IRC log of svg on 2009-02-26

Timestamps are in UTC.

19:26:28 [RRSAgent]
RRSAgent has joined #svg
19:26:28 [RRSAgent]
logging to http://www.w3.org/2009/02/26-svg-irc
19:26:30 [trackbot]
RRSAgent, make logs public
19:26:32 [trackbot]
Zakim, this will be GA_SVGWG
19:26:32 [Zakim]
ok, trackbot; I see GA_SVGWG()2:30PM scheduled to start in 4 minutes
19:26:33 [trackbot]
Meeting: SVG Working Group Teleconference
19:26:33 [trackbot]
Date: 26 February 2009
19:29:56 [ChrisL]
ChrisL has joined #svg
19:30:23 [ed__]
ed__ has joined #svg
19:30:27 [Zakim]
GA_SVGWG()2:30PM has now started
19:30:28 [Zakim]
+Shepazu
19:31:33 [Zakim]
+??P8
19:31:40 [Zakim]
+??P13
19:31:43 [heycam]
Zakim, ??P13 is me
19:31:43 [Zakim]
+heycam; got it
19:31:43 [ed__]
Zakim, ??P8 is me
19:31:44 [Zakim]
+ed__; got it
19:32:43 [Zakim]
+??P14
19:32:56 [ChrisL]
zakim, who is here?
19:32:56 [Zakim]
On the phone I see Shepazu, ed__, heycam, ??P14
19:32:57 [Zakim]
On IRC I see ed__, ChrisL, RRSAgent, Zakim, heycam, anthony, shepazu, ed_work, trackbot
19:33:18 [Zakim]
-heycam
19:33:23 [Zakim]
-??P14
19:33:35 [ChrisL]
present+ a dalek
19:33:44 [Zakim]
+??P2
19:33:46 [heycam]
Zakim, ??P2 is me
19:33:46 [Zakim]
+heycam; got it
19:33:59 [ChrisL]
rrsagent, here
19:33:59 [RRSAgent]
See http://www.w3.org/2009/02/26-svg-irc#T19-33-59
19:34:15 [Zakim]
+??P5
19:34:53 [ChrisL]
zakim, ?/p5 is me
19:34:53 [Zakim]
sorry, ChrisL, I do not recognize a party named '?/p5'
19:35:02 [ChrisL]
zakim, ??p5 is me
19:35:02 [Zakim]
+ChrisL; got it
19:35:38 [ed__]
Zakim, who's here?
19:35:38 [Zakim]
On the phone I see Shepazu, ed__, heycam, ChrisL
19:35:39 [Zakim]
On IRC I see ed__, ChrisL, RRSAgent, Zakim, heycam, anthony, shepazu, ed_work, trackbot
19:36:44 [ChrisL]
zakim, who all is here, y'all
19:36:44 [Zakim]
I don't understand 'who all is here, y'all', ChrisL
19:37:15 [Zakim]
+[IPcaller]
19:37:37 [anthony]
Zakim, [IPcaller is me
19:37:37 [Zakim]
+anthony; got it
19:37:54 [ed__]
Zakim, pick a victim
19:37:54 [Zakim]
Not knowing who is chairing or who scribed recently, I propose ChrisL
19:38:00 [ChrisL]
scribe: chris
19:38:11 [ChrisL]
scribenick: chrisl
19:38:11 [ed__]
Agenda: http://lists.w3.org/Archives/Public/public-svg-wg/2009JanMar/0172.html
19:38:17 [shepazu]
zakim, pick a victim
19:38:17 [Zakim]
Not knowing who is chairing or who scribed recently, I propose ed__
19:38:37 [ChrisL]
Topic: Referencing ISO-32000-1 for blending [Compositing module]
19:39:42 [ChrisL]
cl: is this a normative or informative reference, and will the equations be in our spec of just referenced?
19:39:48 [ChrisL]
cmc: costs money?
19:39:51 [ChrisL]
cl: of course
19:40:13 [ChrisL]
ds: prefer to have the equations in our spec, can reference the ISO as well
19:40:23 [ChrisL]
... and better to be in a RF spec
19:40:54 [ChrisL]
cl: normative or informative?
