<Jemma> Scribe:Jemma
simon: pilot test update - need to write up instruction for tester
system only allows two testers for now.
if that is the concern, we can change that before the pilot test.
simon: we can have multiple test cycles if you want to
mck: I wonder where those constraints are coming from.
simon & matt: it seems that there were miscommunication
simon: I can do some set up for testing
<zcorpan> https://github.com/w3c/aria-at/issues/162
jon: I invited Lucy Greco at UC berkely as the test.
simon: I need tester's github user name to add to the testing.
and invitee needs to accept the invitation.
issac: I had an update I would like to share.
AT and browser version can be added.
decending the result order by date would work.
this version will have independent sorting, not global one. - mainly AT version and ??
pass and fail is kind of strecth goal at this time.
https://github.com/w3c/aria-at/issues/159#issuecomment-628103259
mck: conversation about the result - no output, in correct output, correcct output
both no support and incorrect output are "fail"
incorrect output is a bug, more problematic for users.
mf: it will be ok for now to treat those as failure.
it is related to initial design
mck: we can make this as explicit regarding the concept.
mck: michael merged my changes
https://github.com/w3c/aria-at/issues/52
mck: there is runner bug for VO
mf: we are going to ignore it for now.
mck: runner in staging is using the most recent version, simon and issac?
simon, I can ask the team.
mck: does any one has a chance to check assertions?
simon: runner has the last version.
mck: https://github.com/w3c/aria-at/issues/52#issuecomment-622299312
rob: https://github.com/w3c/aria-at/issues/52#issuecomment-624803452
<zcorpan> scribenick: zcorpan
Matt_King: in every case the SR
can have many ways to support the same semantic
... how many of those we test is the main question
rob-fentress: (missed)
michael_fairchild: I think it's a
good idea to add
... same concern when you're trying to invoke the same
rendering that was tested in different ways
Matt_King: we should invite SR devs to review checkbox test plan
<Jemma> mck: we dont have a way to open the test plan from github, we open from html.
Matt_King: point them to test in GH pages, but isn't able to open the test case
<Jemma> we can use test runner but there is the complexity to do that.
Matt_King: could point to the new
runner, but it's not finished
... the group of checkboxes contains an unordered list. we
don't have any assertions related to the list semantics
michael_fairchild: i'm ok with preceding the way it is, we can add assertions later if needed
Matt_King: i think so
... should we always test every single a11y semantic in every
pattern?
... thinking ahead for some of the patterns, I don't think it's
what we want to do
... for some examples there are things to make the example more
realistic
... we should test all the semantics that are listed in the
design pattern that are included in the exmaple
... if there are additional elements within the example that
aren't explicitly called out they're not included for
assertions
michael_fairchild: the pattern description table in APG?
Matt_King: yes
... property, state, keyboard interface
... some of the listbox examples, with editable listbox, has
unrelated stuff in it
michael_fairchild: fine with that
Zakim: take up item 4
jongund: making progress on data
files
... setup scripts, focus on top level menubar
... updating setup script description
... looks like additional commands need to be added to commands
list
... we want to test review commnad
... review is related to toggling checkbox
... to open the menu, need interaction command to open
... go back to review mode
Matt_King: don't necessarily need
every interaction test in reading mode
... when you are in reading mode, the semantics that are
present are accurately present
... people don't normally operate menubars in reading
mode
... likely to encounter them in that mode though
... jaws has made it impossible to get to reading mode when
you're in a menubar
... i think it's a bug
... wanted to ask them if it's intentional
... might be a feature
... if so the assertions might need to be optional
... does it make sense?
... when you have a reading mode test, instruction is
(...)
... at some point, switch to reading mode
jongund: maybe eliminate test that need interaction before reading mode?
Matt_King: if you're in a menu
but need to read something else on the page...
... or read the menu item itself
... might not be necessary to test
jongund: focus changes may do things
Matt_King: within the submenus,
we don't do reading mode tests
... only on parent
jongund: cuts down differences between screen readers
Matt_King: process question...
number of places, don't know if number of commands is
correct
... should i make the changes in csv file?
... make a PR, do you have an open PR?
... or should i describe what should be changed?
jongund: describe
rob-fentress: my comments in
menu... there's testing of whether the group role and group
name are spoken
... using interaction mode
... in VO, group role and name aren't announced
... if you switch to what might be considered reading mode in
VO
... then it does announce
... if you're not testing in that way, might miss support
Matt_King: fundamentally
important for apple - they regard VO as modeless
... the commands work all the time
... NVDA has an equivalent, but is advanced
... VO ... some people only use VO commands all the time
rob-fentress: in NVDA/JAWS (missed)
Matt_King: we need VO specific
assertions
... related to navigate by item
... VO test might be to ... I don't know how to word what the
test would be
... navigate menu by item
rob-fentress: not using web page commands
Matt_King: right, VO commands
rob-fentress: there's sort of a
mode in VO, group mode
... different results, maybe don't people don't use
... number of items in a menu
... other mode, DOM mode? doesn't seem to
Matt_King: rob-fentress are we on the same page that we need VO-specific tests
rob-fentress: there's a difference so yeah
jongund: item mode?
Matt_King: want to avoid the word
item
... I can propose wording
... something along the lines of navigating the menu...
... next element
... we have sequential navigation
... with arrow keys
jongund: other sequential navigation things?
Matt_King: most of the time we
don't need separate sequential navigation task
... additional task that applies to VO that doesn't apply to
other SRs
... would be an additinoal test
... different from more commands to an existing task
jongund: from a user standpoint, just reading radio items
Matt_King: users navigate in
different ways...
... VO commands as their first preference
... some users learn that they can also navigate using the web
page commands
... because of the way VO works on the mac, and don't learn a
Windows screen reader
... a lot of people use apple's native menus with VO, use VO
commands
michael_fairchild: role might announce that they can use web page commands
RRSAgent: make minutes
This is scribe.perl Revision: 1.154 of Date: 2018/09/25 16:35:56 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00) Succeeded: s/out/output/ Succeeded: s/no output/no support/ Present: Jemma Matt_King shimizuyohta isaacdurazo michael_fairchild jongund rob-fentress zcorpan Found Scribe: Jemma Inferring ScribeNick: Jemma WARNING: No scribe lines found matching previous ScribeNick pattern: <zcorpan> ... Found ScribeNick: zcorpan ScribeNicks: zcorpan, Jemma WARNING: No date found! Assuming today. (Hint: Specify the W3C IRC log URL, and the date will be determined from that.) Or specify the date like this: <dbooth> Date: 12 Sep 2002 People with action items: WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option. WARNING: IRC log location not specified! (You can ignore this warning if you do not want the generated minutes to contain a link to the original IRC log.)[End of scribe.perl diagnostic output]