Minutes, 3 May 2010 FX telcon

Hello public-fx,

In glorious stereo

and scratchy mono

                   CSS-SVG Task Force Teleconference

03 May 2010


      [2] http://lists.w3.org/Archives/Public/public-fx/2010AprJun/0034.html

   See also: [3]IRC log

      [3] http://www.w3.org/2010/05/03-fx-irc


          Shepazu, ed, [IPcaller], anthony, +39.524.9.aaaa, ChrisL,
          smfr, +1.617.584.aabb, fantasai, +1.858.655.aacc, plinss_,
          al_macdonald, dbaron

          Tab, Dean




     * [4]Topics
         1. [5]Intros
         2. [6]'image-fit' vs preserveAspectRatio
         3. [7]transitions of CSS images (including gradients)
         4. [8]svg/css merged transforms spec
         5. [9]harmonizing CSSMatrix and SVGMatrix (2D)
         6. [10]next telcon
     * [11]Summary of Action Items

   <trackbot> Date: 03 May 2010

   <smfr> i can't get dialin to work

   <shepazu> smfr: the conference code is 26632

   <smfr> ah, was looking at the email

   <f1lt3r> hey, im in

   code today is 26632

   <scribe> scribe: Chris

   <scribe> ScribeNick: ChrisL


   (people introduce themselves)

   <smfr> is al on irc?

'image-fit' vs preserveAspectRatio


     [12] http://www.w3.org/mid/25829533.20100331230050@w3.org

   <fantasai> ChrisL: There's been a bit of to-and-fro about the naming
   of these properties

   object-fit: fill | cover | contain

   object-position: <bg-pos>


   <fantasai> ChrisL: That's the most recent proposal from the CSSWG

   <f1lt3r> i am al :)

   elika: bg-pos allows you to position an image in a box

   consistent with background-position in CSS

   e.g. top left, 20% 80%

   doug: depencdency on box model?

   elika: none
   ... its just the names

   erik: read minutes, seem to have had 4 proposals for naming
   ... would be fine with object-fit and object-position
   ... liked others better but fine with this one
   ... fit-position would have been first choice

   elika: proposals were in pairs, so what would the other property be

   erik: fit-scale. ... ok not perfect either

   doug: objec t-fit ok with me

   chris: ok with these names too

   (no objections to proposed names)

   resolved: svg will accept the css proposal of object-fit and

   <smfr> sweeping bird poop, by the sound of it

   <fantasai> ACTION fantasai edit css3-images with resolution

   <trackbot> Created ACTION-3 - Edit css3-images with resolution [on
   Elika Etemad - due 2010-05-10].

transitions of CSS images (including gradients)

   simon: this is my proposal. web authors can animate links and
   colours but can't do cross fades between images and especiall


     [13] http://lists.w3.org/Archives/Public/public-fx/2010AprJun/0007.html

   simon: new functional notation lets you do wipe or grossfade, or
   opocity or a filter
   ... its making a synthetic image that can be used anywhere in css
   ... gradient interoplation is kind of a side issue
   ... if accepted, will write a proposal

   elika: could go in the images draft

   doug: loks good so far, would like to see it in the draft to see it
   in context

   <F1LT3R_> could i get that link again?

   simon: infinite range of effects, so how to bound it
   ... could use svg filters
   ... need a set of wipes/fades

   erik: could do that using filters too


     [14] http://www.w3.org/TR/2005/REC-SMIL2-20050107/smil-transitions.html#TransitionEffects-Appendix

   anthony; some video editing apps have a set of crossfade effects

   chris: smil has a set from video editing

   simon: too big a list

   anthon: canned list is very extensive

   simon: can have canned effects but also a fiunction to transform an

   <ed> erik: mapping more than one input image from css syntax in
   terms of svg filters is possible, if you want to define it that way
   (e.g several feImage elements, or new filter input image-keywords)

   doug: canned effects as syntactic sugar

   anthony: so zoom into images?

   simon: yes, and rotates. wipe is a translation

   chris: like them all defined and then shorthand for common ones

   anthony: need to describe the underlying rendering model. all
   rendered to an image. need to be clear on how it composites

   simon: risk of over complex functions which are hard to author

   doug: if the primitives are the underlying part we can optimise

   simon: suppose you have a high-level push. but a more complext one
   would have three or four levels of nesting

   chris: ok to make the common things easy and the complex things

   doug: need some good examles, especially if there are things that
   will be outside what we plan to do

   <scribe> ACTION: simon to draft examples for transitions of css
   images on a wiki [recorded in

   <trackbot> Created ACTION-4 - Draft examples for transitions of css
   images on a wiki [on Simon Fraser - due 2010-05-10].