19:41:08 [ChrisL]
ds: informative. Useful to know the connection between them
19:41:20 [ChrisL]
cl: agree
19:41:44 [ChrisL]
cmc: if its nortmative then you still should buy it in case
19:41:55 [ChrisL]
ed: good to have all the equations in our spec
19:42:31 [ChrisL]
resolved: equations for compositing will be in the svg spec, with an informative reference to ISO 3200-1
19:43:12 [ChrisL]
action: anthony to check equations for blending and add informative ref to ISO 32000-1
19:43:12 [trackbot]
Created ACTION-2480 - Check equations for blending and add informative ref to ISO 32000-1 [on Anthony Grasso - due 2009-03-05].
19:43:27 [ChrisL]
ag: will do, not checked the equations yet
19:44:24 [ChrisL]
... will check the equations again, needs some testing to verify
19:47:14 [ChrisL]
topic: transforms review
19:47:21 [anthony]
http://www.w3.org/Graphics/SVG/WG/wiki/CSS-Transforms-Review
19:47:50 [ChrisL]
http://dev.w3.org/cvsweb/~checkout~/csswg/css3-2d-transforms/Overview.html?rev=1.3&content-type=text/html;%20charset=iso-8859-1
19:49:38 [ChrisL]
cl: (explains stacking context)
19:49:48 [ChrisL]
ds: liked your email comments, chris
19:50:54 [shepazu]
s/ liked your email comments, chris/ liked your email comments, chris... they seem more likely to be fruitful/
19:50:54 [ChrisL]
before "Currently, Firefox and Safari throw away" say "Since backward compatibility is important we tested current browwsers and currently ...."
19:51:48 [ChrisL]
should probably motivate why openvg is important
19:52:46 [ChrisL]
existing implementations get acceleration from openvg, etc
19:53:31 [ChrisL]
cl: the overasll review document is good, I just had some specific comments
19:55:29 [ChrisL]
ds: its important that we support css in this useful work
19:55:33 [ChrisL]
ed: sounds good
19:55:52 [ChrisL]
ag: been editing live, these are ready to go
19:56:02 [ChrisL]
cmc: i have a few detailed comments in email
19:56:13 [ChrisL]
.. want clarification on matrix sizes and math
19:56:28 [ChrisL]
ag: need to commit the new draft of svg transforms
19:57:24 [ChrisL]
cl: the css spec needs some references to explain transforms and matrices for those unaware
20:00:32 [ChrisL]
ed: we have a minuted resolution to publish
20:00:55 [ChrisL]
ds: plh said recently that tight integration and strong coordination is important
20:02:59 [ChrisL]
ag: did we respond to dean's email?
20:03:20 [ChrisL]
cmc: send a reply to dean thanking him, and point to this review?
20:04:42 [ChrisL]
action: doug to reply to Dean on 3d transforms
20:04:42 [trackbot]
Created ACTION-2481 - Reply to Dean on 3d transforms [on Doug Schepers - due 2009-03-05].
20:05:31 [ChrisL]
referencing http://www.w3.org/Graphics/SVG/WG/wiki/CSS-Transforms-Review
20:06:13 [ChrisL]
cmc: i still have some comments yet to make
20:06:32 [ChrisL]
ds: ok lets talks offline and I will send it over the weekend
20:06:47 [ChrisL]
cl: css wg will likely be discussing these documents next week in japan
20:07:17 [ChrisL]
Topic: Transitions and Animations
20:07:48 [ChrisL]
cmc: so we are going to send outr individual non-svgf related stuff directly, and coordinate on a wg response for the svg specific aspects
20:09:23 [ChrisL]
ds: think individual responses is more timely and allows more discussion
20:10:15 [ChrisL]
... send responses to www-style with cc to www-svg
20:11:06 [ChrisL]
ed: i have a bunch of feedback on that, will send it next week
20:15:16 [ChrisL]
topic: z-index use cases and requirements
20:15:23 [ChrisL]
issue-2226?
20:15:23 [trackbot]
ISSUE-2226 -- Investigate the use cases and requirements for z-index -- RAISED
20:15:23 [trackbot]
http://www.w3.org/Graphics/SVG/WG/track/issues/2226
20:15:56 [ChrisL]
ed: good to have an action on this.
20:16:10 [ChrisL]
... does our spec deal with z-index?
