CSS A11Y TF Inaugural Call -- 31 Jan 2017

<Rossen_> can somebody please scribe?

<scribe> scribe: MichaelC

introductions

ra: Welcome to CSS A11Y TF

thanks for making it on short notice

want to get things moving

Round of introductions

Rossen Atanassov, Microsoft Edge, work on CSS and lots of other things, co-chair CSS WG, involved in SVG

Bogdan Brinza, Rossen´s colleague

Greg Whitworth, also Rossen´s colleague

Ted Drake, Intuit, co-facilitator of this TF, working in APA WG

John Northup, been involved in A11Y for a while

Katie Haritos-Shea, Deque, involved in W3C A11Y since 2000

Michael Cooper, W3C staff contact to ARIA & APA

Rachel Andrew, CSS WG, particularly interested in layout

Michiel_Bijl, TPG, co-editor of ARIA Authoring Practices, also part of APA

fantasai, CSS WG, edit many of the specs

Janina Sajka, chair APA, a lurker to CSS A11Y TF

<Ryladog> Reminds that I will be leaving at the top of the hour

Francois Remy, Microsoft, CSS WG

Alan Stearns, co-chair of CSS WG, also lurking

Scope of our work

ra: initial meeting at TPAC identified areas overlap between CSS and APA that falls into this TF

fantasai has raised some issues to CSS WG

and we´ve started marking some issues with a11y tag

there was discussion of relationship between visual presentation and content structure

this is not a new topic, CSS WG has been dealing with since the advent of absolute positioning and box positioning

fantasai has put together an explanation and guidelines for that

also she and I have an action to write a best practices document for CSS authoring

that could be in remit of this TF

other topics people want group to be considering?

-> https://www.w3.org/WAI/APA/wiki/Category:CSS APA CSS tracking

<Rossen_> MichaelC: has been tracking APA issues related to CSS

-> https://www.w3.org/WAI/APA/wiki/Category:CSS_A11Y_TF APA request list for TF

<MichielBijl> https://www.w3.org/WAI/APA/wiki/Category:CSS_APA_CSS_tracking

<janina> https://github.com/w3c/aria/wiki/CSS-AAM-Potential-Features

js: generated content

ra: two broad categories

Visual display and rearranging compared to document structure

positioning, transforms, animations, etc.

Structural changes introduced by CSS

generated content, which isn´t in DOM, fixups

think the APA navigation category falls into the first category above

including impact on keyboard etc. navigation

think the layout topic is a hybrid

most of the issues are around visual placement of layout boxes

don´t see as much there about generated content

mc: makes sense to me, anything missing

td: what about audio?

ra: speech?

td: there used to be aural style sheets

ra: there is CSS speech module

this is fantasai´s territory

fantasai: speech module, would like to talk about what to do with it

it´s pretty stable, just 2 open issues right now

planning to address and then republish CR

not much implementation interest

minimal feedback for years

<Rossen_> speech module - https://github.com/w3c/csswg-drafts/issues/511

don´t expect much interest will emerge

though the speak and speak-as properties might interest implementers more

wanted to ask this TF whether the property should be separated from the rest of the module

either by module versions, or move speak to display module where it would be more visible to authors and implementers

several issues reported to CSSWG asking for the feature, even though it exists

td: was trying to work on it a while ago

ra: yeah, lots of interest

How and where will we work

ra: propose we do the work in GitHub

<MichielBijl> +1

CSS work does that

was a challenge to transition to that, but has proven successful and adaptive

have one repo available

we should decide if we want to use that, or choose a different one

is everyone ok with that?

mb, td: +1

<rachelandrew> +1

ra: big and small issues can be addressed there

can cross-tag and stuff

there is mailing list

though probably mainly for administrative stuff

don´t want to impose standing teleconferences

but if need arises, we could schedule one-offs

also probably spend some chunks of time at TPAC in FtF discussion

though probably not separate dedicated FtF meeting

any other thoughts?

td: sounds good

ra: for github, currently have https://github.com/w3c/css-aam

is that ok, or do we want a new css-a11y?

