Web & TV IG meeting - TV workshop follow-up

16 Apr 2014

See also: IRC log


Present (identified attendees)
Andy_Hickman_DTVL, bgidon, Bin_Hu, bryan (Bryan_Sullivan), CyrilRa, ddavis, elindstrom, geunhyug, giuseppep, gmandyam, igarashi, jcdufourd, jcverdie, jon, JP Evain, karen, kawada, kaz, MarkS, MarkVickers, pal, paul_higgs, schuki, simon, skim13, wuwei, yosuke, Daniel_Wester
Present (unidentified callers)
+1.818.370.aadd, +325045aacc, +44.178.442.aaff, [GVoice], [IPcaller], [IPcaller.a], [IPcaller.aa], [IPcaller.aaa], [IPcaller.aaaa], ??P54, ??P56, ??P9
Andreas Tai, Giles Godart-Brown, Victor Klos


<giuseppep> https://docs.google.com/presentation/d/1nQGA_g6Zp1mbO75LFVEzzHDOwxDjTjiFRas4KoF1i7M/edit#slide=id.p28

workshop next steps

<giuseppep> https://www.w3.org/2011/webtv/wiki/Agenda_and_Minutes/Agenda_Telco_16th_April_2014

giuseppep: The workshop was held in Munich a few weeks ago
... The slides for the next steps are here: https://docs.google.com/presentation/d/1nQGA_g6Zp1mbO75LFVEzzHDOwxDjTjiFRas4KoF1i7M/edit#slide=id.p28
... Firstly, there was a presentation on a new initiative called GGIE by Glenn.

giuseppep: There was a discussion on use cases and requirements which we also do at W3C.
... I don't we have much next to do with regards to GGIE.

Glenn: The group is meeting in a couple of weeks to discuss the charter and then we can come back to W3C to see if there's any overlap.

paul: Is GGIE a W3C activity?

Glenn: No, it supports other standards bodies including W3C, IETF, etc.
... We're not sure yet where are "home" would be.

giuseppep: The slides are linked from the workshop agenda: http://www.w3.org/2013/10/tv-workshop/agenda.html

<kaz> GGIE slides

giuseppep: Next is synchronization of media and metadata.
... Many of these things are spread across existing specs
... There isn't any significant activity we can do, but the IG mailing list can be used to raise awareness of issues that could be reported as bugs.
... Do people agree?

<jcverdie> +1

Jean-pierre: I don't remember having such a deep discussion about metadata.

giuseppep: This is more about integration of existing specs and HTML5 spec.
... Do you remember something different?
... I'll mail a link to the slides I'm sharing.

N.b. GGIE is "The Glass to Glass Internet Ecosystem"

giuseppep: If we agree on the conclusion, I'm moving to slide #4 - testing.
... Note that the slides are not automatic - you have to click on each slide.
... There was a discussion about testing which is very important.
... There was a request to understand more about what W3C is doing.
... There's a need for more test cases.
... Also, a discussion about what testing actually means.
... Different people may have different goals.
... If resources/tests are out there, do they match testing requirements?
... What possibility is to use the IG or some other group to get a better idea of what's out there
... And maybe have a Test The Web Forward event for TV.
... The conclusion was the IG could facilitate a conversation.
... There is already a working group and interest group.
... Any comments so far?

paul_higgs: We used to have a Testing Task Force within this IG. It produced analysis of what is needed for TV testing.
... A lot of effort was put in - where did it go?

giuseppep: Since the activity of developing steps actually got stopped, there was no progress.

MarkVickers: There was a lot of effort around testing. There was a set of requirements that we published.
... What it said was we need a centralised placed for tests and have them run in a consistent way.
... Some of that work has been done. We have centralised tests.

MarkVickers: The big challenge was getting better coverage of tests and for that, W3C built a detailed test plan.
... This required possibly more staff, definitely more funding but W3C didn't receive any money.
... There is still effort going on and contributions by browser vendors. There is a backlog of tests and a requirement for test reviews.
... So there's the opportunity for people to contribute tests and reviews, or funding for testing.
... The depth of coverage is going to take a lot longer to improve.

paul_higgs: Which group should we take ideas to?

