<scribe> scribe: shadi

Proposal: create "ACT Heuristics" as an "Accessibility Requirements Document" #428

<Wilco> https://github.com/auto-wcag/auto-wcag/issues/428

shadi: we have several checks that are not really conformance checks, but really good approximations
... think we should document these approximations as an output of this group
... and refer to these approximations in the rules, rather than directly to WCAG

wilco: maybe a question of how accurate is accurate enough?

audrey: OK with what shadi says
... can't be a failure but needs human validation
... only a partial check
... agree with shadi

<emma_jpr> Morning ... I can't link into the call, so trying to figure it out

wilco: RGAA has similar checks to these, right?

audrey: RGAA also considers assistive technology support
... different from WCAG in that respect
... we do not yet have a check for CSS lock
... but would likely have something like a composite rule

<emma_jpr> Thanks Shadi

dagfinn: think composite rule is the way to go, like we did for keyboard rule
... when the atomic rule does not cover all aspects of a requirement

wilco: why is assumptions is not enough

<emma_jpr> which issue is the current discussion?

<Wilco> https://github.com/auto-wcag/auto-wcag/issues/428

shadi: think it is a misuse of the assumptions

<emma_jpr> ta

shadi: we are redefining the requirement that way
... better to publish the pattern and refer to that

carlos: think these are two very different examples
... would agree that CSS lock is really screaming for a composite rule

wilco: non of the rules are really conformance tests
... they check if there is a requirement failure
... but they leave out conforming alternative versions for example
... think the CSS lock is similar to conforming alternative
... don't think it is Auto-WCAG role to police what vendors do
... but rather to collecnt and document

shadi: think conforming alternative and CSS lock are very separate
... it's about different degrees of confidence
... image without accessible name is known to not meet an SC
... CSS lock is only likely to be wrong
... about your second point, I think there is a certain level of agreement in this group
... otherwise we don't need to agree on any of the rules
... good to document what we think is a useful check

wilco: as long there are enough implementers content with a rule, then it is good enough
... question is how we deal with objections
... maybe few people objecting to a rule, but others accepting the rule
... how do we address that?

shadi: to be clear, don't disagree with the check itself
... just how to transparently the relation is expressed

carlos: if it is manual or semi-automated, then can refer to the SC

wilco: several ways to deal with such edge cases
... use assumptions, like here
... using composite rule
... or relating to best practice
... comes down to the question of how we address objections
... maybe have a policy that two or more objections requires the rule to change

emma: would always need to be part of a composite rule

wilco: i don't think so but majority decided

emma: we would likely have content that would be hit by such a check
... checking the exception is important

wilco: suggest majority vote to rules

emma: don't know for rules but for guidelines we do it differently

shadi: think voting can be tricky
... how to count active members
... what about minority perspectives
... conflict resolution etc

wilco: our Deque implementation is different
... if we change the rule we would lose that implementation
... useful discussion, need to iron out a proposal

emma: Google Accessibility Scanner has concept of warning

rssagent, make logs world

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)

Default Present: audrey, carlos, shadi, wilco
Present: audrey carlos shadi wilco
Found Scribe: shadi
Inferring ScribeNick: shadi

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: 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.)