Notes 30 October 2014 auto-WCAG call

Notes AUTO-WCAG 30th of October

Jesse: I've got a deadline tomorrow that I need to focus on, sorry!
Hanno will take notes
Annika will chair the meeting
Mathieu joined the team, he works for Tangaguru

1. action points  https://www.w3.org/community/auto-wcag/wiki/Action_items
a. Write eGovMon tests for 3.2 - Annika  -no time previous week to write test for 3.2 
b. Do further research on methodology to check the relative order of links for SC 3.2.3.  - Kamyar: started, next week
c. title is acceptable for an alternative in HTML5 if there is no text at all in alt - Katie:  DONE, see mailing list

note Wilco: we should take reviewed items from the list
Annika: person reponsible for the test should be responsible for the workflow
Decision: test editors are responsible to change the status


2) Automating keyboard accessibility checks
John: 'Automating keyboard accessibility checks': more research needed, brainstorm?
Wilco: check for onclick attributes
Hanno: if a picker (datepicker) has a field alternative, it will pass. we can automate things with a probability
Wilco: clickable div's are certainly a fail
John: we can find some automatic failures
Mathieu: relying on onclick is not reliable, there are other listeners as well

Decision: John will start tests for keyboard accessibility, John wil send a document 
Event listeners will be on the agenda next week and mailing list

3) Review https://www.w3.org/community/auto-wcag/wiki/SC3-1-1-xml-lang <https://www.w3.org/community/auto-wcag/wiki/SC3-1-1-xml-lang>
4) Review https://www.w3.org/community/auto-wcag/wiki/SC3-1-2-xml-lang <https://www.w3.org/community/auto-wcag/wiki/SC3-1-2-xml-lang>

3.1.1 avoid failure overlap with other test another message
sc3-1-1-xml-lang if this pass, still the other test can fail, we need aggregation
A final step should always give an outcome
Annika: But the outcome is not necessary an outsome for the SC, only for the test
https://www.w3.org/community/auto-wcag/wiki/Basic_aggregation_algorithm
Kamyar: how do we aggregate?
Hanno: difference between return a pass and a fail coverage
Wilco: all failures should be covered
Annika: with additional manual checks you can pass
Kamyar: How do we aggregate a situation where a "success criterion is NOT completely covered"
Wico: that should be described on the wiki and generally should lead to a canttell

5) Review http://www.w3.org/community/auto-wcag/wiki/SC1-1-1-css-image <http://www.w3.org/community/auto-wcag/wiki/SC1-1-1-css-image>
Frank: couldnt find css selector, so javascript selector used
Annika: definition of surrounding text, we should have a programmatic defintion
Wilco: there is a wcag defintion for context in links
John: example I saw last week: empty cells in data table, with background image in the cell, it can be information, for example yellow or grey stars in the background image
Decision: John will find the actual case, so we can see if we can automate testing for it
Mathieu: we made some tests with images and use of colors
https://github.com/Tanaguru/Tanaguru-rules-AccessiWeb-2.2-doc/wiki
Mathieu: Loading all images of a website is processor consuming, costed 1 minute or more
Hanno: caching?
Wilco: +1, we can add a note that is it perfomance consuming
Annika: we should also add these tests even if they are undoable processorwise
Mathieu will send a link and we discuss this next week

6) Review http://www.w3.org/community/auto-wcag/wiki/SC1-1-1-img-empty-alt <http://www.w3.org/community/auto-wcag/wiki/SC1-1-1-img-empty-alt>
not dscussed

7) Final comments
Wilco: please take a look at the tests for review right now