MarkVickers: I would contact the people within each working group

bryan: I gave an overview of the testing effort at the workshop.
... The testing IG is not going to do anything for you, and probably not the WGs either.
... This TV IG needs to focus much more specifically about what's needed for certain test cases.
... Just saying fill the gaps is not enough to make headway.
... This TV IG needs someone to focus the work and contribute resources to do that.
... I suggest you take a look at your existing HTML5 STB - map the features to web technologies, e.g. CSS opacity, etc.
... Define which features are used and decide which tests are needed.

giuseppep: What's missing is where the tests are and tools are?

bryan: The community has to serve itself. I said we should do this on the wiki.
... We should detail how to do this for ourselves.
... We have to jump in - the people who know how to run tests have to lead the others. No-one else will do it for you.

paul_higgs: We need to be in charge of our own destiny.
... We should collect the tests for the TV environment.

MarkVickers: I strongly disagree with this. The old testing group was closed down but there is a new one and a structure for submitting tests.

<bryan> what is the active group, mark?

MarkVickers: If you want to submit tests, there is a process and existing group to do that.
... I don't think we should collect our own group of tests.
... There's a good test plan and what's missing is people to do the tests.

paul_higgs: I think we agree but we're saying it differently.

bryan: There's no W3C group any more, it's hidden behind GitHub.
... Right now there is no focussed effort.
... Somebody has to coordinate and organise the effort.

giuseppep: People want to contribute test cases but they want to know where they go. And they want to know what tools are there.
... The requirement is there.
... There is a need for someone to drive this.
... As it stands it's a bit confusing.
... How to move forward on this?
... Maybe Clarke would be willing?

MarkVickers: Are you talking about getting a task force going again?

giuseppep: If everyone works on their own things could fall apart.

<bryan> the only mechanism of communicating with the "group" is the test-infra mailing list, and following the discussions on github (pull requests, bugs etc)

giuseppep: I.e. if we just provide people links.

bryan: I agree we should start with links.
... We have to identify what tests are there.
... Right now there is no guidance.

giuseppep: Please could one of you send an email to the list summarising these points?
... Maybe in the next call we could cover this.

bryan: I said we could do something to help fill this gap.
... Help to explain how to get engaged and what to focus on.

<darobin> may I suggest you email public-test-infra with questions, comments, and grievances?

giuseppep: In two weeks time we'll review it.

andy: There's a view that if you've got a lot of money and engineers, all the problems would be solved.

<darobin> public-test-infra is there to help and is very helpful (also #testing)

andy: But I don't believe that's the case.

<darobin> I strongly suggest that any work on testing in WebTV be coordinated there, it will avoid confusion, duplication, mistakes, etc.

andy: I don't think you'd end up with something that HbbTV or IPTV Forum Japan could use. There's no way to handle test IDs, test waivers and certification.
... Some of the basic infrastructure is not there and I can't see it happening.

<bryan> robin, people don't know about those tools of communication - we can start by letting them know. but what this group needs most is to decide what features are important and to put resources onto those tests, reviewing, developing, whatever

MarkVickers: In the DLNA we've referenced W3C tests for some time.
... It's not a testing organisation so you're not going to get full infrastructure such as waivers. The referencing groups should do that.

<darobin> bryan: I think you'd be more successful deciding what to do by interacting with public-test-infra right away

<darobin> giuseppep: no, but if it's really useful I can kill this other thing I'm doing in parallel

<bryan> robin, I will do that

MarkVickers: W3C is a place where the tests should be reviewed and seen to be technically correct, and hold a repository of the tests.

andy: I agree with that, but what I would hope is that the infrastructure has some support in other organisations.
... E.g. they need unique IDs and version numbers.

