Protocols and Formats Working Group Teleconference

16 Oct 2015

See also: IRC log


fesch, AmeliaBR, Rich_Schwerdtfeger, shepazu, chaals
Rich, fesch


<trackbot> Date: 16 October 2015

<richardschwerdtfeger> Meeting: SVG Accessibility Task Force

<richardschwerdtfeger> scribe: Rich


<fesch> chair: fesch

<fesch> scribe: fesch



reviewing AAM editor's notes



rs: we are reviewing editors notes

<richardschwerdtfeger> https://rawgit.com/w3c/aria/master/svg-aam/svg-aam.html#exclude_elements

chaals: anyone know what happens when HTML is put into a desc

abr: so few browsers use the desc

rs: do the browsers parse them into a DOM - I don't know

<chaals> -> svg-access-w3cg.github.io/svg-a11y-tests/index.html A bunch of tests for what SVG accessibility actually works

chaals: you could put stuff in because it was XML , I was interested in this in 1999...

<AmeliaBR> http://www.w3.org/TR/SVG11/struct.html#DescriptionAndTitleElements

chaals: support the idea there are things that you can't see and things in the DOM you can't get to...
... making a test case for anchors...

<chaals> [For desc/title lets's make tests cases and find out what happens… feel free to put them in svg-access-w3cg.github.io/svg-a11y-tests/index.html ]

fe: there is a difference between hidden and not visible

abr: yes, want to make sure everyone can interact with it (not visible but interactive)

chaals: want Rich to take it up with core

discussing Note

The exception above is required because current best practice for fallback browser support is to use aria-labelledby and aria-describedby to redundantly link to title and desc child elements. These elements should only be used to generate the accessible name/description for the parent element. They should not be included as separate nodes in the tree.

<chaals> Example use case - can we include the content of svg-access-w3cg.github.io/use-case-examples/composed-tree-desc.html as an overall desc for svg-access-w3cg.github.io/use-case-examples/composed-tree.svg since the tree itself is pretty hard to understnd via screenreaders - and it is as simple a tree as you get.

<chaals> s|svg-access-w3c.github|http://svg-access-w3cg.github|G

fe: are you giving descriptive content to AT users and not non AT users

abr: yes

chaals: that is something like what we do in the Yandex browser

<richardschwerdtfeger> <desc xmlns:mydoc="http://example.org/mydoc">

<richardschwerdtfeger> <mydoc:title>This is an example SVG file</mydoc:title>

<richardschwerdtfeger> <mydoc:para>The global description uses markup from the

<richardschwerdtfeger> <mydoc:emph>mydoc</mydoc:emph> namespace.</mydoc:para>

<richardschwerdtfeger> </desc>

chaals: the key is switching to a foreign object

<chaals> I was planning on writing some tests for that...

chaals: bugs all across the board for browser support of SVG...
... works with Chrome, yandex, bad in Firefox, Safari (without voiceOver turned on)

ds: what we are doing here is writing a standard, so we shouldn't pay too much attention to what they are doing today, since we didn't tell them what to do

rs: we need to know what browsers are doing with desc

discussing Editor's Note

[Note 1] Issue 698: Need to address mappings when tabindex or aria relationships are applied to elements whose native host language semantics have a strict role="none". Typically these should not be mapped (elements are never rendered). However, tabindex is technically allowed on all elements even though authors are unlikely to do it.

abr: what if someone put tabindex on gradient or something that is never mapped?

rs: only elements that can be rendered (or whatever term) can be mapped

abr: so anything that can't be mapped - tabindex should be ignored

chaals: should go back to SVG group

rs: we had this discussion in SVG, want it like HTML

abr: HTML is weird but you can see...

chaals: lets go back to SVG and tell them they got it wrong
... needs to be fixed in the spec

abr: we made it clear in the SVG spec that if you put a tabindex on it, the tabindex is ignored

rs: so note can go to text.

discussing: Editor's Note

[Note 2] Issue 699: Address role="none" mapping to concrete objects in some platforms. A role of "none" should not be mapped unless there is an error condition to avoid unnecessary growth of the accessibility tree.


<richardschwerdtfeger> https://rawgit.com/w3c/aria/master/svg-aam/svg-aam.html#mapping_role_table

discussing Editor's Note

The "group" role is used in many cases as a generic container. There may be more appropriate mappings available in some platforms, which should be investigated. The proposed ARIA roles for graphics will replace some of these defaults.

abr: we have changed this a little bit...
... can get rid of note because we are more specific in the table

<richardschwerdtfeger> http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#role-map-text

chaals: seems img did what we thought... nothing I found would let me explore inside an img

abr: difference between symbol role and img role is how it would be announced to the user

jw: j in the hierarchy symbol could be derived from the img role

<chaals> [I don't see that the use cases justify the difference suggested as a reason to have a symbol role different from the img role]

abr: in the mapping table - we have two basic types of notes - one about symbol and the other issue is text

<richardschwerdtfeger> http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#role-map-text

abr: last week suggested we borrow the HTML mappings for text

we will replace groups with text on text, tspan and textPath

abr: may loose extra content
... use case in HTML when you want something to be treated as text, but in SVG we have

ds: tspan will be used less in the future because SVG2 has text wrapping
... anyone that is adding extra description can add it...

<AmeliaBR> http://www.w3.org/TR/wai-aria-1.1/#text

abr: HTML doesn't have roles for stuff like paragraph
... no ARIA roles for paragraph and stuff like that

jw: could you map directly to system API's?

abr: we could tell them to do whatever they do for paragraph..

rs: pinging Joseph

abr, chaals: tspan is like HTML spans

chaals: use should not have the role of img

<chaals> [It is perfectly possible to have <tspan role="heading" aria-level-"2">head</tspan><tspan>the text that gets covered</tspan> ]

chaals: use should clone
... you can't explore the root, but should be like a shadow DOM

abr: lots of issues

fe: is that an SVG level issue?

abr: it is both... tricky want to harmonize with shadow DOM, but that keeps changing - issue is not resolved

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.140 (CVS log)
$Date: 2015/10/16 15:13:10 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.140  of Date: 2014-11-06 18:16:30  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s|-access-w3cg.github.io/svg-a11y-tests/index.html|svg-access-w3cg.github.io/svg-a11y-tests/index.html|
FAILED: s|svg-access-w3c.github|http://svg-access-w3cg.github|G
Succeeded: s/suggested/suggested as a reason to have a symbol role different from the img role/
Found Scribe: Rich
Found Scribe: fesch
Inferring ScribeNick: fesch
Scribes: Rich, fesch
Present: fesch AmeliaBR Rich_Schwerdtfeger shepazu chaals
Found Date: 16 Oct 2015
Guessing minutes URL: http://www.w3.org/2015/10/16-svg-a11y-minutes.html
People with action items: 

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.

[End of scribe.perl diagnostic output]