SVG Working Group Teleconference -- 18 Aug 2011

<trackbot> Date: 18 August 2011

<cyril> thanks heycam

<cyril> heycam, how do you generate the ToC on the group's wiki pages ?

<heycam> cyril, you just use headings like "= Level 1 heading =" or "== blah blah Level 2 heading =="

<cyril> thank you

<cyril> I'll edit the SVG 1.2 Tiny page

<cyril> it's easier to reference to a item by number

<cyril> yes

<cyril> scribe: Cyril Concolato

<cyril> scribenick: cyril

meeting plan around SVG Open and TPAC

CM: we decided not to have an SVG F2F WG meeting in SVG Open but Tav can only be there

DS: having a day is reasonnable

CM: it doesn't even have to be a whole day
... it depends on the topic
... it would be nice to have Tav because he's usually not able to travel
... we can have a meeting and review what was discussed later on in Santa Clara

ED: what day ?

CM: are you involved in the workshop day ? on the thursday ?
... Tav is not involved in the workshops
... I'm wondering about rooms
... it would be good to meet in the same place as the workshops

CC: could we have a bridge to call in ?

CM: someone should contact the organizer to check room and bridge

ED: the 4th day should be fine

<scribe> ACTION: heycam to contact Andreas Neumann about the room and bridge for a short 1 day F2F at SVG Open 2011 [recorded in http://www.w3.org/2011/08/18-svg-minutes.html#action01]

<trackbot> Created ACTION-3109 - Contact Andreas Neumann about the room and bridge for a short 1 day F2F at SVG Open 2011 [on Cameron McCormack - due 2011-08-25].

Feature request list for SVG 2

CM: we have 3 lists on the wiki

<heycam> http://www.w3.org/Graphics/SVG/WG/wiki/SVG2_Planning_Page

CM: we should discuss how we decide on those lists
... especially before the F2F
... we should put comments on the wiki
... it would be good if people could put comments on each feature and put a yes/no

ED: I have an action to go through minutes for resolutions
... I'm not sure about overlaps

CC: we should have a single list

DS: agree, we should also quantify
... putting a numerical value 1-5
... one for priority
... one for complexity

<ed> ACTION-3095?

<trackbot> ACTION-3095 -- Erik Dahlström to go through the last few f2f minutes to find resolutions for SVG2 items, and add them to the wiki page. -- due 2011-08-05 -- OPEN

<trackbot> http://www.w3.org/Graphics/SVG/WG/track/actions/3095

DS: one for usefulness
... those are the 3 key factors
... obviously complexity is for viewers, but it might be also for authors but we can work that out in the text

CM: should one of us go through all and put initial numbers

DS: we can each contribute and then we discuss where we will not agree

CM: would like to put those assessment done before you put your comments ?

DS: not necessarily
... we should start today to put assessments

CM: I think my 2 lists should be free of duplicates
... Issues are vague, coming from the tracker

range from bug to feature

scribe: I did not include bugs or problems with existing features

<heycam> http://www.w3.org/Graphics/SVG/WG/wiki/SVG2_Requirements_Input_From_Tracker

scribe: it would be nice to have the issue in a single list

CM: is it reasonnable to go through the list for next week ?

<scribe> ACTION: heycam to merge the two features/issues lists [recorded in http://www.w3.org/2011/08/18-svg-minutes.html#action02]

<trackbot> Created ACTION-3110 - Merge the two features/issues lists [on Cameron McCormack - due 2011-08-25].

<scribe> ACTION: cyril to merge the SVG Tiny 1.2 features into the results of ACTION-3110 [recorded in http://www.w3.org/2011/08/18-svg-minutes.html#action03]

<trackbot> Created ACTION-3111 - Merge the SVG Tiny 1.2 features into the results of ACTION-3110 [on Cyril Concolato - due 2011-08-25].

