W3C

Accessibility Conformance Testing Teleconference

08 Aug 2019

Attendees

Present
Wilco, KathyEng, MaryJo, Anne, Romain, MoeKraft
Regrets

Chair
MaryJo, Wilco
Scribe
Romain

Contents


ACT-R F2F in Copenhagen - https://www.w3.org/community/act-r/act-rules-cg-face-to-face-meeting-on-24-25-september-2019-in-copenhagen-denmark/

maryjom: have everyone signed up?

Wilco: I am going

romain: I won't be there

anne_thyme: me neither
... but maybe I'll stop by
... who owns the sign up?

Wilco: shadi is organizing

anne_thyme: OK, so we'll wait for him to get back to get an estimate participation number

Reaching out to AC Reps for their approval of spec going to Recommendation.

maryjom: we want to get approval from our AC Reps to go to Rec
... please prop them to vote

Wilco: I've reached out to John

anne_thyme: I'm the AC Rep so I'm getting it done

maryjom: I'll reach out too

Wilco: it may also be worth telling your AC Rep to tell other AC Reps to do the same

Rule review - HTML Page has a title: https://act-rules.github.io/rules/2779a5

link to the survey: https://www.w3.org/2002/09/wbs/93339/PageTitle/

Wilco: before diving in the details, maybe we can talk about what we think with the surveys in general

maryjom: yes, I found the answers didn't always fit
... my answers often went to the open questions category

anne_thyme: it worked pretty well for me
... I did also put a lot in 7 (misc stuff)
... also, the "are there any remaining issues" item may not be needed as we can look that up in the tracker

+1

Wilco: yes, I think the question was to entice people to look at the list and tell if anything was blocking

anne_thyme: OK, then the question might need to be rephrased a bit
... like "which issue do you consider as blocking"?

kathyeng: I wasn't sure whether the implementations were OK

maryjom: yes, I found the same thing
... we can go through the results

https://www.w3.org/2002/09/wbs/93339/PageTitle/results

scribe: about the ACT format consistency I wasn't sure that the order of the sections respected the suggested order in the spec
... also each of the implementations results point to a different result file, so you couldn't tell for each of the test if all implementations achieved the same result
... it took me a while to find there was a description too, as it was burried after metadata

Wilco: would you like to report something to ACT-R?
... I'll pass along the feedback

romain: I thought I had answered this survey, but my answers don't show up there…

maryjom: there is a slight mismatch in the section headings compared to the ACT spec, so it made a little harder to review
... though I'm sure you had a reason to use this order

Wilco: right, I'll take that feedback to the group

maryjom: OK.
... Anne had a comment that the applicability is ambiguous

anne_thyme: I was also looking at the rule "Document title is descriptive", which is a sibling rule to this one and is sructured a bit differently

Wilco: I noticed that too, and think that "page" needs to be defined here

anne_thyme: also, rules talking about the same things should use the same terminology (e.g. page vs document)

romain: I agree with the previous comments about consistency and unambiguous definition

maryjom: about assumptions, Anne can you describe your issue?

anne_thyme: we don't talk about the location of the `title` element in the HTML (in `head` or in `body`)…

maryjom: to me, all the browsers do the same computation and the first title becomes the computed title
... if they all agree, having a second title doesn't matter and it is ignored
... it sounds like a good idea to document it

Wilco: I agree with documenting it, I don't think it's an assumption

anne_thyme: it can be a note

Wilco: right, a note would be better

maryjom: Implementation data
... RGAA passed the last 3 failed/inapplicable tests, can it then be considered a correct implementation?

romain: what do reviewers need to do here about the implementation results? should we re-do the evaluations? check their consistency?

