See also: IRC log
<scribe> scribe: AmeliaBR
<fesch> agenda https://lists.w3.org/Archives/Public/public-svg-a11y/2016Sep/0000.html
LJW: A question came up recently. Do any SVG editors support creating accessible SVG?
ABR: Inkscape allows you to add title and desc in the visual editor. Anything else (ARIA), you'd need to go to the XML editor.
LJW: I think they're using D3.
FE: D3 is a basic JavaScript
library. It doesn't do any accessibility for you, but you can
add the elements and attributes yourself.
... I'd recommend the Brunnel Visualization Language for
accessible chart output.
<chaals> [I'd note that a lot of what you want to do is use links and tabindex, or actual graphics features that any editor can do…]
<shepazu> (there's also ChartML)
FE: Any update from SVG WG?
ABR: Nikos has finished tidying up the SVG 2 spec, and put in the formal request for CR status to PLH. Hopefully that will happen before TPAC.
FE: And updates from ARIA?
RS: We have a few little details
to work on for ARIA 1.1. Main issues are with getting the
mappings correct and up-to-date, for the new roles and for UIA
API.
... And the other thing going on is building the test harness.
Main issue is getting someone to work on it for Mac. But
whatever happens there will be easy to extend to SVG.
<chaals> [Doug and I have done a fair bit of work on that, although we're behind on getting the documentation into decent shape]
RS: The other next step for this group is authoring practices.
FE: I've been talking with Judy about that, and letting them know that we really need more resources, especially with Doug's time, to focus on SVG accessibility and authoring practices.
DS: There is a discussion about
that. But there's lots of chaos with the W3C re-organization,
so I haven't heard back.
... The other issue is that the SVG WG charter runs out in
October. Although SVG 2 should be at CR, it won't be done [ABR:
particularly re test suite].
... I had hoped we'd be able to just extend the existing
charter, but that is not encouraged anymore. So I'm drafting a
new charter, and will make sure it emphasizes the accessibility
issues.
RS: I have some concerns about the cross-dependencies between SVG 2 and ARIA 1.1, to make sure anything that SVG references isn't at-risk in ARIA. Particularly re new attributes, we're trying to get commitments from implementers.
DS: As far as SVG 2, I don't think it specifically calls out any of the ARIA attributes.
ABR: There's a list of attributes, but the next paragraph is a warning that the list may be updated by a future ARIA spec.
DS: Relative to other issues
going on with other W3C specs, and other parts of SVG 2 that
are at risk, it's not very big.
... We expect that SVG 2 is feature-complete, but we reasonably
expect that other details will get modified during the CR
period. There are some at-risk issues with no implementations,
and there are probably details that will need to be clarified
as implementations and test suites are built. So we will likely
have to publish an additional CR draft at some
point.
RS: For ARIA and SVG, we've had
good implementer support, from Google, and Apple. Haven't heard
a lot from Mozilla, unfortunately, but they have at least been
working on tabindex support.
... Thanks to everyone who has been working on these specs. We
really have made headway.
DS: Update on spec publication. There had been some questions, but I think we've got them sorted out.
<LJWatson> MS has SVG AAM as a "medium" priority on its Edge roadmap https://developer.microsoft.com/en-us/microsoft-edge/platform/status/svgaccessibilityapimappings/?q=svg
ABR: That's good to know. Good to remember that those are still updated working drafts. We still need a lot of work on Graphics-AAM mapping. Also for SVG-AAM, I've got some changes in the editor's draft that won't be in that working draft, but I also have open issues that we need to address now that SVG 2 and ARIA 1.1 are stable.
FE: I've updated the script I
have that parses the tests in the wiki and converts them to the
harness.
... It integrates the ARIA mapping data with the API mapping
data to convert from the SVG tests to actual API outputs for
each API.
... But the limitation is that there are still lots of TBD in
the API mappings.
RS: We need to get Microsoft to
spend some time on it, and they have not been able to do that
so far.
... My focus is on getting the core ARIA 1.1 mappings first,
then SVG and HTML.
FE: Tomorrow, we have one last conference call about the testing harness. Any final questions or concerns to bring up there?
ABR: So what are you checking? Whether an element is exposed or not, and the computed name and description, what else?
FE: Well, for relationships like labelledby, we currently just define "empty" or "not-empty".
ABR: So the harness doesn't confirm whether the pointer is to the correct labelling elements?
FE: There are cases where we have aria-labelledby="", and I'm not sure the mapping is correct.
RS: That sounds like a simple error case testing, there shouldn't be any label in that case.
(more narrow discussion of specific tests from the above links)
<Rich> https://www.w3.org/TR/accname-aam-1.1/
RS: In Section 5 of the Acc Name spec, it specifies which fields in UIA should be used. The name property and the description property. They were adding a new feature to UIA for description, though, I'm not sure what the state is.
LJW: As I posted the link earlier, Microsoft has put SVG-AAM on their roadmap, with "medium" priority. From what I've heard, the main obstacle to further work is that it's still a working draft. They want something more stable before they work on it.
<richardschwerdtfeger> http://rawgit.com/w3c/aria/master/accname-aam/accname-aam.html
ABR: That's a good argument. We've re-written it considerably. Even the WD that will be published next week is now well out of date. But we also need implementer feedback to address the remaining issues, so we need some involvement.
RS: Here (above link) is the latest draft of Acc Name. They've filled in those TBD's, Fred.
FE: My other concern is whether any of the changes I have just made will change Joanie's ATK layout tests. E.g., with respect to empty aria-labelledby.
ABR: Check with Joanie. She might have already got that change integrated.
FE: Should I tell the ARIA testing call about these choices? Because I'm not sure they have any tests that specifically look at whether relationships are empty or don't exist.
ABR: Makes sense. You need positive & negative tests.
FE: What about the language switches? Is it worth following up with that?
ABR: I think so. Having a test
suite would be very helpful for encouraging
implementations.
... The issue is that those test results are dependent on
additional user agent state, as far as what the preferred
languages are.
... The browsers must have tests that use that state, for the
switch/systemLanguage effect on rendering, but I don't know how
to indicate that within your testing harness.
RS: Who will be at TPAC?
DS: I will. [Silence from everyone else]
<chaals> [Chaals will be at TPAC, but flat out]
RS: So, it's probably not worth scheduling any time for SVG-A11y.
ABR: It would be useful if you could spend a few minutes of the main ARIA call updating folks on our progress, and asking for feedback on our remaining spec issues. But probably not worth discussing the issues too much, if no one has had a chance to review them before hand.
RS: OK, I'll schedule in a short SVG-A11y update.
trackbot, end telcon
This is scribe.perl Revision: 1.144 of Date: 2015/11/17 08:39:34 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/Library/Language/ Succeeded: s/In Section/RS: In Section/ Found Scribe: AmeliaBR Inferring ScribeNick: AmeliaBR Default Present: LJWatson, AmeliaBR, fesch, richardschwerdtfeger, shepazu, chaals Present: LJWatson AmeliaBR fesch richardschwerdtfeger shepazu chaals WARNING: No meeting title found! You should specify the meeting title like this: <dbooth> Meeting: Weekly Baking Club Meeting Got date from IRC log name: 06 Sep 2016 Guessing minutes URL: http://www.w3.org/2016/09/06-svg-a11y-minutes.html People with action items:[End of scribe.perl diagnostic output]