20:16:19 [ChrisL]
ag: we have discussed it, in Sydney
20:16:50 [ChrisL]
http://www.w3.org/2009/02/15-svg-minutes.html#item05
20:18:39 [ChrisL]
pointer to stacking context http://www.w3.org/TR/CSS21/zindex.html
20:19:57 [ChrisL]
ed: we could define z-index the same way so that it also applies to svg as well as being compatible with html
20:20:14 [ChrisL]
'z-index'
20:20:15 [ChrisL]
Value: auto | <integer> | inherit
20:20:15 [ChrisL]
Initial: auto
20:20:15 [ChrisL]
Applies to: positioned elements
20:20:15 [ChrisL]
Inherited: no
20:20:15 [ChrisL]
Percentages: N/A
20:20:17 [ChrisL]
Media: visual
20:20:19 [ChrisL]
Computed value: as specified
20:21:10 [ChrisL]
ag: instead of z-index, could also use the transform-style property, to switch on z-ordering for a group of objects
20:22:16 [ChrisL]
cl: interesting option
20:22:28 [ChrisL]
ag: yes as it could apply only to container elements
20:22:34 [ChrisL]
... similar to layered groups
20:23:05 [ed___]
ed___ has joined #svg
20:23:32 [ed____]
ed____ has joined #svg
20:23:52 [Zakim]
-ed__
20:23:54 [ChrisL]
ag: as soon as document order is lost, impact on streaming and rendering speed so being able to isolate the effects is beneficial
20:24:25 [ChrisL]
ds: would that need a whole shadow tree?
20:24:30 [Zakim]
+??P1
20:24:30 [ed_____]
ed_____ has joined #svg
20:24:52 [ChrisL]
zakim, ??p1 is erik
20:24:52 [Zakim]
+erik; got it
20:25:34 [Zakim]
-erik
20:25:40 [ChrisL]
ds: if we have layered groups, it could be implemented by making a shadow tree that replaces the docuent order subtree. but that could be expensive on memory
20:26:26 [ed___]
ed___ has joined #svg
20:26:39 [ChrisL]
cl: impact on filter effects / render background, and on event bubbling .....
20:26:54 [Zakim]
+??P1
20:26:57 [ChrisL]
cmc: would complicate document updates as well
20:27:02 [ChrisL]
zakim, ??p1 is erik
20:27:02 [Zakim]
+erik; got it
20:27:19 [ed___]
rrsagent, pointer?
20:27:19 [RRSAgent]
See http://www.w3.org/2009/02/26-svg-irc#T20-27-19
20:27:51 [ChrisL]
ds: maybe contact andrew and ask about layeredg
20:28:14 [ChrisL]
ed: think you just need a different tree traversal order, not a whole copy of the subtree
20:28:35 [ChrisL]
ds; impact of that on keyboard navigation / docuent order?
20:29:58 [ChrisL]
ds: would this affect hierarchy eg with a use element that points outside the z-ordered group
20:30:32 [ChrisL]
... example of two circles bing twiddled on z-order, then two use elements that point to them
20:31:04 [ChrisL]
ag: i would say no, it gets really complex. import the transform not the layered group it could get really hairy
20:31:26 [ChrisL]
ds: right, using them outside the layering context gets document order not this new rendering order
20:32:23 [ChrisL]
ag: trying to ease implementor burden while keepin git simple and powerful for authors
20:32:37 [shepazu]
<layer><circle id="c1" .../><circle id="c2" .../></layer><use xlink:href="#c1" .../><use xlink:href="#c2" .../>
20:34:00 [ChrisL]
ag; a reference to that ISO spec (for full title, for the references) http://www.iso.org/iso/iso_catalogue/catalogue_tc/catalogue_detail.htm?csnumber=51502
20:34:31 [ChrisL]
ds: need to name this carefully, eg inkscape already has laters that are really more like groups
20:35:10 [ChrisL]
costs CHF 380,00
20:35:56 [ChrisL]
ed: so its good to gather use cases and requirements for z-index
20:36:07 [ChrisL]
ag: does this impact the transforms UC&R?
20:36:20 [ChrisL]
ed: should still examone the use case for it
20:37:08 [ChrisL]
cl: good in the transforms UC&R to discuss how this is similar to, and different from, xss 2.1 z-index
20:37:46 [ed___]
s/xss/css/
20:38:29 [anthony]
s/examone/examine/
20:40:31 [ChrisL]
action: anthony draft use cases and requirements for z-index and compare and contrast with transforms and critically discuss with reference to a neo-marxist dialictic
20:40:31 [trackbot]
Created ACTION-2482 - Draft use cases and requirements for z-index and compare and contrast with transforms and critically discuss with reference to a neo-marxist dialictic [on Anthony Grasso - due 2009-03-05].
