SVG Working Group Teleconference -- 02 Dec 2010

<trackbot> Date: 02 December 2010

<shepazu> http://www.nasa.gov/topics/universe/features/astrobiology_toxic_chemical.html

<scribe> scribeNick: pdengler

SVG 1.1 2nd Edition

heycam: requires last call comments by 12/15
... 12/15 is the deadline for asking for publication

<heycam> http://www.w3.org/Graphics/SVG/WG/wiki/Full_11#Remaining_work_for_SVG1.1F2

heycam: Need to resolve above linked issue and the test suite
... We could prioritize some of these issues (e.g. improve pass critiera)
... I've been updating tests and could give a concerted effort to finalize that in the next week; that should make us OK from the test suite perspective

pdengler: After the test suite, we need to run test suites and fiinalize open issues?

adrianba: Are you intending to move to CR or Proposed Rec?
... You need implementation reports for Proposed Rect

ed: Our plan was to do proposed REC and have test results

heycam: How long would it take for folks to run tests?

ed: It's not much to run through; it's only if we discover we don't have enough passing we're in trouble

heycam: We could drop tests that no one passes.

pdengler: I'd like to consider dropping specific tests on a case by case basis
... But in the interest in time, and the quality of our tests, I support the general thinking

heycam: It's important to get this done and not have it our charter

shepazu: It's bogging us down

heycam: We should consider simplifying the tests that we need to review

ed: There are some tests that require particular fonts to be installed, but batik on osx doesn't read those fonts for some reason. It only works with SVG fonts.

tbah: Does work on Linux on Batik

heycam: Have you been updating SVG Fonts?

ed: No, i've just been removing explicit use of Arial systemfonts, then it falls back to "SVGFreeSans, sans-serif" But there are instances where the test requires downloading of Fonts

ChrisL: One of the problem is that you have the right to download them, but not convert them

heycam: In the webpage we could provide a patch file PNG replacement

ChrisL: The problem with that is that you have to regenerate the image each time

pdengler: So let's run them on Linux

ed: Pretty sure I generated reference images on both

heycam: Do you have a list of these tests?
... If you can come up with a list of them...

ed: Probably get to them today

heycam: Then I will run that list on Linux/Batik and see the result
... Target early next week to be done and will email to let folks know it is ready to run

pdengler: How many implementations do we have available to run?

ChrisL: As long as independent verifications can be run by, for example, W3C, that should work

pdengler: If the tests are ready by the 8th, we can run them quite quickly after that (within a day or two)

heycam: Eric and I will target these by the 8th then

ChrisL: I can help with this as well as I have cleared up all of my actions

heycam: It turns of putting together the implementation report, do we have document filled in?

anthony: THe implementation report is ready, unless we included new tests

heycam: In that case on Wednesday if it is finalized I'll ask you to regnereate the template

ChrisL: We should run any implementations we can find

Remaining Issues

<ChrisL> I got around to all but one of my actions, now closed

pdengler: Can we put an end date on these

heycam: Let's get all actions done by end of next week (by next telecon)

<ChrisL> issue-2346?

<trackbot> ISSUE-2346 -- Last Call Comment: previous discussion about filterRes -- raised

<trackbot> http://www.w3.org/Graphics/SVG/WG/track/issues/2346

<heycam> http://www.w3.org/mid/op.vmxjwavjgeuyw5@localhost

ed: I had some progress on 2346 (ablove). I will make the edit to the spec

ChrisL: Sounds reasonable

<anthony> http://lists.w3.org/Archives/Public/www-svg/2010Dec/0001.html

feConvolveMatrix bias property

anthony: I sent some details on results and proposed formulas for feConvolveMatrix

ed: I'd like to see transparent image as well. It's not so interesting when its only opaque images

anthony: So you would like a PNG with a gradient, masks or transparency?
... What I did find is that all of the implementaitons vary. I liked the output from Firefox
... But I made the note that we should settle on something

ed: The proposal is OK with me as long as there are tests for the transparent images as well

anthony: I will create a test similar to the one sent to the list which includes the transparent image

<scribe> ACTION: anthony to update the feConvolveMatrix test to include transparent image and upate the spec to clarify the algorithm [recorded in http://www.w3.org/2010/12/02-svg-minutes.html#action01]

<trackbot> Created ACTION-2916 - Update the feConvolveMatrix test to include transparent image and upate the spec to clarify the algorithm [on Anthony Grasso - due 2010-12-09].

heycam: Tav, do you have everything you need to get your actions to get done

anthony: I will be working on my action items and have them done by next week

shepazu: Chris, do you mind working with Tav to get CVS access set up ?

heycam: If for some reason that doesn't come through in time, we can send patches through email

<ChrisL> Tav, your cvs access has been asked for. Pat, Sandy's as well

<adrianba> ACTION-2843?

