Re: Agenda, 1 March 2012 SVG WG telcon

Minutes from today's SVG WG telcon are here:

   http://www.w3.org/2012/03/01-svg-minutes.html

and below as text.


    [1]W3C

       [1] http://www.w3.org/

                                - DRAFT -

                    SVG Working Group Teleconference

01 Mar 2012

    [2]Agenda

       [2] 
http://lists.w3.org/Archives/Public/public-svg-wg/2012JanMar/0139.html

    See also: [3]IRC log

       [3] http://www.w3.org/2012/03/01-svg-irc

Attendees

    Present
           plinss, cabanier, [IPcaller], heycam, krit, ed, [Microsoft],
           Tav, cyril, Doug_Schepers

    Regrets
    Chair
           ed

    Scribe
           cabanier, shepazu

Contents

      * [4]Topics
          1. [5]the animation sandwich model/presentation attributes
          2. [6]testing requirements
          3. [7]CSS3 Transforms WD does not cover SVG changes
      * [8]Summary of Action Items
      _________________________________________________________

    <trackbot> Date: 01 March 2012

    <cabanier> scribenick: cabanier

the animation sandwich model/presentation attributes

    <krit>
    [9]http://lists.w3.org/Archives/Public/www-svg/2012Feb/att-0083/sand
    wich-svg.svg

       [9] 
http://lists.w3.org/Archives/Public/www-svg/2012Feb/att-0083/sandwich-svg.svg

    <ed>
    [10]http://lists.w3.org/Archives/Public/www-svg/2012Feb/0083.html

      [10] http://lists.w3.org/Archives/Public/www-svg/2012Feb/0083.html

    dirk: the question: should SVG be on top of CSS animation or the
    other way around

    cameron: can you summarize?

    dirk: in the model you see the CSS change the intrinsic style? How
    do you apply SVG animation on top of that?

    cabanier: Brian, Shane and I have been having talks about how to
    progress animations

    <ed>
    [11]http://lists.w3.org/Archives/Public/www-svg/2012Feb/0084.html
    says that the CSS animations overrides in gecko

      [11] http://lists.w3.org/Archives/Public/www-svg/2012Feb/0084.html

    cabanier: we were thinking that maybe there should be the ability to
    specify the order that animations are applied

    cameron: you mean that the order lets you mix svg and css animation?

    cabanier: yes
    ... that's an idea we discussed

    cameron: with the sandwich model you look at document order

    dirk: yes. if the animations are applied on the same time, then
    document is chosen

    cameron: so is the proposal for both SVG and CSS

    cabanier: yes

    dirk: mozilla currently divides them
    ... the problem with specifiyng an order is that they run into
    different time container

    cabanier: yes, we're aware of that
    ... it's not a complete proposal at this point.
    ... just something we were thinking about this week
    ... we're hoping to have a proposal by the Hamburg F2F

    dirk: I believe that will be too late.
    ... do you have something in the mean time that works?

    cabanier: I think what you propose now is reasonable

    cameron: I think I'm OK with that as well. So you can specify
    behavior if you don't have the new keyword
    ... that way we're also not held up by the new css spec

    dirk: In the mean time, shoud the CSS animation mention SMIL

    cameron/cabanier: no

    cameron: are the notes of the animation meeting online?

    cabanier: yes, they are on the w3 server

    meeting notes:
    [12]http://www.w3.org/Graphics/fx/wiki/Web_Animations/Meetings

      [12] http://www.w3.org/Graphics/fx/wiki/Web_Animations/Meetings

    cabanier: overview:
    [13]http://www.w3.org/Graphics/fx/wiki/Web_Animations

      [13] http://www.w3.org/Graphics/fx/wiki/Web_Animations

    ed: do we have a resolution

    dirk: should we say that CSS animation don't mention SMIL?

    cabanier: yes
    ... and your proposed order is reasonable

    dirk: I believe it's OK to leave that unspecified at the moment

    cameron: so how do other browsers than firefox do this?

    dirk: we just landed a path to match WebKit with FireFox

    cameron: and opera?