20:41:18 [ChrisL]
topic: tpac 2009
20:41:23 [ed___]
http://lists.w3.org/Archives/Public/public-svg-wg/2009JanMar/0174.html
20:43:10 [ChrisL]
cl: (explains tpac question - group attendance, timing, cost and location)
20:43:54 [ChrisL]
ds: an advantage of the proposed location is silicon valley so more particiation from browser vendors like mozilla, applle etc
20:44:29 [ChrisL]
http://lists.w3.org/Archives/Member/chairs/2009JanMar/0059
20:44:48 [ChrisL]
ds: its one month after svg open in the same area
20:45:25 [ChrisL]
cl; prefer the cost-effective location
20:45:50 [ChrisL]
ds: perhaps shift svg open?
20:46:12 [ChrisL]
cl: not clear that the tpac will happen. depends on how many groups confirm
20:47:20 [ChrisL]
... plh suggested a mini-tpac for browser-based technologies like html, css, svg, webapps and so forth
20:48:53 [ChrisL]
ed: mini-tech plenarry sounds good
20:49:12 [ChrisL]
s/rry/ry/
20:50:22 [ChrisL]
ds: downside is that we miss being able to meet with other groups
20:52:34 [ChrisL]
http://www.w3.org/2002/09/wbs/34786/TPAC2009/results
20:58:42 [ChrisL]
ds: this has an impact on level of detail
20:58:51 [ChrisL]
ag: we need near and far clipping really
20:59:15 [ChrisL]
ag: this is 3d transforms applied toa 2d object then flattened to give a 2d result
21:00:10 [ChrisL]
topic; back to transforms
21:00:24 [shepazu]
s/this has an impact on level of detail/3d Transforms have an impact on level of detail/
21:00:26 [ChrisL]
oops, should have changed topic sorrry
21:00:45 [Zakim]
-anthony
21:01:43 [ChrisL]
topic; odf and svg
21:01:49 [ChrisL]
s/;/:/g
21:02:25 [ChrisL]
ds: it would be a win if odf were to use svg natively. currently they don't. its being discussed in svg ig
21:02:36 [ChrisL]
... who is interested?
21:02:41 [ChrisL]
ed; yes, sure
21:02:48 [ChrisL]
s/;/:/g
21:02:59 [ChrisL]
zakim, list attendees
21:02:59 [Zakim]
As of this point the attendees have been Shepazu, heycam, ed__, ChrisL, [IPcaller], anthony, erik
21:03:00 [anthony]
anthony has joined #svg
21:03:25 [Zakim]
-ChrisL
21:03:27 [Zakim]
-heycam
21:03:29 [Zakim]
-erik
21:03:30 [ChrisL]
rrsagent, make minutes
21:03:30 [RRSAgent]
I have made the request to generate http://www.w3.org/2009/02/26-svg-minutes.html ChrisL
21:03:37 [ChrisL]
chair: erik
21:03:40 [ChrisL]
rrsagent, make minutes
21:03:40 [RRSAgent]
I have made the request to generate http://www.w3.org/2009/02/26-svg-minutes.html ChrisL
21:04:04 [ChrisL]
rrsagent, make logs public
21:04:08 [ChrisL]
rrsagent, make minutes
21:04:08 [RRSAgent]
I have made the request to generate http://www.w3.org/2009/02/26-svg-minutes.html ChrisL
21:04:20 [Zakim]
-Shepazu
21:04:22 [Zakim]
GA_SVGWG()2:30PM has ended
21:04:24 [Zakim]
Attendees were Shepazu, heycam, ed__, ChrisL, [IPcaller], anthony, erik
21:07:50 [anthony]
anthony has joined #svg
21:16:09 [ed__]
ed__ has joined #svg
22:06:44 [heycam]
heycam has joined #svg
22:26:15 [shepazu]
heycam: who's this "https" guy?
22:27:28 [heycam]
yeah :)
22:27:56 [heycam]
i noticed that in one of our minutes from last week
22:53:57 [heycam]
anthony, thanks
22:54:09 [heycam]
i'm reading through ChrisL's pdf at the moment
23:01:26 [heycam]
anthony, what happens if we use a negative number for the 'perspective' property?
