See also: IRC log
<trackbot> Date: 14 March 2011
<scribe> scribenick: smfr
<scribe> Scribenick: smfr
heycam: summarizes css transitoins and animations applying to svg attributes
first proposal: turn all attributes into properties with the same name
(aussies all sound the same)
controvecial to convert all attributes to properties
may end up with property name conflicts; css properties are global, so property with name 'r' is not acceptable
second proposal: new properties for attributes, but use different names
e.g. r -> circle-radius
<heycam> <circle circle-radius="100"/>
but this breaks correspondence between properties and presentation attributes, so confusing
would need to define how e.g. circle-radius and r would interact
third option: add extra syntax to css transitions/animation to allow targeting attributes
SVG WG preferred the functional syntax in the conclusion section
wouldn't be able to transition svg attributes directly in css, because you can't use attr(x) on the LHS outside the keyframes
smfr: there should be nothing
special about css parsing inside of keyframes, so i don't like
... functional syntax on LHS is weird
TabAtkins: points to dean's proposal in email
heycam: strange to extract special meaning out of "-attr-" identifiers on the LHS
heycam: prefer the second
... it's more of a wholesale change to svg to make a bunch of attributes into newly named properties
TabAtkins: this proposal seems
similar to -attr-r
... attribute-targeting as a magic hack is untenable if general enough for transitions as well as animations
... OK with option 2 or -attr-
smfr: -attr may conflict with vendor prefixes
TabAtkins: could use attr- and claim it as a unique prefix
anthony_work: can you use this to apply style to content?
... like setting HTML attributes which map to CSS
like the normal cascade; the presentational attribute loses
(who just spoke?)
<cabanier> is 'display' animatable?
TabAtkins: we talked about all properties being animatable
smfr: i was going to try it and see if there were perf issues
TabAtkins: there's an issue to change transiton-property default to 'none'
heycam: does it make sense to
target every attribute, or only animatable ones?
... don't want to modify the DOM attribute property value on animations
TabAtkins: say that only subset of attributes are mapped to properties
heycam: limit to simple values;
lengths, numbers etc
... we've heard concrete arguments about option 3
... haven't heard technical arguments against property promotion idea
ChrisL: CSS WG doesn't like
adding lots of properties, and element-specific
... adding cx and cy properties that only apply to circles would not be liked
heycam: don't want to expand out property tables on every element
smfr: how many properties are we talking about?
heycam: about 45 new
... roc suggested that rather than adding e.g. circle-radius and ellipse-x-radius, see where properties are similar and merge them
TabAtkins: but this list has cx,
x, rx so this list hasn't been folded
... if implementations can only use memory for SVG elements, then the number of new properties shouldn't be an issue
heycam: was this the main objection?
TabAtkins: do we want to go with the attr- prefix or more specialized property names
heycam: maybe don't need to
... x can be a single value on rect, but a list on text
... so need to be careful with possibly conflicting values
<ed> dx,dy has the same issue (for filters vs text elements)
TabAtkins: mostly an issue for the spec, not authors
heycam: if we go the attr- route then we could not create new presentation attributes on the elements
TabAtkins: wouldn't add new attributes even with the specialized names
heycam: what to do with SVG DOM obects like SVGAnimatedLength
TabAtkins: issue seeing the animating value from JS
heycam: SVG is interested in something easier to use than SVGAnimatedLength
<scribe> ACTION: take attr- proposal to CSS WG [recorded in http://www.w3.org/2011/03/14-fx-minutes.html#action01]
<trackbot> Sorry, couldn't find user - take
<scribe> ACTION: TabAtkins take attr- proposal to CSS WG [recorded in http://www.w3.org/2011/03/14-fx-minutes.html#action02]
<trackbot> Sorry, couldn't find user - TabAtkins
<scribe> ACTION: Tab take attr- proposal to CSS WG [recorded in http://www.w3.org/2011/03/14-fx-minutes.html#action03]
<trackbot> Sorry, couldn't find user - Tab
<heycam> trackbot, users?
<trackbot> Sorry, heycam, I don't understand 'trackbot, users?'. Please refer to http://www.w3.org/2005/06/tracker/irc for help
<heycam> trackbot, who do you know about?
<trackbot> Sorry, heycam, I don't understand 'trackbot, who do you know about?'. Please refer to http://www.w3.org/2005/06/tracker/irc for help
<ChrisL> trackbot, status
<scribe> ACTION: cameron to take attr- proposal to SVG WG [recorded in http://www.w3.org/2011/03/14-fx-minutes.html#action04]
<trackbot> Created ACTION-26 - Take attr- proposal to SVG WG [on Cameron McCormack - due 2011-03-21].
anthony_work: made a TODO list
heycam: how do we proceed through this list?
smfr: take to email, the discuss things we need to
anthony_work: if you see issues that are easy to address, go in and edit the wiki page or let me know
smfr: what is "add IDL"
... issue is that CSS OM is in flux, and the IDL that used to be there relied on something like CSSValues that were removed?
heycam: does transform become attr-transform?
anthony_work: was assuming it
would stay as transform
... what does this mean for the SVG animation interfaces?
... some content uses them, we can't just drop them
heycam: we should discuss what
happens with SVG DOM, and if it changes we will need IDL for
... wiki should make it clear that we won't add the CSS OM interfaces, just the SVG ones
... SVG DOM interfaces could just be defined in SVG, and not in this spec
anthony_work: i like that
ed: i would like to see css2d transforms interfaces, e.g CSSMatrix interface compatible with svg objects
smfr: we should strive to make a
single matrix class
... issue with 2d vs. 3d transform matrix
TabAtkins: we should start with a 4x4 matrix
smfr: you have to say what happens when you apply a 3d matrix to something that can only render 2d
TabAtkins: you can specify what happens (flatten or whatever)
anthony_work: would prefer the transform doesn't apply, to avoid behavioral differences
smfr describes issues around supporting 3d css properties but not being able to render 3d
<ed> trackbot, end telcon
This is scribe.perl Revision: 1.135 of Date: 2009/03/02 03:52:20 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/properties/attributes/ Succeeded: s/anthony_work:/heycam:/ Succeeded: s/c-x and c-y/cx and cy/ Succeeded: s/css matrix/css2d transforms interfaces, e.g CSSMatrix/ Found ScribeNick: smfr Found ScribeNick: smfr Inferring Scribes: smfr Present: heycam ed smfr rik_cabanier hober anthony TabAtkins CL Agenda: http://lists.w3.org/Archives/Public/public-fx/2011JanMar/0124.html Found Date: 14 Mar 2011 Guessing minutes URL: http://www.w3.org/2011/03/14-fx-minutes.html People with action items: attr- cameron proposal tab tabatkins take WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option.[End of scribe.perl diagnostic output]