See also: IRC log
<mck> trackbot start meeting
<trackbot> Meeting: Protocols and Formats Working Group Teleconference
<trackbot> Date: 16 November 2015
<jamesn> scribe: jamesn
MK: should we have something in APG about browser support?
MB: limit to words but not
numbers
... going to be a chore to update to browser versions
MK: we didn't make a decision to
support major bersions of browsers
... our task is to provide practices which support th
specification
... we need to make some sort of statement - find where we last
discussed this
... pretty sure we made a resolution that we would not make an
effort to update every example in response to eery screen
reader or browser version change
... examples are not how to work around quirks
... best practices for compliance to spec. our expectation that
browser and screen reader devs would use them to guauage their
support for the design patterns
... recall that if there is some importatnt functionality that
is broadly used and have to do something quirky we might
include that but would not be our general practice
JN: 2 different examples - when things dont work completely and when things are completely broken
MK: i think the aria-labelledby
example is slightly better for both NVDA and JAWS but the other
works in some other cases and should be sufficient
... shoudl we create a bug for a section to draft our browser
and AT approach
https://www.w3.org/Bugs/Public/show_bug.cgi?id=29297
MK: in the example there is @alt
on the mixed checkbox image
... there really should be null alt text
Jemma: google doc for screen reader testing
JG: sometimes NVDA does read the
grouping label, sometimes it doesn't
... with radios seems consistent, but checkboxes seems to be
not consistent
... consensus to use fieldset/legend
... ?
MK: both are valid
JG: fieldset legend leaves it up
to AT
... the 2 techniques will affect the user experience
MK: don't know if that will always be that way
<MichielBijl> JN: did you test between fieldset legend and role=group labelledby?
MK: looking at the spex we were 99% certain we can use fieldset/legend in this way
JG: that is intersting
... in the a11y api how is the FS/Legend exposed to at
... the AT finds the grouping label and then reads it
... think that is an interesting authoring issue
... the 2 techniques are saying different things about the
accessible name
... the grouping thing is important in certain contexts
MK: the ATs should always provide
access to the grouping label in all contexts even if they don't
read it automatically
... all ATs have a command for reading the current element and
its label and grouping container should probably get read
there
... the only one that does that consistently is JAWS but not
even sure that is accurate now
JG: shouldn't be comparing aria-labelledby with FS/Legend but should instead be comparing with role=group
discussion on if group works
JN: they should all be valid
MK: in the wild we see all 3 approached
JG: newbie devs want to
overcontrol the experience
... they will latch on to the aria-labelledby example
MK: none of the screen readers are reading the grouping label from the currently focussed item
discussion of aria-describedby
JG: does role=group work for links
JN: no idea
MK: if you wanted the group to be
a single tab stop
... Jon and Jemma will add a 3rd version with role=group
<MichielBijl> http://s.codepen.io/boomerang/a62554037b0f8d25269b06c245ea8e031447699204201/index.html
<MichielBijl> scribe: MichielBijl
MB: is this what we mean?
MK: we want to show the different
ways to achieve the same thing
... we don't want to say how particular cases work
... what it's doing is explaining redundant code per spec
... redundant coding can produce redundant speech
... that could be a warning
JG: developers don't want to be redundant
http://s.codepen.io/boomerang/a478ae4e07dc1cb260899459db247af21447699453614/index.html
http://s.codepen.io/Michiel/debug/mevpMa
Zakim nextitem
Zakim next item
MK: any updates?
MB: my name still shows after
alert
... that is done
MK: no edits?
BG: there are more ways to
trigger a alert
... you can add a role of alert to a div that is already
there
... you can also add content to a div with role=alert
MB: do we want to implement different ways?
MK: alert developers of ways that don't work?
BG: if the alert is off screen it doesn't work
MK: maybe put in a note?
MB: rather show different ways of doing it than a note; that list might grow
MK: another way is to just unhide
it with CSS
... is html5 hidden ready?
MB: you can use that, and have a fallback with a CSS attribute selector [hidden] { display: none }
MK: can you make that?
MB: sure
MK: for all intends and purposes
this section is blank slate
... one choice is to not have a specific design pattern for
landmarks
... Ann was pretty opposed
... we could have a design pattern here.
MB: landmark navigation could have inline links
MK: other parts of the APG
JG: I'm willing to take this
on
... I can use this for examples
MK: do you wan't to work on the description?
JG: I will, and see what people think
MK: just take out te keyboard interaction section?
JG: yes, or provide a note
JN: “this is not your
problem”
... we have much more important things to do and I'm in favour
of removing it.
JG: I think it's important, people are confused about landmarks
MK: propose a description. I do tink we have some example work that is higher on the list.
JG: over the summer we have a new student helping us with the examples
<jamesn> don't forget we also have this in the APG
<jamesn> https://rawgit.com/w3c/aria/master/practices/aria-practices.html#h-kbd_layout
JN: can I get a task and just remove (X) from the spec?
MK: whole section doesn't really make sense
MB: I'm interested in what Jon comes up with
JG: I'll try to harmonise
MK: no, don't look at this. Write
something, and we might delete this section
... do we need to discuss headings and heading level in the
guide?
... we need to review section 4.2
Next item: be done!
MK: probably no meeting 21st and 28th of December
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/JG/MK/ Found embedded ScribeOptions: -final *** RESTARTING DUE TO EMBEDDED OPTIONS *** Found Scribe: jamesn Inferring ScribeNick: jamesn Found Scribe: MichielBijl Inferring ScribeNick: MichielBijl Scribes: jamesn, MichielBijl ScribeNicks: jamesn, MichielBijl Present: matt_king James_Nurthen jaeunjemmaku Michiel_Bijl Bryan_Garaventa LJWatson Found Date: 16 Nov 2015 Guessing minutes URL: http://www.w3.org/2015/11/16-aria-apg-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]