SV_MEETING_TITLE -- 21 Nov 2012

<olivier> Scribe: Matt

<olivier> ScribeNick: mdjp

<chrislowis> Zakim: +[IPcaller] is me

<chrislowis> :D

<chrislowis> olivier: I've connected telephone with irc already :)

<olivier> good :)

Testing

<olivier> http://lists.w3.org/Archives/Public/public-audio/2012AprJun/thread.html#msg739

olivier: discussion on testing - approaches to be taken - how will we manage test approach, (manual or automated) and submission

<olivier> http://w3c-test.org/

<olivier> http://w3c-test.org/ttml/ttml11/testsuite/web-framework/START.html

<chris> WebKit Web Audio tests: http://svn.webkit.org/repository/webkit/trunk/LayoutTests/webaudio/

olivier: w3c maintain a test site http://w3c-test.org/ contains interesting example (ttml)

<Zakim> ChrisLilley, you wanted to discuss link between testable assertions in spec and testability and to discuss link between testable assertions in spec and testability and also to

<olivier> (WOFF Test suite. link?)

ChrisLilley - recommends use of W3C test harness, been used in Fonts WG for WOFF. Link betwen how spec is written and tests - easier to check what a test is testing if phrasing is directly testable, testing should not be left too late

chris - peter lynn working on shepard, testing tool used by css team - not fully integrated with w3c instance

<olivier> (Peter Linss is working on Shepherd)

<ChrisLilley> http://www.w3c-test.org/framework/app/suite/woff-ua/results for user agents

<ChrisLilley> Peter Linss

<ChrisLilley> http://www.w3c-test.org/framework/app/suite/woff-at/run test harness

chris let tests drive the spec. If feel there are gaps and wording is not right we can change the spec over however 75 test written in webkit so far

chris recommend we look at tests written so far which run evey day

chrislowis: do webkit test depend onan implementation?

olivier thanks

<ChrisLilley> link to existing webkit tests?

<olivier> WebKit Web Audio tests: http://svn.webkit.org/repository/webkit/trunk/LayoutTests/webaudio/

<chrislowis> Zakim: IPcaller is me

chris: result of tests produces a wavefile and compares with reference. Example spatialization - compare panned file against reference file

olivier: we should do this manually as wave file depends on the platform - can't be fully automated?

chris: webkit automated but different on other browsers, quite painful to manage baselines for other platforms.

olivier: testing performance vs regression etc?

ChrisLilley: svg heading towards quick testing to check a number of conditions aiming at automation, preferable for anything that can be automated to be automated

olivier: We will have to have a mix of manual and automated testing

first….

olivier: Thierry has time to take care of coordinating the test submission - agrees with chris tests will drive how we change the specs, not just a translation.
... next step look at which of the webkit tests could be submitted as a first step

chris: seems like a good first step

Ehsan: all current tests written for Mozilla can be submitted to the test kit

<chrislowis> I'm happy to help with this testing effort too.

Face-to-face meeting

<ChrisLilley> happy to help with web midi testing

olivier: havent talked about midi testing - need to think about this also

<ChrisLilley> https://www.w3.org/2002/09/wbs/46884/f2f-2013-1/

<jussi> ChrisLilley: :) let's talk more about that on the list

<ChrisLilley> jussi - yes, sure

<ChrisLilley> https://www.w3.org/2002/09/wbs/46884/f2f-2013-1/results

olivier: f2f meeting couldn't meet at TPAC due to various reasons, poll in the w3c site to collect dates and locations for next f2f. Please answer if you have not done so already
... preferred time at the moment is late or early March, location mixed preference for location (US)

<chrislowis> Let's have it at SXSW :)

<shepazu> March 8-12 SXSW

<ChrisLilley> http://sxsw.com/ 8-17 March

Audio WG Document Milestones

<olivier> http://www.w3.org/2011/audio/wiki/W3C_Audio_Publications_and_Milestones

<chrislowis> Joe: this page: http://www.w3.org/2011/audio/wiki/W3C_Audio_Publications_and_Milestones

olivier: milestones from recharter available on wiki - milestones may be ambitious, require test suite to be ready very soon. Added updated estimates for api

<joe> chrislowis: thx

olivier: need a clear an honest view about our progres at any time. Some groups don't include any milestones, we have tried to estimate, need to talk about current estimases from audio and midi api
... could reach last call march 2013 - means we have solved all main issues internally and push for public to review
... is this still reasonable?

jussi: still realistic but needs discussion

chris: Don't know - fast progress on adressing issues but many less significant issues. More issues will arise, 1st Q 2013 still a posibillity but may be ambitious

jussi: Number of issues from Mozilla will increase

olivier: leave at Q1 2013 and reassess

chris: we should publish a working draft, discussing using web IDL enums.

olivier: due for working draft soon, last was in August.

<Zakim> ChrisLilley, you wanted to mention that linking met milestones to the actual specs would be helpful

ChrisLilley: helpful for people who want to look at milestones if they show visibly progress (style completed differently)

<olivier> ACTION: olivier to update milestones tables with links to FPWD [recorded in http://www.w3.org/2012/11/21-audio-minutes.html#action01]

<trackbot> Created ACTION-52 - Update milestones tables with links to FPWD [on Olivier Thereaux - due 2012-11-28].

shepazu: public depending on this spec, may not be tracking the day to day work of the group but still invested in the work.

Face-to-face meeting

state of Mozilla implementation

shepazu: how do you feel about the spec so far, how long will it take to get to a state when people can start to test

ehsan: expect to output audio within the next couple of weeks - the spec in its current form is not really well suited to be implemented by people without specific knowledge of audio processing
... doesn't really cover edge cases eg no mention of invalid property values, types of exception not specified. Treating webkit as defacto, not ideal.

doug: good assesment, how many months will it take Moz to get to the point where people can start "playing" with it

ehsan: nightly channel by end 2012, working on a list of 7-10 nodes, putting off the code audio data api, expect to be finished this week and will be first part of the api available. Estimate that end of 2012 still reasonable.
... audio api will be hidden behind a preference - no concrete plans on when this will be enabled by default

<Zakim> ChrisLilley, you wanted to ask if there is reference material (like graphics gems) to link to for algorithmic details

<ChrisLilley> https://wiki.mozilla.org/RapidRelease/Calendar release date FF20 2012-Nov-20

chris: thanks - understand how parts of implementation could be difficult. Would like to continue getting feedback. Can add more details where relevant.

ehsan: spec should reference literature on digital audio,

ChrisLilley: svg had similar issues with expert knowledge during implementation. Its helpful to reference literature.

<ChrisLilley> separate glossary with links sounds like a good idea

olivier: for definitions, reference literature - think about having seperate glossary of terms if required. Would i make more sense to have a glossary or links to reference material

<olivier> [discussion on level of detail in algorithms. CLilley says it doesn't have to be a foolproof guide, but make it reasonably doable for an apt programmer]

AOB

chris: Is Moz interested in the web midi api and implementing it

ehsan: no one currently woking on it and not sure if they would be interested

<olivier> Next Meeting: 5th December, same time

- DRAFT -

SV_MEETING_TITLE

21 Nov 2012

Attendees

Contents

Testing

Face-to-face meeting

Audio WG Document Milestones

Face-to-face meeting

state of Mozilla implementation

AOB

Summary of Action Items

Scribe.perl diagnostic output