CM: I'll put a template also to enable people to comment
... what about "Feedback from Jonathan Watt: https://wiki.mozilla.org/SVG:Language:Regrets"

ED: I wonder how many are already covered by issues ...

CM: at least one of those is already there (xlink)

<scribe> ACTION: heycam to go through jwatt's SVG regret page and add any feature from there to our requirements wiki page [recorded in http://www.w3.org/2011/08/18-svg-minutes.html#action04]

<trackbot> Created ACTION-3112 - Go through jwatt's SVG regret page and add any feature from there to our requirements wiki page [on Cameron McCormack - due 2011-08-25].

<heycam> http://www.w3.org/Graphics/SVG/WG/wiki/SVG2_Requirements_Mailing_List_Feedback

CM: it's probably hard to accept something without a concrete proposal

CC: we shouldn't definitely reject features without concrete proposal
... we should give an idea if a proposal in a given area is acceptable to the group or not

<heycam> ScribeNick: heycam

<scribe> Scribe: Cameron

radial gradients

CC: this is related to an action erik has

ACTION-3097?

<trackbot> ACTION-3097 -- Erik Dahlström to propose wording for the edge case where a radialGradient's focal point sits on the edge of the circle and the gradient repeats. the spec. should say that when the focal point is on the circle edge, with repeat, then the distance between the first and last stop for the repeating colors is 0 and the paint should generate a color that is the average of all the gradient stops. -- due 2011-08-05 -- OPEN

<trackbot> http://www.w3.org/Graphics/SVG/WG/track/actions/3097

CC: In seattle we discussed content that has radial gradient noninteroperable behaviour
... I tried to check with the spec what the problem was
... if the spec was actually specifying something about the behaviour of the gradient outside of the circle, but there's no formula for how the gradient should be built
... I only found one in the canvas spec

http://lists.w3.org/Archives/Public/public-svg-wg/2011JulSep/0095.html

scribe: so in the canvas spec it clearly defines how to build the gradient
... but it doesn't say what to do when the centre of the circle is outside the main circle

ED: canvas doesn't have reflect and repeat

CC: I think as a basis we could adopt this definition for the inside of the circle
... it would clarify how we build the gradient
... then we should define how we fill the outside the circle
... in the email I have two examples

<cyril> http://perso.telecom-paristech.fr/~concolat/SVG/radial_gradient_construction_animation.svg

CC: that shows a radialGradient with a repeat
... it shows that if you continue the formula defined in the canvas spec, you don't have a problem, the gradient is well defined

<cyril> http://perso.telecom-paristech.fr/~concolat/SVG/radial_gradient_construction_animation_edge_case.svg

CC: in this one the focal point is on the edge of the circle defining the gradient
... you can still follow the canvas spec formula, but it only fills half the plane
... as you animate the circle you'll never go beyond the normal to the focal point

CM: so that'd be a straight vertical line

<krit> Firefox and WebKit don't support setting focal point on the edge

CM: for reflect we could just mirror the image around that line
... we could just define this not to work

CC: or choose a solid colour to fill the rest of the plane with

ED: I think IE does this

CC: I'm just checking IE9, they don't use a solid colour

CM: so I think it would be good for us to have a formula for computing gradient colors

CC: the one in the canvas spec is the one that we want

ED: the action I have is to make the half plane fill with a solid colour, average of all the stops in the gradient

CC: I don't understand why though
... it's not even a colour that the author specified
... I could understand the colour at stop 0 or 1, but why the average?

ED: probably because it's hard to pick any colour that matches

CM: I think Tab was using average of stop colours in certain cases in css3 image values, but I don't remember which cases they were

ED: if you have a line perpendicular to the radius of the circle, and keep very tiny steps between each straight line, you'll get something like an average
... tightly packed colours from the gradient

CM: we could flip the well defined half of the image to the undefined part

ED: it seems like new behaviour though

CM: does it make sense as a "reflect" or "repeat"?

ED: not so much repeat, reflect yes. it's different from the linear gradient though.