   doug: if its on a wiki we can all contribute

   action simon to implement the whole web

   <trackbot> Created ACTION-5 - Implement the whole web [on Simon
   Fraser - due 2010-05-10].

   close action-5

   <trackbot> ACTION-5 Implement the whole web closed

   <smfr> heh

   <smfr> phew

svg/css merged transforms spec

   erik: any progress on that?

   anthony: committed a merged spec but its very vasic and needs work.
   need help with it

   <smfr> anthony: is it this? [16]http://www.w3.org/TR/SVG-Transforms/

     [16] http://www.w3.org/TR/SVG-Transforms/

   anthony: need somewhere to put it

   action chris to make an fx area on dev.w3.org

   <trackbot> Created ACTION-6 - Make an fx area on dev.w3.org [on
   Chris Lilley - due 2010-05-10].

   anthony: no, starting with the 2d ones

   action doug to ask for an FX wiki

   <trackbot> Created ACTION-7 - Ask for an FX wiki [on Doug Schepers -
   due 2010-05-10].

   <shepazu> [17]http://www.w3.org/Graphics/fx/track/

     [17] http://www.w3.org/Graphics/fx/track/

harmonizing CSSMatrix and SVGMatrix (2D)

   simon: css matrixc was wrong and svg was right so we have corrected


     [18] http://lists.w3.org/Archives/Public/public-fx/2010AprJun/0022.html

   simon: the way we implemented css matrix multioply, needed to do a
   mult right not mult left
   ... that change is now in webkit. no other changes are needed.
   ... we use row major because C++ does that
   ... spec needs to be updated. removed multiply-left
   ... pseudocode is in c/c++ style, need to be more explicit to be
   column major

   anthony: doesn;t matter on internal format as long as spec gives
   right results

   <dbaron> Talking about left and right is a bit painful because they
   have opposite meanings with row-major and column-major

   anthony: ok with actual matrix examples. should clarify that
   internal format is implementor chpice

   chris: agree with david

   david; cairo and mozilla do it the opposite way round for example

   <dbaron> cairo and the spec, anyway

   <dbaron> and in some places mozilla follows the spec and in some
   follows cairo

   erik; dom interfaces have access properties, will these be the same?

   simon: yes
   ... so spec will be updated to use the svg notation
   ... remaining piece is decomposition pseudocode, - can leave it and
   say it refers to coloumn-major, or replace it with the actual math

   chris: ok to keep pseudocode as long as its prefaced by an
   explanation of the two ways to do it

   actiuon: simon add explanatory material about row and column major
   options to spec

   actiion: simon add explanatory material about row and column major
   options to spec

   <scribe> ACTION: simon add explanatory material about row and column
   major options to spec [recorded in

   <trackbot> Created ACTION-8 - Add explanatory material about row and
   column major options to spec [on Simon Fraser - due 2010-05-10].

   <ed> s/dom interfaces have access properties/the SVG/CSSMatrix dom
   interfaces have access properties [a-f] for matrix values/

next telcon

   <shepazu> [20]http://www.w3.org/Graphics/fx/wiki/Main_Page

     [20] http://www.w3.org/Graphics/fx/wiki/Main_Page

   simon: prefer biweekly andf 60 minutes

   <smfr> "The URL path in your request doesn't match anything we have

   (general agreement on 60 minutes biweekly)

   <scribe> ACTION: chris to book a regular biweekly telcon [recorded
   in [21]http://www.w3.org/2010/05/03-fx-minutes.html#action03]

   <trackbot> Created ACTION-9 - Book a regular biweekly telcon [on
   Chris Lilley - due 2010-05-10].

   elika: simin mentioned a colour function, what is it for?

   david: for cross fade t a solid colour

   elika: why not just put a colour value directly?

   simon: ambiguity in background shorthand
   ... will see if we can do just a colour

   elika: or wrap in image notation

   doug: will transitions also work with svg, withing svg and also for
   svg images used as backgrounds in html?

   simon: will work on animatable css properties, and will work with
   svg images same as any other image

   erik: agree

   doug: does the spec make it clear that this is allowed?


Summary of Action Items

   [NEW] ACTION: chris to book a regular biweekly telcon [recorded in
   [NEW] ACTION: simon add explanatory material about row and column
   major options to spec [recorded in
   [NEW] ACTION: simon to draft examples for transitions of css images
   on a wiki [recorded in

   [End of minutes]

 Chris Lilley                    mailto:chris@w3.org
 Technical Director, Interaction Domain
 W3C Graphics Activity Lead
 Co-Chair, W3C Hypertext CG

Received on Monday, 3 May 2010 21:14:52 UTC