23:03:59 [anthony]
means that the viewing distance is on the other side of the plane... but I'll have to test what happens exactly - I think the perspective effect is essentially flipped
23:04:23 [anthony]
plane ---> X-Y plane where Z=0
23:04:55 [heycam]
yeah. i guess layeredG or whatever mechanism used to decide the painting order would need to take that into account.
23:05:47 [heycam]
so with our spec, the only projection transform you can do is the perspective one where the viewing point is (0,0,z)
23:06:15 [heycam]
but you can always use translate/rotate to choose a new viewing point, right?
23:06:47 [anthony]
correct in both cases. The syntax is actually perspective(x y z)
23:07:04 [anthony]
so you can move the viewing point on the X Y plan if you wish
23:07:15 [heycam]
oh ok
23:07:22 [heycam]
for the css one?
23:07:48 [anthony]
CSS has perspective(z) and perspective-origin(x,y)
23:07:53 [heycam]
oh
23:08:05 [heycam]
so we allow a perspective(x y z)?
23:08:26 [anthony]
yeah, we effectively combine the two
23:08:35 [heycam]
oh in our spec it sometimes talks about 'perspective' and sometimes 'perspective-point'
23:08:45 [anthony]
nah, I've made changes to fix that
23:08:50 [anthony]
just need to commit them
23:08:55 [heycam]
k
23:09:06 [anthony]
the old spec was terrible very confusing
23:09:16 [anthony]
ooops missed a ","
23:09:16 [Zakim]
Zakim has left #svg
23:09:26 [heycam]
and we can only do a single vanishing point, whereas the 4x4 matrix can do two and three
23:09:39 [heycam]
what are multiple vanishing points useful for anyway?
23:10:32 [anthony]
one example is a corner street view
23:10:34 [anthony]
http://gatewayhsart.files.wordpress.com/2008/12/perspective-street_corner.jpg
23:11:27 [heycam]
ah, interesting
23:11:34 [heycam]
so we couldn't do that
23:12:48 [anthony]
in short no. but you could construct that scene still using what we have
23:13:09 [anthony]
given that the objects we deal with are 2D in a 3D space I'm not sure of the use case for 2 or 3 vanishing points
23:13:35 [anthony]
like if we were dealing with 3D objects in 3D space then... yeah big limitation
23:13:37 [heycam]
ok, you could have two groups with different perspective transforms
23:13:44 [anthony]
correct
23:13:47 [anthony]
also
23:13:47 [heycam]
one group for each street
23:13:52 [anthony]
here is an example of 3 point
23:13:54 [anthony]
http://www.atpm.com/9.09/images/design-three_point.gif
23:14:35 [anthony]
having said that - I'm still open to suggestions though
23:14:37 [heycam]
oh, cool
23:14:44 [heycam]
so what does normal eyesight do?
23:14:54 [anthony]
1 point perspective
23:15:04 [heycam]
in the direction that you're facing, i suppose
23:15:08 [anthony]
yup
23:15:10 [heycam]
well, looking :)
23:15:13 [heycam]
ok
23:15:34 [heycam]
so it would be good to ask csswg if it's worth supporting these multiple vanishing points then
23:16:18 [anthony]
well actually in a way we do see two points if we are looking at a corner
23:16:52 [anthony]
but it's hard to focus on both
23:19:49 [anthony]
but yeah, would like to see the use case for their proposal
23:26:08 [anthony]
also, heycam, another big problem is bounding box calculation - CSS hasn't defined this
23:26:20 [anthony]
well, neither have we, but I have an idea of how to
23:53:14 [heycam]
2d bounding box?
23:53:16 [heycam]
after projection?
23:53:31 [heycam]
yeah i guess you might want to get it before or after projection
23:57:41 [anthony]
nah, the thing is a bounding box is just that
23:57:43 [anthony]
a box
23:58:06 [anthony]
so one way of doing it is multiply the bounding box by the Transform matrix ONLY
23:58:17 [anthony]
that way you still end up with a box
23:58:51 [heycam]
does that give you the same box as doing the projection, then taking the bounding box of that?
23:59:18 [anthony]
although one problem with multiplying it by the transform matrix is it could rotate...
23:59:37 [heycam]
i mean i guess it's the same as with 2d transforms
23:59:44 [heycam]
you get the bbox before the transform, yeah?
23:59:53 [anthony]
I guess so