<jon> The speaker is Andy Hickman from DTVL.

bryan: HbbTV has an exacting requirement for tests.
... That level of mapping, etc. does not exist in W3C.
... That needs to exist to make it happen.

andy: It's hard for third parties to reference in a robust way.

giuseppep: So it's not an issue of contribution but of how they are maintained.

<bryan> the problem is that W3C tests has been developed in an adhoc way with little specific guidelines or consistency on metadata, linking to spec/test assertions, etc. that is one of the key gaps for utility in certification programs. this was a comment I made at the workshop.

giuseppep: (scribe missed that)

<darobin> bryan: that is not true, they are largely consistent — metadata is simply considered to be of lesser importance, and something that can easily be provided externally

<bryan> it was also one of the goals for the Web Testing IG that have so far not been realized - how to change the process of test documentation and development to obtain more rigor.

giuseppep: Next step - Bryan will send out information then I'll try to summarise the issues that have been raised.

andy: Do you know what the next step is?

giuseppep: My understanding was that if there was no clear direction then it's not worth taking this to the testing group.

<bryan> robin, OK I agree that if metadata needs to be provided externally, that is a solution if the underlying tests are structured in at least a stable, granular way. This group can provide that "external" augmentation. That's the "focus" I am talking about.

giuseppep: My proposal would be to not rush anything until we see that there are good tools out there that could be used.

<darobin> bryan, that's certainly something we can handle

darobin: Robin here. I work for W3C managing publication of HTML spec and help with testing effort.
... It would be useful if any discussion of testing could be coordinated with public-test-infra mailing list.
... There's been a lot of hearsay and misunderstanding.
... The testing list is friendly and helpful.

<MarkVickers> public-test-infra@w3.org

darobin: Things are moving fast in terms of infrastructure, resources, and documentation.

<bryan> internally, we are starting here... how to build test runner scripts based upon the existing tests in the repository, that execute the tests we are more interested in - that's one of the basic metadata items - what set of tests do you want to run. Then when we have found gaps, reviewing/developing the tests for them.

darobin: Don't hesitate to ask someone from the testing group to join in your discussion.

<kaz> testing infra email archive

giuseppep: When you send that mail, Bryan, please include the testing mailing list.

bryan: Will do. Sorry if it seems I don't appreciate the effort - I do but I think it's better if we work as a test.

<MarkVickers> I encourage anyone to take any questions to public-test-infra@w3.org

<darobin> +1 to working as a team, hence the suggestion to discuss with the testing team :)

giuseppep: Next is rendering of linear video, aka Tuner API
... We discussed this previously with Bin leading the discussion

Bin_Hu: We discussed the need to take action quickly if we want a Community Group
... There are lots of people interested - strong interest in driving this forward.
... We need five supporters. Last week we had a role call and got more than five supporters.
... As a result, this will be moving forward but still no volunteer to drive the effort.
... As a next step I'm thinking maybe we can help start off the work and see how the group proceeds.
... There are other groups that have similar technologies so we can see how those specifications can contribute to our efforts.

giuseppep: So we're still missing someone to lead the group.

Bin_Hu: We have jcverdie happy to co-chair

jcverdie: I'd be happy to do it with Bin.

Bin_Hu: OK.

jcverdie: Great.

giuseppep: So the people who expressed an interest can propose the group.
... We have provided a lot of use cases in the past year so it's a case of deciding which one to use for the group, based on existing specs and requirements.
... So the Community Group will be independent of the IG and also open to anyone, not just W3C members.
... Once it's created let's announce it on the IG and maybe the CG chairs could report to the IG every month, for example.

<kaz> Community Groups page

Bin_Hu: I think that's important because of the heritage of the CG.

giuseppep: Any other comments?

<jcverdie> +1 to reporting progress to IG obviously

giuseppep: Next is slide #6 - mostly regarding bugs around video element.
... The IG should maybe get some feedback, maybe comment on the bugs.
... Do people feel our involvement would be useful?
... At present we don't have task forces, just one IG call every two weeks.
... The agenda for each call is open so if there's a specific spec that could benefit from being presented to the group, that's fine.

