<fesch> chairL fred

<scribe> scribe: Brian

Nested Links

fesch: We had a discussions thread on ARIA
... many of the things worthy of note are that nested links are not allowed in HTML
... nested links don't have particualrly good support, because it's not allowed
... not sure if you read all the threads, but some of them got philosophical about what people would like
... In the earlier email threads, some of them like Matt, pointed out

<fesch> Fred, screen readers, at least, do not have a way of rendering a link that is nested inside another link. So, yes, there are problems. A good question is whether nesting links provide sufficient benefits to merit solving those problems … in ARIA 2.0, of course. Matt

fesch: that was matt's input - the main point is that screen readers don't know how to handle it

<fesch> I see a lot of problems with this. First, HTML doesn't have nested links,

<fesch> so why should SVG if they want to duplicate HTML features?

<fesch> Second, since there are no nested links in HTML, ATs have no support for

<fesch> them.

richardschwerdtfeger: they flateen them out, that's what the browsers doing

<fesch> Third, nested links in themselves sound like a bad idea. Where does one

<fesch> link end, where does the next one start, what's the keyboard model like,

<fesch> where do clicks go, do we have two tab stops or three (in terms of two

<fesch> links nested in one another) etc. etc.

<fesch> But basically, the discussion ends at point 2, if not even point 1.

<fesch> My suggestion is that SVG should not have nested links. No support for them

<fesch> in HTML or ATs, and lots of unanswered questions that we didn't have to

<fesch> deal with in HTML because nested links aren't supported.

<fesch> Marco Zehe

fesch: several things are happening, they would drop stuff after the second link. There's a lot of different behavior
... above is marco's email - ATs don't support them, why does SVG want them

richardschwerdtfeger: I agree, it's confusing

<fesch> Amelia states she will ask the SVG WG to neuter nested links.

AmeliaBR: There are a couple issues here that really come down to ATs have no idea how to handle this, becasue it's not something they ever expect to come across
... when you have a link and structured content inside the link they should be able to expose that. But a link inside a link, we can say that shouldn't function. Right now it's inconsistently supported
... we wuld need to submit bugs to various browsers

richardschwerdtfeger: there's too many inconsistencies with SVG across browsers. We don't need to introduce anything as flaky as this

AmeliaBR: I haven't reverted the changes yet. We have a resolution next week in the SVG working group to formally remove this. I don't expect anyone to vigorous defense of nested links
... there needs to be special error handlng rules for nested links so that child content will be presented in browsers that do not support nested links

fesch: does anyone here support nested links in SVG? Actual silence, in this group?

AmeliaBR: I like the idea, but not right now

richardschwerdtfeger: SVG 6

fesch: since no one supports it, and we know there's technical issues lets make a resolution
... any discussion, anyone opposed?

RESOLUTION: Recommend the SVG Workign Group Withdraw Support for Nested Links in SVG

fesch: Amelia, if you want ot relay that to the SVG group, we'd appreciate it
... now, let's talk about your changes

Amelia's Changes

<fesch> https://lists.w3.org/Archives/Public/public-svg-a11y/2016Jun/0013.html

AmeliaBR: we did a face to face, we made lots of different changes
... these are the ones I thought had accessibility implications

<AmeliaBR> https://svgwg.org/svg2-draft/struct.html#DescriptionAndTitleElements

AmeliaBR: I finally rewrote the section on description adn title, here's the current text
... a few changes we previously talked about. Clarifying the rules about multi-lingual switching
... a mention that browsers can user user's preferences to pick the best match, rather than simply the first match
... there are rules for when lang is an empty string, which is valid for unknown or no language
... an example is included with a star, this will match as the lowest priority match for everyone
... a warning about against using an empty title and desc

fesch: before we go on, it looks like this may need a new testable statement

AmeliaBR: probably yeah, in your section dealing with language switching. You'd probably want to use that example I've got in the spec
... that's something Joni brought up, what to do with empty language

<fesch> action fesch add testable statement for blank lang

<trackbot> Created ACTION-2027 - Add testable statement for blank lang [on Fred Esch - due 2016-07-06].

AmeliaBR: warnings about title and desc, and about using title and desc if something is semantically important
... the other change is making it clear that title and desc can be used on any element except for a few exceptions. The switch element; script and style are treated as character data. Other than that you can have title and desc on anything
... for anything not rendered you can only use it as labelledby or describedby

fesch: a title on a filter?

AmeliaBR: yes, you can put a title on a filter but it won't add it to the accessibility tree
... this is more for authors, the AAM is more for user agents

fesch: this does not change anything in the SVG-AAm?

AmeliaBR: it shouldn't
... the only normative changes in the AAM are the language switching rules, because it defers to SVG in picking the correct language option

fesch: any questions?

AmeliaBR: the next one on my lsit was nested links, but we already talked about it
... number 3, we may need some changes to AAM, but that should be minor
... the mesh element was a paint server and a shape. I will need to add an entry for the AAM for the new element, but I'm holding off as we're still deciding about the best name for the new element.

fesch: once the name's decdied we'll add a testable statement for the new element

AmeliaBR: the swtich element now has better language selection rules. This now allows using the preferred language if multiple acceptable languages are provided.
... I don't know if you have any tests involving switch. Not stirctly an accessibility feature.
... final change is not in that email, but I put through yesterday a section on aria- attributes and allowed roles

<AmeliaBR> https://svgwg.org/svg2-draft/struct.html#WAIARIAAttributes

AmeliaBR: I was gonna rewrite these to use categories. For allowed rules and default roles, rather than using explicit values for each element. This should be up to date according to the SVG-AAM
... this is now using the new graphics roles.

fesch: do we have a catch-22 situation? Do you reference the roles in the graphics module AAM?

AmeliaBR: I don't reference the graphics AAM, I refrence the graphics ARIA module
... practially, before we can test this, we need the graphics-AAM
... saying it should use this role, but not how it should be exposed means we don't have something we can test

fesch: that's all the changes so far?

AmeliaBR: There were some editorial changes. I'd appreciate someone reading through it quickly for anything problematic.
... the normative change was changing the table to match the chagnes we've already got in AAM

fesch: any questions for Amelia?
... hearing none, I think we're done with that subject. Does anyone have any news or anything like that?

Publishing Drafts

AmeliaBR: I think we resolved on this last time around. Graphics-aam needs a lot of work, but I think we should publish it as a first working draft

fesch: a workign draft can have changes, we expect it
... I'm going to propose a resolution we publish a first draft of the graphics-aam and a heartbeat of the other two documents

<richardschwerdtfeger> +1

RESOLUTION: Publish a First Working Draft of the Graphics Module-AAM and the Heartbeat of the SVG-AAM and the Graphics Module

fesch: anyone have any news or other things they want to talk about? Agenda items?

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)

Found Scribe: Brian
Inferring ScribeNick: Brian
Present: AmeliaBR

WARNING: Fewer than 3 people found for Present list!

Found Date: 29 Jun 2016
Guessing minutes URL: http://www.w3.org/2016/06/29-svg-a11y-minutes.html
People with action items: