See also: IRC log
Mary-Jo: The public draft is ready to publish, and there are still a few minor things to sort out before it will be published on Wednesday/Thursday
<scribe> ACTION: Shadi: where are we with the accessibility issues in the draft [recorded in http://www.w3.org/2017/09/11-wcag-act-minutes.html#action01]
<trackbot> Created ACTION-22 - Where are we with the accessibility issues in the draft [on Shadi Abou-Zahra - due 2017-09-18].
<maryjom> https://www.w3.org/WAI/GL/task-forces/conformance-testing/wiki/ACT_Review_Process
MaryJo: Shadi wanted to have a short
discussion on the review process - i.e. the process of making sure that
the rules are ready for publication as ACT-rules. Contribution, review,
review with the AG etc.
... We have to find the balance between openness and rigourous
submission and review process, and wants to know what peoples views are
on this
Charu likes the flow where rules go through review, then implemented and then wetted and tested. On the comment: can we reject something? If something doesn't meet the requirements it automatically gets rejected
Would we accept something that doesn't contain all the parts from the rule format
MaryJo: do we want to document these things in the process document - i.e. potential reasons for not accepting
Stein Erik: Thinks it is important to require a minimum of information to ensure a higher standard to the ocntributions
MaryJo: Thinks we should accept only submissions that contain the required parts of the ACT format
Charu: We currently have certain points, but doesn't think they are explicit enough to discard contributions or to decide if they are accurate enough to be accepted
MaryJo: we need to define a level of requirements for accpetance
<scribe> ACTION: Shadi: Establish a minimum level of requirements for accepting a contribution [recorded in http://www.w3.org/2017/09/11-wcag-act-minutes.html#action02]
<trackbot> Created ACTION-23 - Establish a minimum level of requirements for accepting a contribution [on Shadi Abou-Zahra - due 2017-09-18].
MaryJo: Shadi didn't put in an action number of implementation. We would need a number of implementations before passing on to the AGWG. Typically for most things in the W3C two implementations are required.
Mo: Isn't it two for a success criteria or specification. Should this apply to a rule? It could be accepted as a draft until a second implementation is in place
Sharu, Shadis responce is that we should work into the criteria; reasons for acceptance/non-acceptance.
Charu: We have to define what lack of
quality means
... criteria would fall under submission
MaryJo: reasons for rejecting could fall under the peer review to know what we are looking for
Charu: should go in both. We do not want to waste everyones time if it doesn't conform to certain criteria - also after peer review
Anne: does it say anywhere that we are looking for the balance between false positive and false negatives?
MaryJo: where the rule make generate more false positives the number of implementation may have to be higher
Mo: suggests to include in the maintainance section a clause on the process if implementations of rules show too many false positives/negatives
Anne: Going back to adding the reasons to drop a rule if it lacks quality. Should that be its own sections since rules could be dropped at all stages in the process - e.g. in maintainance phase
Mo: Instead of deleted we deprecate or quarantine
MaryJo: Would be nice to have states of the rule going through this process. By attaching to the rule in submission which state it is in. Are there issues? Is it in implementation etc
Anne: Didn't we agree to use tagging for the state rules are in?
MaryJo: It should be added to this review process document
<maryjom> https://github.com/w3c/wcag-act-rules
MaryJo: Looking for volunteers to expand the README with a section on how to contribute
Mo: We do have a contribute.md for the rules format to use as a guide for the rules
<MoeKraft> https://github.com/w3c/wcag-act/blob/master/contribute.md
MaryJo: looking for volunteers to ajust this. We would want to point to the review process and define what are the required parts of a rule som submitters will know what we are expecting
Stein Erik: Suggests criteria should be kept in the review process document
Mo: It is important to have a reference to the requirements and keep it fairly short
MaryJo: give contributors q pointer to the
resources they will need, an example, and anything to make the process
simpler. There has to be a balance between "learn by doing" and being to
prescriptive
... TPAC registration still open with a deadline for early registration
in October.
... Our next meeting will be on 18th of September