MarkVickers: I like a single call and having the content decided as needed.

paul_higgs: I think I heard you say we could discuss the bugs on the call. But how should SDOs report bugs?

giuseppep: Bugs should be reported to the WGs, e.g., HTML WG
... If there's a problem with a spec, report it to the spec on Bugzilla - https://www.w3.org/Bugs/Public/
... If there's a problem then raise it with this IG.
... Slide #7
... Communication between two UAs
... We discussed whether there's anything more the IG can be doing.
... The answer was no, because there are already groups doing this work.
... See slide #7, and if you're not sure how to join in, ask the list.
... Next is #8 - performance measurement.
... There was a discussion about what's needed and what W3C can do.
... Is now a time to go to the web performance WG with requirements?
... If so, there's a need for people to drive this discussion.
... Is there anyone on the call that feels this is important enough to drive forward?
... Is this related to the testing activity or outside that scope?

bryan: Performance measurement and benchmarks is always a tricky area. So far it's outside W3C's area.
... Functional tests are the main focus.

MarkVickers: There's a performance WG, right?

giuseppep: Yes, but they focus on performance specs.
... If someone wants to discuss this, we're open.

<MarkVickers> Web Performance Working Group: http://www.w3.org/2010/webperf/

bryan: If it's just discussing common ideas, that's OK, but don't build your hopes up with benchmarks.

giuseppep: I think you need APIs for browsers to measure performance.

bryan: That's worthwhile talking about here.

giuseppep: Let's see if there's enough interest to continue the discussion.

yosuke: What we've learned (from HybridCast) is performance is still really important for actual production.

<bryan> I think experience with the video API will help identify that performance data gaps exist - testing may also uncover gaps in what people were expecting the API to provide, and that can drive bugs.

yosuke: If no-one's willing to kick off the discussion, I'd be happy to.

giuseppep: It would be helpful to send a mail to the list and if there's interest it can be followed up in the next call.
... If you can compile a list, e.g. of HybridCast issues, that would help start the discussion.
... Slide #9 - there was a discussion about accessibility.
... There's more than just subtitles and captioning. We didn't really go into details, just continue in the IG if there are enough contributors.
... If you're interested, send a mail to the list and we can discuss it on the call.

Pierre: I saw a number of issues filed by Jon Piesing for HTML, e.g. caption/subtitle selection.
... Have they been addressed or resolved?

Jon: They are being addressed, slowly.

Pierre: Do you feel there's more that could be done?

Jon: It feels like it's purely a discussion between Silvia Pfeiffer and myself.
... It's hard to know what other people think of the discussion.
... You can assume that silence is consent, but the lack of people makes me nervous.

Pierre: Do you have a record of all the issues that are relevant to this group?

Jon: I can send something to Giuseppe who can forward to the list. There'll be more to come.

Pierre: Feel free to send it directly to me as well.

<MarkVickers> I believe anyone can send to the public list.

<kaz> right

giuseppep: It's OK to send a mail to the group asking for other people to comment.

Pierre: Jon, I'm looking forward to that and helping if I can.

MarkVickers: It is a pattern for these things that one of the editors will lead the discussion. In this case it's Silvia as she's an HTML editor.
... The bugs you reported were well written. I agree it's not a fast process but when it gets done it gets adopted by all the browsers.

Pierre: Now that HTML is taking over video content, that video content comes from places outside the browser.
... I'm sure many of us in the TV industry have experience we can contribute.

MarkVickers: I agree. The HTML spec a few years ago was very basic for video and this IG has contributed a lot. Also, individuals from this group have gone on to work on specs directly.
... Web browser experts are not necessarily media experts so they need us to chime in on these issues. Our role will continue to be important.

giuseppep: Slide #10 - Pluggable CDM for EME
... Jan is not on the call. I didn't feel there was anything that could be done within the IG.
... It's something that could be mentioned in the HTML WG Media Task Force.

MarkVickers: In general, the W3C specs just cover the interface between the application and the browser.
... The W3C takes the position that specs don't touch the plugin or anything above the browser interface.
... The W3C in this area hasn't taken on any work defining CDMs or plugins.