    <heycam> ed,
    [14]http://people.mozilla.org/~dholbert/tests/smil/compat_tests/tran
    sitionsVsSMIL_v1.svg

      [14] 
http://people.mozilla.org/~dholbert/tests/smil/compat_tests/transitionsVsSMIL_v1.svg

    ed: I don't know. Is there a test?

    dirk: yes, Daniel Holbert supplied one

    ed: tested in Opera 12, seems that SMIL wins over the css
    transitions

    cameron: It seems that we're in a situation that the public versions
    are different

    cabanier: if everyone converging, shouldn't that become the
    standard?

    cameron/ed/Dirk: that would be OK for me

    cameron: to be sure: this is only when CSS and SMIL start at the
    same time
    ... is it ignoring the start times of the SVG or the CSS?
    ... if there are CSS and SVG, SVG will always come second and CSS
    first
    ... regardless of start time

    <scribe> ACTION: cabanier to come up with a solution to specify
    ordering of animations [recorded in
    [15]http://www.w3.org/2012/03/01-svg-minutes.html#action01]

    <trackbot> Created ACTION-3241 - Come up with a solution to specify
    ordering of animations [on Rik Cabanier - due 2012-03-08].

testing requirements

    <ed>
    [16]http://www.w3.org/Graphics/SVG/WG/wiki/SVG2/Testing_Requirements

      [16] http://www.w3.org/Graphics/SVG/WG/wiki/SVG2/Testing_Requirements

    dirk: for SMIL animations, we need JavaScript

    ed: I agree. I have an action to commit some examples of converted
    SVG 1.1 animation tests using the js testing framework
    ... this is something we do internally

    cameron: we set the current animation time and look at the current
    values
    ... tav has an action on this
    ... is this easily doable or does it require a lot of work?

    tav: currently it is set up to test for environments that don't have
    javascript
    ... we like the test to have links so you can link to the spec (and
    vice versa)
    ... (going over the test suite)
    ... we have limited tests for CSS
    ... you have tests that require pixel perfect result
    ... for certain tests, we don't require that (ie gradients)

    cameron: last time we looked, the automation wasn't there

    dirk: you want to have it cross browser
    ... can you use the canvas tag/

    plinss: some people have looked into that

    shepazu: hixie is looking into ways to draw svg paths directly into
    canvas
    ... you could draw the same SVG and canvas using the same API
    ... for a vast majority of content

    <krit>
    [17]http://www.whatwg.org/specs/web-apps/current-work/multipage/the-
    canvas-element.html

      [17] 
http://www.whatwg.org/specs/web-apps/current-work/multipage/the-canvas-element.html

    <heycam> This is with drawImage

    cameron: it works as long as you don't have external references

    cabanier: I believe it's the same in webkit

    dirk: I have to check that

    cameron: automation would be quite beneficial

    plinss: you could set it up so that you can say that a test must
    matchs all references or any
    ... so you can have browser specific reference files

    cameron: how are css animations tested?

    plinss: we don't have any

    ed: do you want to add more to the wiki page tav?
    ... for example, that the css testing framework documentation said
    you can't use plain SVG?

    tav: we can look at the script and see if you can use plain svg

    plinss: that should not be a problem

    ed: the metadata was supposed to be a HTML

    plinss: the harnass itself doesn't really care where the data comes
    from

    tav: can you extract the metadata from the SVG itself

    pllins: if it's XML, we open it and look HTML namespace head element
    inside of it

    plinsss: I parse the file and create the DOM and look for the head
    tag and expect the metadata to be there

    tav: right now it's all in the namespace "d"

    <ed> here's an example:
    [18]http://dev.w3.org/SVG/profiles/1.1F2/test/svg/animate-dom-01-f.s
    vg

      [18] 
http://dev.w3.org/SVG/profiles/1.1F2/test/svg/animate-dom-01-f.svg

    plinss: if you use it consistently, it would be trivial to add to
    the build system

    tav: how would I find this?

    plinss: there is no good documentation on how to set it up

    <plinss> [19]http://wiki.csswg.org/test/css2.1/harness

      [19] http://wiki.csswg.org/test/css2.1/harness

    plinss: this is our wiki that has everything you need

    <plinss> [20]http://hg.csswg.org/test

      [20] http://hg.csswg.org/test

    <plinss> [21]http://hg.csswg.org/test/file/tip/tools

      [21] http://hg.csswg.org/test/file/tip/tools

    <plinss> [22]http://hg.csswg.org/dev/w3ctestlib/

      [22] http://hg.csswg.org/dev/w3ctestlib/

    ed: what would you recommend, going with the html:head structure, or
    keep and extend the existing svg test metadata?

    plinss: whatever is convenient
    ... it would be easiest if the metadata itself is the same

    <plinss> [23]http://wiki.csswg.org/test/format

      [23] http://wiki.csswg.org/test/format

    plinss: what you wrap it in is not relevant
    ... but it would be easiest if the data itself is similar

    cameron: we should be able to do it

    <plinss> [24]http://test.csswg.org/shepherd/

      [24] http://test.csswg.org/shepherd/

    plinss: it's like a bugzilla for your testsuite

    cameron: pretty cool

    plinss: it shows your revision history, who approved the tests,
    flags tests as needing work, etc
    ... it's still in development but you can use it now

    tav: where would the SVG test suite be located?