CC: another point was talking about a new value, a smaller radius fr=""
... I think we can already do it by adding an initial stop colour
... same offset colour for 0 and fr/r
... then you get the same effect

CM: unless you want to repeat/reflect into that smaller circle

CC: the pdf spec has an fr

<cyril> Figure L.11

CC: they have some interesting examples

<cyril> http://wwwimages.adobe.com/www.adobe.com/content/dam/Adobe/en/devnet/pdf/pdfs/PDF32000_2008.pdf

CM: ok so they don't fill anything in the inner circle

CC: let's see what erik proposes
... the adobe guys might have an opinion

ED: the canvas one, does that support things outside the circle?

CC: no

ED: I will consider this when making the proposal

propagation of events

http://lists.w3.org/Archives/Public/public-svg-wg/2011JulSep/0085.html

CC: I don't know if you understood my problem
... I think this is applicable to 1.1
... when you have an SVG linked from another SVG, two separate documents
... this describes a single application
... you want some key events to be processed by the parent document, and some by the child document
... you want to design your documents so they don't know exactly what the other document is doing with the keys
... e.g. if the subdocument has the focus, and you press "a", if it's used by the subdocument it's fine, if not then it could go to the parent document
... a bit like mouse events -- when you click on the subdocument and there's nothing to be clicked on, transparent background, then the parent document will take that event

CM: I think they're slightly different cases
... I'm thinking about iframes, events don't propagate across those

CC: what do inner documents do with a key that's not been used?
... I'd like some text to describe that

[discussion of animation vs image]

ED: I would expect svg <image> to behave the same as html <img>

CC: the question then is how should it work with an iframe
... I'll make some tests

CM: if we wanted to allow propagation across windows, it's broader than just svg

CC: yes it's more of platform thing than just svg

CM: I think it is defined that it shouldn't propagate currently
... per dom events flow

text tracks and webvtt

CC: I was arguing that you could do the same thing in SVG, with synchronised graphics
... I wanted to check whether there was any interest in having synchronised graphics in this context
... I found a page from Charles on how to do subtitling in SVG in Opera

s/Chaals/Chaal/s

<cyril> http://dev.opera.com/forums/topic/370891

CC: This is a year and a half old example
... sylvia was saying that browsers will implement webvtt, and not svg as tracks

ED: webvtt is probably a lot simpler to write

CC: yes but it's more limited
... my guess, since it's how we implemented in gpac, is that once you have svg and video support, adding synchronised display of svg on top of video is not complex

CM: does webvtt fire events?

<krit> it does

CC: ontrackcuechange

CM: so maybe you could have svg animation triggered off those

CC: all you need is to use the same time base as the video
... don't even need the events
... see the example from chaals
... you can structure your svg file in a way that is frame based, it's really equivalent to a vtt file
... one use case is having clickable shapes that come up
... with a link to a web page for an advertisement for example
... the media pipeline taskforce are only considering rectangular regions

<cyril> http://lists.w3.org/Archives/Public/public-web-and-tv/2011Aug/0023.html

CM: one question would be whether to use the time of the video to drive smil animations in an svg document, or just to use webvtt cue times to show/hide svg fragments

<cyril> http://www.w3.org/TR/html5/video.html#handler-texttrackcue-onenter

CC: that's the actual events for the cues

DS: I agree with cyril that there should be non-rectangular regions

CC: how about the timing?

DS: no opinion

CC: silvia was mentioning another option would be to embed svg content in the webvtt file, and retrieve the svg from the cue event, put that svg inside the html document by script
... that would work, but it's a bit complicated
... if there is any interest in this, we can discuss this

trackbot, end telcon

- DRAFT -

SVG Working Group Teleconference

18 Aug 2011

Attendees

Contents

meeting plan around SVG Open and TPAC

Feature request list for SVG 2

radial gradients

propagation of events

text tracks and webvtt

Summary of Action Items

Scribe.perl diagnostic output