<bryan> sorry, gotta drop - late for WebMob

MarkVickers: The question of whether there's a plugin API that would work cross-browser is still an open issues.
... You could raise it in the HTML WG Media TF and see the response.

giuseppep: So if Jan and the people who raised the issue are not here I'll leave it but they can raise it in the IG if they want to.
... That's all I have for the workshop next steps. Anything I forgot?

kaz: Maybe Daniel or I should talk with Mark Sadecki about the possible collaboration (maybe including his direct participation in the IG)?

giuseppep: OK.

<giuseppep> http://lists.w3.org/Archives/Public/public-web-and-tv/2014Apr/0025.html

Request from DAP WG

<inserted> DAP Request

giuseppep: They want feedback on the Network Service Discovery API
... They want to know if the spec covers our use cases
... If people are interesting in implementing it and can share plans.
... And if there are additional requirements, such as if there's no support for legacy devices will they still implement it/not implement it?
... As an IG we can't answer this but we can send a mail to get responses.
... So I'll follow up and draft an email, sending it to our liaison contacts.

Bin_Hu: Based on the email that they sent to everybody, looks like if there are no further implementations then their work will be shelved.
... It's a good time to speak up if you're interested in it or in implementing it.

paul_higgs: What do you mean by implementing it?
... TV manufacturers are interested in it but it's implemented in the browser, so it's up to browser vendors.

giuseppep: It's not limited to the browser vendors.
... Forks of WebKit would still count as implementations.

Also, it's up to TV manufacturers to request requirements as customers of browser vendors.

Update from the Media API TF

giuseppep: Is there a volunteer to answer requests for an update on the status of the group?

MarkVickers: I can draft an email.

giuseppep: So let's close the call there.
... Anything else?
... Next call in two weeks then and I'll send a summary to the IG.

<jcverdie> thanks giuseppep

<jcverdie> bye all

kaz: Next call will be a normal TV IG call, right?

giuseppep: Yes, and sorry I got the timezones wrong.

giuseppep: If you want an invite, please let me know and I'll send you one.
... Goodbye everyone.

Meeting adjourned

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.138 (CVS log)
$Date: 2014-04-16 15:28:11 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.138  of Date: 2013-04-25 13:59:11  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/@@@/bryan/
Succeeded: s/@@@/bryan/
Succeeded: s/bryan:/@@@:/
Succeeded: s/bryan:/@@@:/
Succeeded: s/@@@:/paul:/g
Succeeded: s/@@@:/andy:/
Succeeded: s/@@@:/andy:/g
Succeeded: s/@@@:/andy:/
Succeeded: i/workshop was held/scribenick: ddavis
Succeeded: i/There was a discussion on/scribenick: ddavis
Succeeded: s/resources/infrastructure, resources,/
Succeeded: s/IG/CG/
Succeeded: s/the e.g./the WGs, e.g.,/
Succeeded: s/there's/there are/
Succeeded: s/about his participation in the IG/about the possible collaboration (maybe including his direct participation in the IG)/
Succeeded: i|They want feedback|-> http://lists.w3.org/Archives/Public/public-web-and-tv/2014Apr/0025.html DAP Request
Found Scribe: Daniel
WARNING: No scribe lines found matching ScribeNick pattern: <Daniel> ...
Found ScribeNick: ddavis
Found ScribeNick: ddavis
Found ScribeNick: Daniel Davis
WARNING: No scribe lines found matching ScribeNick pattern: <Daniel\ Davis> ...
Found ScribeNick: ddavis
ScribeNicks: ddavis, Daniel Davis
Present: Bryan_Sullivan JeanClaude_Dufourd JC_Verdie_(on_IRC phone_bridge_is_full) Yosuke Funahashi
Got date from IRC log name: 16 Apr 2014
Guessing minutes URL: http://www.w3.org/2014/04/16-webtv-minutes.html
People with action items: 

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.

[End of scribe.perl diagnostic output]