SVG Working Group Teleconference

11 Oct 2012


See also: IRC log


birtles, [IPcaller], ed, +1.415.832.aaaa, krit1, Tav, Doug_Schepers, cabanier
cyril, heycam


<trackbot> Date: 11 October 2012

<krit1> ed: and heycam

while we're waiting, if you haven't filled out the telcon time page, please do (and make sure your UTC offset is correct for *after* nov 6)

<krit1> ed: where is this page?



if you're looking for UTC offsets, see here e.g: http://timeanddate.com/worldclock/meetingdetails.html?year=2012&month=11&day=8&hour=19&min=0&sec=0&p1=195&p2=224&p3=207&p4=240&p5=248

so, 21 UTC looks pretty ok


proposal: http://timeanddate.com/worldclock/meetingdetails.html?year=2012&month=11&day=8&hour=21&min=0&sec=0&p1=195&p2=224&p3=207&p4=240&p5=248

RESOLUTION: we'll have the telcon thursdays 21.00-22.30 UTC starting 8 November 2012

<scribe> scribeNick: ed

TPAC telcon (nov 1)

dirk: i won't be at tpac thu-fri

ds: i'll be there

ED: let's cancel the nov 1 telcon then

Filter Effects 1.0 publication status

ed: ready to be published, transition request approved

<scribe> ... pending pubrules checks and actual publication

dirk: think the spec is in good shape

radialGradient @fr


dirk: not very controversial, can fr be larger than r?
... think it should be able to
... canvas allows it

tav: didn't we already decide to follow canvas?

ed: did you find any content relying on the old behaviour (apart from the testsuite)?

dirk: one comment from dr olaf hoffmann
... and maybe there was something from dailey too

ed: so, in the webkit impl now there's no clamping, just the same as in canvas?

dirk: correct

ed: do we want to move the testcase in the testsuite back to draft status, or just leave it as-is?

dirk: let's leave it and then make new testcases instead

ed: will check if we have any content relying on the old behaviour internally

css clip property

dirk: csswg resolved to move clip to the css masking spec
... clip is different in SVG 1.1, due to reference to CSS2.0
... the behaviour for auto is different, we should follow css2.1

ED: we've already decided to follow css2.1 I think, so that should be fine
... my experience with 'clip' in svg content is that it's unreliable across browsers

dirk: it's also limited to applying to elemetns that establish viewports
... if csswg wants to keep it then maybe it should apply to more elements
... but if they want to deprecate it, then leave it more or less as it is

ed: there's a lack of tests for 'clip' in the svg testsuite

dirk: we should start writing more tests for svg
... can we resolve that clip: auto has no effect in svg?

ed: does it have an effect in html?

dirk: no, no effect

ed: ok, in that case svg should do the same (that is have no effect)

RESOLUTION: clip: 'auto' should have no effect in svg2

removing masking & clipping from svg2 spec

dirk: can i remove the parts in svg2, and reference css masking instead?
... like we do for filter effects

rik: sounds reasonable

ed: as long as it keeps backwards compat that should be fine

tav: should have a short paragraph of text for masking/clipping in svg2 still

dirk: ok, i can do that

RESOLUTION: svg2 will reference CSS Masking


tav: we talked about ports being defined in objbbox space
... should be defined in userspace so that if a rect gets skewed the ports move with the shape
... so that if you transform the shape the ports get transformed too

shepazu: was thinking precentages as being defined in css, relative to bbox

dirk: that's a different thing
... i'd agree with tav

shepazu: if i said something was 50% x and 100% y, and it was in a rect, i'd expect that to be in the 50%,100% bbox of the shape not of the parent

tav: ok, but if you skew the rect, so the points move with the sides

shepazu: if you say that's what it means that's fine
... what would you expect if i had a rect 100x50
... and a point inside that, x,y = 50,50
... where would you expect that to be, if the rect's x,y was 100,100

tav: the center of the bottom of the box

<shepazu> <rect x=100 y=100 width=100 height=50><point x=50 y=50/></rect>

shepazu: in the example i just pasted in, you'd expect the point in abs coords 150,150?

tav: yes

shepazu: that's what i'd want

tav: the pretransformed bbox
... the connector shouldn't be defined in terms of the object
... they don't belong to the object

<krit1> shepazu: do you mean 150,125?

shepazu: they do, but it's different

RESOLUTION: points inside shape elements are in the coordinate space of the shape element

tav: shepazu you're working on the connectors spec now?

shepazu: yes, cleaned up on the flight back

wpd (svgwow)

shepazu: svg wow demos?

ed: had a plan to move the published content to github, will try to do that soon
... should make it easier for ppl


dirk: svg parts are lacking polish

shepazu: please help out with that :)

dirk: it's hard to find things

shepazu: it will get better

trackbot, end telcon

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.136 (CVS log)
$Date: 2011-05-12 12:01:43 $