HTML Media Task Force Teleconference -- 11 Nov 2014

<trackbot> Date: 11 November 2014

<scribe> scribe: joesteele

<scribe> agenda: http://lists.w3.org/Archives/Public/public-html-media/2014Nov/0015.html

Agenda, Role Call

Previous Minutes - http://www.w3.org/2014/10/30-html-wg-minutes.html

paulc: agenda looks long - trying to figure out where we have made progress

Media Task Force meetings

paulc: Possible F2F mtg -- trying to find a date and a host for January
... would like to discuss today what the plan is for TF meetings through the end of the year

<paulc> Nov 11, Nov 18, Nov 25, Dec 2, Dec 9, Dec 23, Dec 30

<paulc> Nov 11: today

paulc: these are the mtg dates

<paulc> Nov 18: David Dorwin will be absent

Dec 23 and 30 I will be absent

<paulc> Nov 25 and Dec 2; Paul is on vacation

paulc: could do MSE next week

<paulc> One proposal would be to do MSE on Nov 18

paulc: any objections?

Nov 25 is Thanksgiving in the US

scribe: Paul will miss that date -- or could cancel
... prefer not to cancel Dec 2nd if possible

<markw> I'd also prefer not to cancel

I would prefer not to cancel as well

<paulc> Proposal to cancel Dec 23 and Dec 30

paulc: we can cancel Dec 23 and 30

<paulc> We need an alternate Chair for Dec 2 and possibly Dec 9 (due to Paul's late arrival from his vacation)

<paulc> Nov 18: MSE

<paulc> Nov 25: Cancel due to USA holiday week

<paulc> Dec 2 and Dec 9: EME with alternate Chair

<paulc> Dec 16: EME and/or MSE

<paulc> Dec 23 and Dec 30: cancel

paulc: how does that sound?
... I will find alt chair for those mtgs

New EME bugs since last mtg relating to bug 26332

paulc: collected together bugs that are related in the agenda
... Henri's reply in bug 26332 pointing forward to his intent to file these bugs

<paulc> https://www.w3.org/Bugs/Public/show_bug.cgi?id=26332#c132 mentions the 5 new bugs

paulc: would like to know how we should process those 5 bugs
... more information needed?

ddorwin: lot of the text looks good
... happy to put in and iterate over it - possibly with issue boxes
... looks like a little duplication

paulc: are you proposing putting in the editors draft?

ddorwin: yes -- easier to see in-document what this looks like

paulc: Henri do you want to say anything

hsivonen: agree that one of the bug is a dup, Domenic bug does not contain proposed text, mine does
... ok with duping as long as my text is not lost
... I think the definition should be added first and we can iterate over that

<paulc> Bug 27270 may be a duplicate of Bug 27166

hsivonen: not proposing removing the "May" proposal on https - the bug that applies depends on which proposal you take

<hsivonen> https://www.w3.org/Bugs/Public/show_bug.cgi?id=27270 is a duplicate of https://www.w3.org/Bugs/Public/show_bug.cgi?id=27166 but latter doesn't have text

paulc: last two bugs are related to bug 26332
... David, do you want to assign all 5 bugs?

jdsmith: we have not resolved our position on bug 27269 -- would like to review that first

<hsivonen> https://www.w3.org/Bugs/Public/show_bug.cgi?id=27271 applies if the impl takes the https MAY and https://www.w3.org/Bugs/Public/show_bug.cgi?id=27272 applies when the MAY is not taken (though it still suggests doing it anyway even when the MAY is taken)

jdsmith: we have discussed, don't feel that we have closed on it internally

paulc: when?

jdsmith: this week

<paulc> https://www.w3.org/Bugs/Public/show_bug.cgi?id=27269 will be reviewed by MSFT this week

paulc: any other comments?
... on master bug 26332 or subsidiary bugs
... maybe since Microsoft has asked for a week, David can you send a report on where we stand this Friday?

ddorwin: yes -- will do the definition one first and then try to review the others

paulc: will put this batch of bugs on next weeks agenda
... would help Jerry if you can respond on the list or in the bug