js: suggest keep AAM specific for AAPI mappings

<Rossen_> how about a new repo called CAT

mc: could create new repo for the primary issue discussion, and initial drafts of deliverables

keep CSS-AAM for that specific deliverable

could spin off other repos as needed

mb: better to split now than later if we´re gonna manage issues

RESOLUTION: create css-a11y repo as primary repo for this TF

<Rossen_> https://github.com/w3c/csswg-drafts/issues/511

speak property

ra: proposal is to split the speech module into two, one which could go to Rec and get implemented soon, but only have the speak property

<Zakim> MichaelC, you wanted to say a11y interested in speech but had kinda given up and to talk about issue vs development

mc: think a11y would be happy to see speech revived

think better to split the module than merge into display, but defer to CSS WG judgement

elika: want to sort that, and also ask whether speak-as should go along with it?

ra: given multiple years of lack of implementer interest, cautious about bogging down display module with it

so I vote for split as well

mc: what does speak-as do?

<fantasai> https://www.w3.org/TR/css-speech/#speaking-props-speak-as

fantasai: specify whether word should be spelled or pronounced

<fantasai> speak-as: normal | spell-out || digits || [ literal-punctuation | no-punctuation ]

mb: relationship to display: none?

<fantasai> speak: auto | never | always

fantasai: speak property allows auto, never, always

auto keys off of display / visibility

bb: could imagine this applies to name and description

and @@

why would you want to tie to display forcefully outside of those use cases?

fantasai: auto defines ¨normal¨ behaviour, if that defined behavoiur doesn´t seem intuitive we should refine the spec

use case of decorative element using text, might not want it spelled out

or asterisk placeholders

or supplemental navigation elements

<MichielBijl> So, el { speak: always; display: hidden; } is equal to el { position: absolute; right: 100%; clip: …; blah }? That’s really cool!

mc: kinda think speak-as more directly helpful to AT users

but should check with AT users more

makes sense to me to bring both those properties forward if feasible

but not sure one should die on the sword of the other

js: would like to canvas other AT users and developers

see good value in speak-as

bb: it used to be AT would get content out of the browser

nowadays get it more from a11y layer

means the property is less visible to them

so can´t build experience around it

and don´t think should encourage in browser

ra: subtree doesn´t always have rich markup

assumptions are built in around use of display:none

fantasai: there is a use case, that we should solve

there is a property

not sure how to handle, but multiple requests

<Zakim> MichielBijl, you wanted to say speak-as: literal-punctuation; would be great for specs / pre | code elements.

mb: see good value for literal pronunciation

use case, contractions make content more readable but AT over-expose the punctuation

so like speak-as

mc: may want to explore pushing the a11y layer before determining something shouldn´t be implemented

ra: not not implementable, but expect more UA enforcing of content integrity

do have concern about forcing element that´s not normally accessible to be made available introduces complexitiies

@@ linking to name description computation

think existing ARIA features cover the use cases

<fantasai> You should read https://github.com/w3c/csswg-drafts/issues/560 which is asking for this feature.

js: more than just AT are consumers

e.g., Alexa

say w-i-n-d pronounces differently depending on whether it´s action on rope or movement of air

fantasai: that´s a semantic issue out of scope for CSS

should handle at markup level

js: so speak-as is more restricted?

fantasai: yeah, doesn´t interpret, just says how to render

it´s possible the speak-as aspect [s/sh]ould be handled in markup

mb: +1 one markup for JS use case

ra: so good discussion

need to close the telecon though

Next steps

<Rossen_> https://github.com/w3c/csswg-drafts/issues

will set up css-a11y github repo

suggest participants follow CSSWG issues

in particular look for ones tagged for a11y

you can comment, cross-link with issues in css-a11y repo, etc.

thanks everyone for participating in this call and the TF

js: will it help for APA to do a prioritization round?

ra: would like to see all the issues in the new github repo

they will self-prioritize organically

so talk to you on github, we´ll do a call if needed!

CSS A11Y TF Inaugural Call

31 Jan 2017

Attendees

Contents

introductions

Round of introductions

Scope of our work

How and where will we work

speak property

Next steps

Summary of Action Items

Summary of Resolutions

Scribe.perl diagnostic output