<trackbot> Date: 26 September 2014
<richardschwerdtfeger> scribe: MichaelC
<richardschwerdtfeger> meeting: ARIA Editors on Accessible Name Computation
<clown> http://rawgit.com/w3c/aria/master/accessiblename/accName.html
rs: doc on github for name and description?
<clown> https://github.com/w3c/aria/commits/master/accessiblename
clown: ^
have been doing a bunch of stuff with it
rs: input from cynthia?
clown: no, hasn´t been at AAPI meetings
due to work conflicts
I´ve added introduction (still room for improvement)
in algorithm, have removed references to HTML or made clearly example or added equivalent SVG ref
rs: got input from from JamesC
ARIA takes precedence
then you drop to native host features to compute name / desc
need to refer to host language specs for those
sounds logical?
clown: to paraphrase
after core a11y computation
there would be a section on changes due to [HTML | host language]
rs: Steve was trying to simplify
authors having a hard time
clown: but not for authors
rs: yeah, but...
clown: do need author guidance somewhere
<SteveF> http://rawgit.com/w3c/aria/master/html-aam/html-aam.html#accessible-name-and-description-calculation
sf: my simplification wasn´t for authors, was based on implementer feedback
if you have aria or aria-label, use name computation
otherwise use x-and-such in a particular order
the attributes and order is pretty much how it´s implemented based on my testing
clown: which user agents?
sf: the four majors
most of it is pretty standard, though a few differences
need to do more testing and documentation
placeholders still problematic
clown: still exist
sf: yah-huh
widespread
rs: know why people don´t want it, but on mobile authors will frequently prefer it
throwing label off screen doesn´t help sighted people
clown: sometimes it´s more guidance than label, such as ¨put stuff here¨
anyway, it´s low in precedence list
sf: yes, based on my testing
rs: I bet labelledby and label will be consistent
then you get into specific elements
SVG doesn´t have that because they don´t have form controls
though they have title and desc descendant elements
<SteveF> http://www.html5accessibility.com/tests/placeholder-labelling.html
if we cover this in the main spec
then we have to get into the recursion with labelledby
in SF doc, see it´s under each element
can it be in its own section?
sf: ?
rs: oh, I see, it is under a name and description section
Ok, so that seems to be the break-down
first aria-labelledby, and talk about concatenation process
Another wrinkle is CSS ::before and ::after
doesn´t work in SVG
clown: I´ve addressed already
doesn´t work on elements w/o a content model
e.g., doesn´t work on <img>
might be why it doesn´t work in SVG
so my spec wording uses that assumption
of course, possible it´s just an imcomplete implementation issue
So, if it turns out it works only with HTML
so put in HTML spec?
what if other techs later implement?
mc: we can move which spec it´s in at that time
clown: <concatenation algorithm>
rs: So, there are two things to do
tell people about name computation - authoring practices?
<clown> http://www.w3.org/TR/2009/WD-wai-aria-implementation-20090224/#mapping_special_te
clown: ^ is very first published version of the algorithm
it´s similar to what Steve recently did
would like to take the current version and put it more like that style
rs: that´s fine
plus section for host languages
sounds reasonable?
sf: at quick glance looks good
want to look more closely
clown: want to take the best of that style with the best of the current algorithm and smoosh them together
sf: let´s set up some test cases to see what´s actually being done
find common ground areas, and file bugs against implementations to steer towards a common denominator
then document that
rs: so where we refer to the algorithm
add a step for the native host language
<SteveF> Accessibility for Docs editors https://support.google.com/docs/topic/6039805?hl=en&ref_topic=1382883
<SteveF> any of those docs "If you're using NVDA, we recommend using Mozilla Firefox as your browser."
MC, what is status of repository?
mc: have been waiting on Jason to re-org
I can if he´s ok with it
clown: need to freeze edits during that
mc: let me ask Jason and then put on agenda for next week´s editors´ meeting
then we can schedule the freeze
clown, sf: <looking at details>
mc: do we want to steer authors to the algorithm?
or create an easier to consume version for authors?
clown, rs: something for authors separately
mc: loop in Authoring Practices editors
rs: where do we put it?
authoring practices is large
should refactor
mc: the editors wanted to refactor
clown: would like an author-oriented section in our doc
rs: works for the core, but what about host languages?
prefer in the authoring practices, and pull in both the core and language-specific stuff for that
<scribe> (new authoring practices needs SVG stuff)
or have language-specific authoring practice versions
also noting existence of Using ARIA in HTML
<clown> accname-aam-1.1
rs: what´s the name of the doc?
(shortname)
mc: placeholder for now
will have to solidify on FPWD
don´t worry about filename in repository, can move
mc: one doc with sections for the techs, or separate for each tech?
rs: sounds like we have agreement on way forward
some details to be sorted
clown: feel free to add to the doc
rs: what do we do in the ARIA spec?
mc: refer to the new doc, remove the existing section
clown: supposed to give JamesC an action to do
want to wait until after re-org