Additional new EME bugs since the last meeting

<paulc> https://www.w3.org/Bugs/Public/show_bug.cgi?id=27283

New Bug 27283 - InvalidAccessError usage is questionable; use TypeError instead?

paulc: filed by Domenic - no responses yet
... have editors looked?

ddorwin: I think originally we used invalid access error since that is what other specs used

paulc: any other editors to respond?

jdsmith: I can take this one

<scribe> ACTION: jdsmith to respond to bug 27283 [recorded in http://www.w3.org/2014/11/11-html-media-minutes.html#action01]

<trackbot> Created ACTION-72 - Respond to bug 27283 [on Jerry Smith - due 2014-11-18].

Bugs not discussed at F2F

Bug 27067 - Define what to do when CDM becomes unavailable

https://www.w3.org/Bugs/Public/show_bug.cgi?id=27067#c11

paulc: some recent discussion
... where do we stand?

ddorwin: the current status is that we want to do a MEDIA_DECODE_ERROR is CDM is lost
... also discussion of a close() event
... just working through the details now
... other part is adding a close event to MediaKeys

paulc: both of these are documented in the bug .. is this on your list David?

ddorwin: just thinking a bit more about it -- I will take it

Bug 27093 - Support for proprietary/system-specific formats in initData should be discouraged/deprecated

https://www.w3.org/Bugs/Public/show_bug.cgi?id=27093#c10

paulc: comments with a question to Joe --

joesteele: planning to supply text if possible, waiting on internal feedback

Bug 27124 - Add "individualizationrequest" to the MediaKeyMessageType enum

https://www.w3.org/Bugs/Public/show_bug.cgi?id=27124#c14

paulc: some discussion in comment 14 -- Mark was asking why this is blocked before F2F

ddorwin: there is also a thread related to this on the mailing list -- just as long
... working through the privacy issues
... can probably reference the work Henri has done when addressing this

<markw_> Works for me

<paulc> TF email list discusssion starts at http://lists.w3.org/Archives/Public/public-html-media/2014Oct/0064.html

paulc: does this fall out naturally?

ddorwin: will be a minor change to add the enum and then just a dd a reference to Henri's definition

hsivonen: a bit concerned about blocking this bug on the privacy language from the other bugs, since that might take weeks
... might get past the point where we have to ship something, don't want to ship something with the wrong enum
... prefer to get the enum in the spec first and then get the privacy language in after

hsivonen: if everyone agrees that this can be added, don't want to ship with the wrong enum because it was not in early enough

<paulc> enum value proposed is "initializationRequest"?

paulc: is that correct?

hsivonen: was going for "individualizationRequest"

paulc: ah yes -- that was in the original bug

<hsivonen> get the bikeshed of whether it is going to be called individualizationrequest or something else over with without blocking on privacy definitions

<paulc> So we have two alternatives:

paulc: so we have two alternatives?
... anyone that objects to individualizationRequest?
... hearing silence -- and Henri asking we implement this sooner rather than later

ddorwin: yes we can do it, looking for enum naming in WebIDL
... 'R' may not be capitalized

paulc: point out that there was a further discussion with another enum

<hsivonen> the "licenserequest" enum item is not camelCase

joesteele: that was me -- dropping for now.

paulc: can you implement this now with the correct case?

hsivonen: I am satisfied

paulc: you would need to indicate that this will point to text elsewhere

Bug 27138 - Consider changing how the MediaKeySession method algorithms run other algorithms to more accurately reflect implementations

https://www.w3.org/Bugs/Public/show_bug.cgi?id=27138

paulc: no responses since Davids original post -- what is the status?

ddorwin: we have inconsistent algorithms where we fire other algorithms, current order is intentionally defined that way, but might be hard to implement
... not super urgent, but would urge folks to read it

paulc: would be useful to have implementer feedback then on options I and II?

ddorwin: these are likely proposals

paulc: anyone who can review?

jdsmith: think this calls for broad review -- I will review

paulc: thansk Jerry -- hopefully others will review as well
... in agenda item 9 -- long list of items we need progress on