    plinss: you set up your own repository and shepherd will refer to it

    <heycam> there's an empty repo here:
    [25]http://svgwg.org/hg/svg2-tests/

      [25] http://svgwg.org/hg/svg2-tests/

    ed: we already have a svg2-tests repository

    <plinss> [26]http://w3c-test.org/framework/

      [26] http://w3c-test.org/framework/

    <RikCabanier> (discussion on using test repository)

    <plinss> [27]http://hg.csswg.org/test

      [27] http://hg.csswg.org/test

    <RikCabanier> ed: I propose that we move to this structure

    <RikCabanier> cameron: I agree

    <RikCabanier> ACTION: cameron to create directories in the SVG2 test
    repository [recorded in
    [28]http://www.w3.org/2012/03/01-svg-minutes.html#action02]

    <trackbot> Created ACTION-3242 - Create directories in the SVG2 test
    repository [on Cameron McCormack - due 2012-03-08].

    <RikCabanier> plinss: within the approved directory, there's a data
    and src directory

    <RikCabanier> plinss: src has the test and data contains the spec
    section manifest files

    <RikCabanier> ACTION: tav to mirror css test metadata setup in SVG
    and come up with a template for new tests [recorded in
    [29]http://www.w3.org/2012/03/01-svg-minutes.html#action03]

    <trackbot> Created ACTION-3243 - Mirror css test metadata setup in
    SVG and come up with a template for new tests [on Tavmjong Bah - due
    2012-03-08].

    <krit> [30]http://www.w3.org/TR/css3-transforms/

      [30] http://www.w3.org/TR/css3-transforms/

    <krit> [31]http://dev.w3.org/csswg/css3-transforms/

      [31] http://dev.w3.org/csswg/css3-transforms/

CSS3 Transforms WD does not cover SVG changes

    <RikCabanier> dirk: since the published draft is wrong, I need to
    ask everyone to approve a new working draft

    <RikCabanier> cyril: I don't see a difference

    <RikCabanier> dirk: look at chapter 7

    <RikCabanier> cyril: o I see

    <RikCabanier> cameron: what is the current thinking on adding the
    SVG to the CSS spec?

    <RikCabanier> cameron: are people happy to include it in the spec?

    <RikCabanier> dirk: absolutely

    <RikCabanier> dirk: there are concerns that it slows down, but I'm
    committed to advancing it quickly

    <RikCabanier> dirk: we can go to CR in May

    <ed> [32]http://dev.w3.org/csswg/css3-transforms/#svg-transform

      [32] http://dev.w3.org/csswg/css3-transforms/#svg-transform

    <RikCabanier> ACTION: cameron to review section 7 in new CSS
    transform spec [recorded in
    [33]http://www.w3.org/2012/03/01-svg-minutes.html#action04]

    <trackbot> Created ACTION-3244 - Review section 7 in new CSS
    transform spec [on Cameron McCormack - due 2012-03-08].

    <RikCabanier> shepazu: sorry, CR in May seems arbitrary

    <RikCabanier> dirk: the CSS working group will bring up issues in
    the meeting

    <RikCabanier> dirk: and if there are none, it will be accepted in
    may

    <RikCabanier> shepazu: when is last call supposed to be?

    <cyril> especially when it says "The merge is in progress and the
    specification is not yet ready for review."

    <RikCabanier> plinss: that is what we want too. If we drop prefixes
    early, we want to see proof that the implementations are identical

    <shepazu> scribenick: shepazu

    shepazu: I want to make sure that we aren't putting the cart before
    the horse… dropping vendor prefixes should be done after there are
    functional signs of stability and interoperability, like a full test
    suite and implementation report, not an arbitrary point of progress
    along the Rec track

    plinss: yes, that's also one of our criteria… but we do want to go
    to CR for the patent commitments

    shepazu: ok, I agree with that

    trackbot, end telcon

Summary of Action Items

    [NEW] ACTION: cabanier to come up with a solution to specify
    ordering of animations [recorded in
    [34]http://www.w3.org/2012/03/01-svg-minutes.html#action01]
    [NEW] ACTION: cameron to create directories in the SVG2 test
    repository [recorded in
    [35]http://www.w3.org/2012/03/01-svg-minutes.html#action02]
    [NEW] ACTION: cameron to review section 7 in new CSS transform spec
    [recorded in
    [36]http://www.w3.org/2012/03/01-svg-minutes.html#action04]
    [NEW] ACTION: tav to mirror css test metadata setup in SVG and come
    up with a template for new tests [recorded in
    [37]http://www.w3.org/2012/03/01-svg-minutes.html#action03]

    [End of minutes]

Received on Thursday, 1 March 2012 21:47:16 UTC