Wilco: part of this comes down to trust, do we trust the group or people who provided the data?
... in the system, there was a bug that something wasn't implemented, and we could review that

maryjom: it would be nice to have a table that put all the results together

Wilco: a table with all the test cases and the outcomes?

maryjom: yes, it would be easier to review
... and we could even highlight the ones that have inconsistent results

<anne_thyme> example of table listing implementations (in browsers): https://www.google.com/search?q=browser+support+table&tbm=isch&source=iu&ictx=1&fir=OQpRDFjyhYW0cM%253A%252Cg2fWkKzLs3NLTM%252C_&vet=1&usg=AI4_-kSN9bx2q8ZZJmqGaV8b67htlI6w2A&sa=X&ved=2ahUKEwjTosTLsfPjAhVt16YKHRFFAdAQ9QEwA3oECAYQDA#imgrc=OQpRDFjyhYW0cM:&vet=1

[pic of the implementation tables used on MDN]

anne_thyme: it seems that if multiple implementations got it wrong, then there's probably something wrong in the rule itself

Wilco: putting my ACT-R CG chair hat on, one of the bigger concern I have is that people will be hesitant to submit their data if we make it obvious who doesn't implement this and that
... tool developers are volunteeringly contributing the data

kathyeng: when I was submitting our results, seeing others' results was actually useful

maryjom: I think it would be interesting to know how they feel about it
... it's gonna point out flaws, but is useful for the developers to know what needs to be improved, and for rule authors too

kathyeng: maybe if we don't hilight it in red, or don't bring attention to it

Wilco: OK, I'll give that feedback to ACT-R

<Wilco> https://act-rules.github.io/pages/implementations/mapping/#automated-mapping

kathyeng: in the wording '"is the implementation data correct", I didn't know what "correct" means: is the data correct? are the implementations correctly implementing the rule?

Wilco: the page linked above defines what the ACT-R community considers a correct implementation

maryjom: is "untested" and "inapplicable" the same thing?

Wilco: I think there are two things that you can call "untested": new test cases, and also the tests that are not covered by the tool, which would be "inapplicable"

kathyeng: looking at RGAA, they would be a partial implementation for two test cases, so would they be a partial implementation for the rule?

Wilco: yes, and that they appear is a bug in the system
... an implemenetation needs to be complete to be listed
... I think that a better solution is that if you looked at it and decided it cannot be tested, then mark it as "inapplicable"

kathyeng: in the question for number 4., can you link to this ACT-R page for reference?

Wilco: yes, can do. but do we agree with this mapping?
... this is something I came up with and proposed to ACT-R, so we need to agree in the TF too

kathyeng: I would want to discuss why Passed is allowing the outcome inapplicable

Wilco: do we want to do this now?

maryjom: I wonder if we can table that question a little and go through the rest of the survey first
... the next question was 5. Consistent with WCAG
... [explains a question with iframes]

anne_thyme: it's the page that must have a title, not iframes

maryjom: I was looking at test case #2

anne_thyme: I think the test description is misleading

Wilco: OK, we can fix the description there

maryjom: Kathy made the comment that the WCAG techniques say a non-empty title appear in the _head_ section

Wilco: this is a sufficient technique
... if you put a title, you definitely should make it in the head, but it will also work if you don't

kathyeng: yes, I think the previous discussion about the informative note will clarify that

maryjom: the next question was "6. Remaining open issues"
... Anne had a comment about the naming conventions with the sibling rule "Document title is descriptive"

anne_thyme: right, it's not listed in the issue list, it's about all the rules

maryjom: ok, it's a general statement
... 7. other questions or concerns

Wilco: I agree with Mary Jo's comment

anne_thyme: my question was about pointing to WCAG 2.1 or 2.0

maryjom: also, the change log is very long

Wilco: right, we can put it somewhere else or hide it

maryjom: 8. Readiness for publishing
... Kathy is asking if we can publish a rule when it doesn't have enough implementations

Wilco: that needs to be changed

maryjom: we must have two correct implementations?

Wilco: 3 is what we're targeting, but the TF can decide otherwise
... the CG targets 3 before we decide it's ready to be shared with W3C

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version (CVS log)
$Date: 2019/08/11 11:19:54 $