Bug 27166

Bug 27166 - All identifiers associated with a user should be clearable in the same way cookies are

https://www.w3.org/Bugs/Public/show_bug.cgi?id=27166

hsivonen: had not seen this bug when I wrote my text on the identifier
... think my bug includes more actionable text than this one

Bug 26887 Allowing license servers and CDMs to control data persistence and secure release

https://www.w3.org/Bugs/Public/show_bug.cgi?id=26887

paulc: Jerry you were going to make a concrete proposal

jdsmith: having some offline discussions about this -- have an action to split out part of this into another bug

Bug 26372 Report issues/events not related to a specific method call

https://www.w3.org/Bugs/Public/show_bug.cgi?id=26372

ddorwin: I had an update about maplike in comment 34

<ddorwin> starting at https://www.w3.org/Bugs/Public/show_bug.cgi?id=26372#c34

ddorwin: took the thread discussion and added the implicit methods
... one of the issues with adding this is that MediaKeySession becomes maplike
... in comment 35 -- proposal is to define a new method that is maplike
... some discussion of returning a copy mentioned in comment 24
... did not respond to Joe - but still have the keyschanged event
... Object.observe() we decide not to do - not sure now

markw: I was expecting that we would just have a new attribute that exposes the maplike interface
... and the CDM would always just inform the application when we have a new keys status
... not sure why the UA could not just keep an up to date status of all the keys

ddorwin: the new and separate member is comment 35 -- think we are going that way
... exposing as a member is nice and easy, but could lead to unexpected behavior
... whether the method returns a Promise or not is not a huge issue, but other methods do return a Promise
... allows more flexibility. If application is just ignoring, no reason to copy these across
... most method will probably use Promises going forward

markw: so you will always get a copy when you access attributes, in WebCrypto we decided that there was one copy that was created and that is the onle that is always returned

ddorwin: in comment 24 - it says that any sequence returned from this must be a copy -- doesn't have to be but would be a fixed length
... we are defining our own structure so we can do what we want
... i.e. continuously updated, but could be complex
... don't really care -- if folks think that is OK we could do that

paulc: anyone else care?

<hsivonen> I think returning a copy is better than having something change under iteration

paulc: Mark did that answer your questions?

markw: not sure I can distinguish all the differences, not sure which would be easier for developers
... as far as changing underneath you, not sure it makes a difference when the change happens

paulc: so you are ok with a copy then?

markw: I am not sure how to discriminate between the two -- maybe someone more familiar with WebIDL should comment.

ddorwin: we should get feedback from WebIDL and developers

markw: I will followup with our developers

paulc: need folks to review this proposal from David

Bug 24874 Positive isTypeSupported() may be misleading (MSE vs. .src=)

https://www.w3.org/Bugs/Public/show_bug.cgi?id=24874#c4

paulc: Joe you commented and there is a reply from David

ddorwin: question is whether we want to allow you to detect whether .src and EME are not supported
... Joe says yes - no other feedback
... want this to be correct by default for applications
... probably means this has to be MSE only not .src
... mostly been talking about what to do after setting the .src

joesteele: my question back to David is what the behavior would be

ddorwin: setting video.src after mediakeys is an expected use case -- can't do that
... at a high level we can discuss what would happen if we set a src that would not work with the mediakeys set

joesteele: so this issue is that the UA may not support .srcs with or without EME
... not whether that actual content is encrypted or not

paulc: David was suggesting adding Chris Pearce to the bug and then responding with a question, or I would add to the email list directly

ddorwin: might be better to add to the list since this is old bug

joesteele: I will send ChrisP an email on this

jdsmith: I recall that we had a conversation on this and did not prioritize, but we said it was a good thing to be able to differentiate between these cases using capabilities.
... would like to be able to distinguish this for websites
... seems like a logical request

ddorwin: before we only had isTypeSupported -- is better now with a dictionary and capability queries
... need to determine what the default is here
... feedback is welcome

paulc: was not trying to change priority just wanted to bring it to the attention of members
... got through the items I wanted to discuss
... thanks everyone!

;-)

- DRAFT -

HTML Media Task Force Teleconference

11 Nov 2014

Attendees

Contents