<trackbot> ACTION-2843 -- Cyril Concolato to propose new wording for the spec for ISSUE-2368 -- due 2010-09-13 -- OPEN

<trackbot> http://www.w3.org/Graphics/SVG/WG/track/actions/2843

<heycam> ISSUE-2384?

<trackbot> ISSUE-2384 -- Order of rx / ry computation for rounded rects -- raised

<trackbot> http://www.w3.org/Graphics/SVG/WG/track/issues/2384

ChrisL: We will get the WOFF fall back fixed by test suite complete

heycam: There was a test that relied on text metrics that I thought would benefit from an SVG FOnt with a WOFF font

<scribe> ACTION: pdengler to Review Issue 2384 and decide if the change to the test and the text is reasonable [recorded in http://www.w3.org/2010/12/02-svg-minutes.html#action02]

<trackbot> Created ACTION-2917 - Review Issue 2384 and decide if the change to the test and the text is reasonable [on Patrick Dengler - due 2010-12-09].

<scribe> topic : Test suites with red fringing through on tests with same sized shapes

heycam: I would like to add a stroke so it doesn't appear that they fail (with the red pushing through). Is that OK with Patrick (and others)

pdengler: Yes

<adrianba> ACTION-2834?

<trackbot> ACTION-2834 -- Doug Schepers to clarify section 16.5 Processing order for user interface events relating to problems in ISSUE-2364 -- due 2010-09-10 -- OPEN

<trackbot> http://www.w3.org/Graphics/SVG/WG/track/actions/2834

<scribe> ACTION: heycam to respond to Cyril on action 2843 [recorded in http://www.w3.org/2010/12/02-svg-minutes.html#action03]

<trackbot> Created ACTION-2918 - Respond to Cyril on action 2843 [on Cameron McCormack - due 2010-12-09].

getComputedText Length

pdengler: I am happy with the response that textLength should be taken into account

<ed> http://dev.w3.org/SVG/profiles/1.1F2/test/svg/paths-dom-01-f.svg

ed: There is a test also on pathLength

<ed> http://dev.w3.org/SVG/profiles/1.1F2/test/svg/text-dom-01-f.svg

<text id="test" font-size="30pt" font-family="Arial" x="20px" y="200px" letter-spacing="5px" textLength="100">Filler Text</text>

document.getElementById("test").getComputedTextLength();

ed: My expectation is that getComputedTextLength which would what it thinks that the text length is, but not take into account what the textLength value at all
... But when it renders, it renders into 100

<heycam> ok right, that makes sense

<heycam> so then you do get differences in results due to fonts etc.

adrianba: What's the use case if it is not to return the size of the actual element

ed: It can differ from what the author expect or what it should be

<heycam> you could have a test where you give an outrageous value for textLength, and then test that getComputedTextLength returns a different value from .textLength

<heycam> or you could add svg/woff fonts

adrianba: My use case is that I can use getComputedTextLength to get the computed text length
... Then I can walk up to any other element and so something else to it konwing its size
... Regardless of the format or attribute on the element

ChrisL: If I was doing that, I would do the other way around, I would read the value of the attribute instead

adrianba: Either way seems right, I just don't know the use case where I asked for getComputedTextLength and get something other than what is drawn

ed: Using textLength is not that useful since it stretches the text which is behaved differently in all implementation

pdengler: Ok, we will keep it as is and address it in the next version of the specification.

shepazu: I raised an issue in SVG 2.0 that includes a font metrics API that is more comprehensive than textLength
... Should probably discussed in FX task foce

Filter spec clarification

http://www.w3.org/Graphics/SVG/WG/track/actions/2815

<adrianba> ACTION-2911?

<trackbot> ACTION-2911 -- Tavmjong Bah to finish the response to Dr Olaf regarding ISSUE-2338 and fix the issues highlighted by Dr Olaf -- due 2010-11-25 -- OPEN

<trackbot> http://www.w3.org/Graphics/SVG/WG/track/actions/2911

tbah: I think the clarification is simpler and i have an action for that
... I'll propose a rewritten section

MIME Type Wording Update

ChrisL: This is all dealt with now and is a non-issue as the WG will recongize the svg/svgz

- DRAFT -

SVG Working Group Teleconference

02 Dec 2010

Attendees

Contents