SV_MEETING_TITLE -- 21 Jan 2016

<scribe> scribenick: mgylling

Schedule changes (based on ARIA 1.1 moving)

tzviya: yesterday, you decided to push out the pseudo-last-call

rich: we’re gonna do a heartbeat instead

tzviya: by when?

rich: by CSUN

tzviya: will that included information re replacing describedAt?
... the core spec has a trickle-down effect for both DPUB and EPUB

rich: I logged an issue after the meeting for extended descriptions, going to look at link-type as well

tzviya: I think Joanie was working on that

janina: issue 2006

<janina> http://www.w3.org/WAI/ARIA/track/actions/2006

<janina> Above opened by Joanie

rich: I also put one in for extended descs

<janina> Rich's issue is:

<janina> http://www.w3.org/WAI/ARIA/track/issues/1009

tzviya: so the plan is that both of these will be in heartbeat by CSUN?

rich: provided we can get the group to sign off, yes
... we have an errormessage attribute, and I think we might follow a similar strategy there
... we can either do two things: either an errormessage that takes an ID, and we will say that it has to be visible
... we can actually annotate the target with a special attribute. Question I have, are extended descs always invisible?

Tzviya: no

janina: use MQ for that

rich: ok so thats a different strategy then
... we have to determine if its hidden, what do we do with it

tzviya: we talked about a toggle option

rich: if its invisible, do we not map it?

<tzviya> extended description analysis: http://w3c.github.io/dpub-accessibility/extended-description-analysis.html

rich: if we have a relationship that goes to an extended desc that is invisible, we do nothing with it
... if the MQ is in place and its visible, then we show the relationship

tzviya: we are trying to accomplish that its accessible to AT, but not necessarily displayed

janina: only for non-visual AT users?
... sighted users are going to want to see the text

tzviya: thats why we talked about the toggle

s/visible to AT/exposed to AT

rich: the importance of the extended desc is that its not just text, contains structure, if you make it hidden its just a string
... for it to be available it needs to visible in some way, and that would be the MQ
... if there is an extended desc, and we want to be able to say there’s an ED, and you can’t get at it unless you turn on the MQ.

janina: this would support having markup available
... because we’re hiding with the MQ

<tzviya> https://lists.w3.org/Archives/Public/public-aria/2016Jan/0031.html

tzviya: we’ve talked about a theoretical MQ, what does it look like?

rich: the MQ would be something on the order of “show extended descriptions”

tzviya: most MQs I’ve seen has to do with screen size or device...

ivan: for me MQ is a CSS feature, we cannot define something for CSS

rich: this has to be done in the CSS WG

tzviya: so we dont really have a solution yet, if we dont have a functional MQ

janina: we need someone to get busy in CSSWG to do this

ivan: the TODO list of the CSSWG is enormous

https://www.w3.org/TR/mediaqueries-4/#custom-mq

ivan: the point is that at the moment, we are trying to push reading systems to rely on provided CSS, IDPF wants to avoid extensions
... its a very long process

rich: we’ve got a binary MQ for inverted colors

matt: is it valid to do user preference in MQ though?

janina: sure

<richardschwerdtfeger> https://drafts.csswg.org/mediaqueries-4/#inverted

<ivan> @media monochrome { ...}

rich: jcraig offered to do the ED media query for us

janina: we want to do this easy now and get some implementation experience
... I can check with jcraig if he still is willing to do this

tzviya: I like to think about this as for ARIA in general, not for DPUB only, it will go into the ARIA spec

ivan: I think we should involve Florian as well, he’s an MQ editor and on DPUB IG

rich: one more thing: the OS does not have this functionality, so the setting would have to be in the browser
... the ability to tell a web app that I want ED exposed has to be set somewhere
... there’s got to be a mechanism to make the change
... it would take a lot longer to convince OS owners
... reading systems would then need to have the ability to go in and turn the setting off or on in browsers

janina: in IndieUI, we were thinking of a set of personalization settings like these
... but that WG hasnt kicked off yet
... the personalization stuff ended up in the ARIA charter

tzviya: DPUB will get involved once ARIA personalization starts
... ARIA 1.1 will remove describedAt, in the near term, what are authors supposed to do?
... we dont have a mechanism for distinguishing arbitrary details elements from an extended descripton details

janina: we were going to disambiguate that right?

rich: yes I logged an issue

janina: a bunch of stuff is stacked up against it
... change the definition of alt, etc

rich: let Joanie know if you have more values to add into her work on linktype
... the CSS work should come from APA, not ARIA

ivan: coming back to where we are: details is i the WD of HTML 5.1. This is as far as I understood to get it out as REC at the end of this year?

janina: yes

ivan: caniuse claims that details is implemented only by webkit

<tzviya> https://lists.w3.org/Archives/Public/public-aria/2016Jan/0031.html

<tzviya> wroing link

<tzviya> http://caniuse.com/#search=details

janina: think we got message from MS that it is one the agenda for Edge

rich: and firefox is working on it

tzviya: even if FF is working on it now, we dont have a timeline do we?

s/<aria/<a aria

markus: details may work even if its not implemented natively; we dont use any of its intrinsic behaviors do we?

ivan: we need to talk to Florian about timeline for MQ4
... I would not rely on having MQ in place for this

<richardschwerdtfeger> https://bugzilla.mozilla.org/show_bug.cgi?id=591737

markus: Ivan right, but at least some of the use cases are covered even without it

<ivan> markus, I agree

- DRAFT -

SV_MEETING_TITLE

21 Jan 2016

Attendees

Contents

Schedule changes (based on ARIA 1.1 moving)

Summary of Action Items

Summary of Resolutions

Scribe.perl diagnostic output