - i was about to start on more and the discussion can help drive that

Actions from F2F meeting

paulc: i said in the agenda that we should discuss items that are still pending
... but i'd like to go to topic 6 on bugs because it's mostly the same
... i'd like to step us through some of the items and record what is outstanding and next steps

Discussion of outstanding bugs

paulc: the first pair of bugs we've partially touched on 18960 and 17002
... 17002 first, adrianba took an action item to implement a change removing the dependency on 18960

adrianba: i updated the bug this morning and assigned it to me

https://www.w3.org/Bugs/Public/show_bug.cgi?id=17002#c9

paulc: now 18960
... wasn't obvious from the meeting minutes what action we took

https://www.w3.org/Bugs/Public/show_bug.cgi?id=18960

acolwell: i believe the result was that MediaSource should generate unique IDs because the consensus was keeping them stable was more important than linking to the media
... think simon suggested we could add another property for the media id
... let me check my notes

<pal> P.S.: i need to relocate and drop off IRC

acolwell: david singer suggested unique and generated by media source object
... i think mark suggested a way to pass IDs to media source but don't think that ended up going anywhere

paulc: can you do 18960 in two stages: put the F2F recommendation as a comment in the bug so it is recorded then is it on your TODO list?

acolwell: if people are okay saying they are generated and unique then i can do that

paulc: yes, but we need to put that in the bug

adrianba: i think it is fine to add the unique IDs - we can always change later if we find we need the link back to media IDs

acolwell: i agree

paulc: then let's include this comment in the bug - we're not adding the link until we get more implementation experience
... if someone sees that and gives a reason for why it should be mandatory

acolwell: part of the reason why simon mentioned having a separate field is because this allows adding without violating HTML5

paulc: next is 18962
... this is pending a change from adrianba

<paulc> See https://www.w3.org/Bugs/Public/show_bug.cgi?id=18962#c3

adrianba: this is XHR append
... i started work on the related actions from WebApps
... creating new XHR spec and updating Streams spec which we'll need to link to
... we can then add the append() method for Stream
... and we'll need two events for complete and error
... but i have a separate question on events

paulc: do you have an ETA?

adrianba: i have friday morning to work on this but i need to coordinate with aaron on the changes he's making

paulc: 18963 - looks like the F2F minutes say we will throw an exception
... is this on your pending list?

acolwell: yes, i was thinking of adding that as part of the next change

paulc: 19531, which we've already discussed
... 17094 - we have a proposal from Bob Lund - there was an attempt to have an action on glenn to do something here but it didn't get recorded because tracker didn't recognise glenn
... i'd like to know the status of this because we have a concrete proposal
... glenn reached out to bob and myself and we've been having a side discussion about the outstanding issues
... we've been iterating on that - yesterday i proposed an alternate proposal for how to do appending that restricts out of order appending for TS
... you can't do out of order append without intervening abort call
... which is different to the other formats but the only way i can see to make TS sane
... we're making progress on it
... i need to send out a proposal to the list saying where we are

acolwell: is that fair, Bob?

BobLund: yes

adrianba: can you put a link to the mail in the bug once sent?
... also can we get something into the spec sooner rather than later and then iterate on it separately

acolwell: i'll try to get something in there then

Progression to First Public Working Draft

paulc: chairs met yesterday and the question i was asked by the team was when are we going to see the FPWD of media specs
... w3c is starting to get questions from outside about progress of the work
... despite all the public information about bugs, drafts, etc. but the team is asking what are the next steps
... so the question is of the pending bugs which are blocking bugs that this group wants to process before a CfC for FPWD inside the Task Force subsequently followed by sending to the WG for a CfC for publication at the WG level
... i'll give a candidate list

acolwell: what are the criteria for FPWD?
... is it okay to add details but not scope?

paulc: i'll try to answer but the answer is subjective
... you only have to agreement on going to FPWD not on all of the content
... many FPWD have links to bugs in the document to emphasise that some things are not final
... the second item we spoke about before the summer
... general agreement on EME and MSE - we wanted the API design to be right because it would be misleading to public a FPWD of the old design and then immediately change to a new one
... i think that item subject to some of the work in 18575 is largely done
... so the only other matter is the patent policy trigger that fires on FPWD
... members have 150 days to disclose or exclude patents on the draft
... so you want the scope to be clear to people so that if they do an active patent search then the material in the document is a good indication of where the group intends to take the document
... i think we've done the API design although there are some interesting bugs
... and on the last item, scope, we're close to done
... we're receiving pressure to show visible progress i.e. FPWD
... we don't have to be bug clean - that's Last Call
... so my question is which of the outstanding bugs does the TF want to get done before sending to WG with a list of outstanding bugs and say these are items we continue to work on but want to publish FPWD

pal: one question for the editors is whether the resolution of some of these bugs will cause a substantial change to the document
... perhaps we can see if those bugs will cause major changes

acolwell: i don't anticipate major changes from the remaining bugs

pal: for example on the seamless transition requires structural changes that is one thing but it might just be an annexe

acolwell: i can give an initial list
... 19531 (mime type), 18963 (rate limit appending), 18960 (id generation), 18962 (XHR), 18615 (buffered), 17094 (TS), 17002 (tracks), 17006 (language/kind)
... those seem structural to the API and aren't just detail

paulc: that's pretty close to the list that adrian and i had

acolwell: will check on 18615

paulc: does anyone want to propose something else?

pal: 19784 timestamp offset in the case of a multiplex - i'd like to get an idea of how the editors plan to address that

paulc: this wasn't discussed at F2F

acolwell: think we didn't get to it - came in right before TPAC

paulc: pal, you're saying it's hard to answer the question without knowing the outcome of this bug

pal: yes

acolwell: isn't this just related to considering the start of the segment where video is not audio

pal: this is if audio starts before video and you want to sync on video boundary

acolwell: this is about UA figuring out which timestamp to use
... i don't think this changes the append method signature

adrianba: when asked how should the editors process a bug i look for the spec text in the bug
... when i don't find some then my question is to turn it back to the group and ask for a concrete proposal

pal: can the other bug on seamless go in FPWD?

paulc: we have to be careful because we can't put everything in or else we'll never get there

acolwell: i don't think these are blocking because they won't change the API shape
... we definitely need to address the bugs but i don't think they are blocking

pal: if the threshold is whether an API change is needed then i can look at the bugs and see if i think they will change API

paulc: even if they will affect the API, we might still not need to take them if they're not of the same magnitude as the changes in the summer
... i'm going to take aaron's list, pal wants to consider 19673 and 19784
... i'd like to see a plan from the editors on the list saying what order they are going to address them in and by when
... then i can go back to the team with a plan
... if it turns out one of the bugs will take a long time perhaps we'll change our mind on if we need it
... i want to see the plan in two weeks and preferably well before then
... if we have something on the list i can point people outside the TF to that
... is that okay?

adrianba: yes

acolwell: i think this list is easy to knock out, xhr is probably the most difficult

paulc: for everyone involved in EME, you can expect the same questions next week

Other business

paulc: none

Chair and Scribe for next meeting

paulc: next meeting is dec 18 and then would be jan 1, when we probably won't meet, so it's important to make progress before the next meeting

Adjournment

paulc: good progress today, we're adjourned

trackbot, end meeting

- DRAFT -

HTML Media Task Force Teleconference

04 Dec 2012

Attendees

Contents