<shadi> https://lists.w3.org/Archives/Public/w3c-wai-ig/2018JulSep/0001.html
Wilco: The wide review draft is out. Please review thoroughly and pass it on to others for review
Shadi: Yes, this is the wide review, so please spread it out widely. And review in details
<shadi> https://www.w3.org/community/auto-wcag/2018/07/03/workshop-2018/
Wilco: There will be an Auto-WCAG face 2 face meeting in London the week after TPAC.
<shadi> https://www.w3.org/2002/09/wbs/1/autoWCAG2018/
Wilco: everyone from ACT TF is invited. Please fill out registration
Wilco: When Shadi, Kasper and I met 2 weeks
ago, we discussed possibly extending EARL to cover the needs of the ACT
Format
... it's an open question if we want to do this and what the format will
be
... Shadi, you suggested that a working group note might be the format
Shadi: Yes
Wilco: And it makes sense since EARL is also a working group note
Shadi: We can publish an erratatory, saying
"these are the things that are wrong with EARL", leaving it up to future
work to address. Or we could be bold and update it ourselves.
... (missed by scribe)
Wilco: There are other things that I might
want to do with EARL, and WAI-Tools project might have some resources
for it
... There is also the Accessibility Statement to consider
... It woulc be nice to have a shared vocabulary for all of these
things, bringing it all together
<shadi> https://www.w3.org/TR/act-rules-format/#output
Shadi: From the ACT perspective, if I look
at section 14 of the spec, the output format (see Shadi's link). The
syntax needs to be better described
... This is the minimum set of elements or vocabulary that needs to be
used by the ACT. This is one requirement
... everything else is up to us to decide, but we need to think about
the timeline
Wilco: I like the idea of starting a note
... I want to understand the requirements better
... You say it can't go to a recommendation the way it is formatted now?
Shadi: Yes, the way it is put now in the ACT
Rules Format is not normative. We need to define that syntax in some
form. We could define the minimal syntax in this document, or we could
refer as a working group note
... But it can't be normative and reference something outside of the
format
... EARL is a working group note, so it's not stable in any way
... There is a dependency between this document and EARL, and EARL could
be changed any day, since it's just a working group note
... Whenever there is a dependency between two specs, it's problematic,
even more when one is not a recommendation
... We need to look more carefully into this part. Maybe we just need to
formulate something differently
... This was just from the ACT perspective
... I assume the JSON serialization of EARL could be buggy. Deque and
Siteimprove have the pleasure of starting to implement using it
... File issues on same repository
... I will start by adding those comments from the delivery to the
European Commission
... When we have this, we can consider whether we just want a simple
update to EARL and how we want to do it
Wilco: It will be interesting to see if there is enough interest in an update to EARL
Shadi: I will start that errata file, or "Proposed changes to EARL"
Anne: Kasper also started filing issues for EARL in the ACT github space
Wilco: I think that is the right place for
it
... There is two things here, first what we need for the ACT Format to
go into recommendation. And then secondly what would be nice to have
otherwise
Wilco: I think we should have "updating
EARL" and "exit criteria" and "collaborating with Silver TF"
... I have reached out to Silver chairs to find out how we can work a
little more closely together
... They are both a Task Force and a Community Group now
... They are also looking at some less quantifiable things that they
want to include into Silver. We will have a conversation about what we
in ACT TF can help with
Shadi: I think what they are looking at is less binary, more gradients or prioritization
Anne: We suggest looking at the review process or defining gatekeepers for ACT rules
Wilco: This also depends on how we end up
relating to failure techniques
... I hope that this will be solved before TPAC
Shadi: I don't think the failure technique
discussion is the most important thing, I think the process update in
Auto-WCAG is more important
... If we let Auto-WCAG handle the development, we still need
gatekeepers at certain points
Wilco: I put down "process"
Shadi: This is something that would be good to involve AGWG people in at TPAC
Anne: Are we completely done with the Atomic vs. Composed rules discussion?
Shadi: I reached out to Alistair already for formal comments on public draft. I will ask Kathy too
<shadi> https://www.w3.org/WAI/GL/wiki/WCAG_2.1_CR_Exit_Criteria
Wilco: We need to come up with a plan, a list of things to do to prove that the ACT Rules Format works
Shadi: Here (see Shadi's link) is a list of
exit criteria for the WCAG 2.1 recommendation, as inspiration
... We have to think about our own conformance requirements. Features
that could be on or off. Obviously rules that conform
... We could have "X no. of rules that are manual, automated,
semi-automated"
... "X no. of rules that have some property to them"
... which output format they support
... that have optional things implemented
... We should have a good, representative set, that demonstrates the
applicability in different settings
<Wilco> https://www.w3.org/TR/act-rules-format/#quality-accuracy
Wilco: Could it be some of what we have in the Accuracy Benchmarking section already?
Shadi: According to the documentation, each feature (SHOULD, MAY, MUSTs) needs to have to implementations
Wilco: So we need to come up with a list of
these features
... and from there, we can write our exit criteria
... I will start a github issue for this
... Input for exit criteria can just go as comments for the Github issue
... Do we just for all of the MUST, COULD, SHOULD?
Anne: We talked about having a specific number of atomic + composed rules
Wilco: Do we need to have a certain number of implementations?
Shadi: That has been asked
... Ideally, we should show implementations of ACT rules
... I think we have a fair amount of implementations
... We shouldn't aim too high, reviewers are always tempted to ask for
more. We can always surpass the tresh hold
Wilco: Timeline wise, we need to go into CR before TPAC, right?
<shadi> https://www.w3.org/WAI/GL/task-forces/conformance-testing/wiki/Month_by_month_plan
Shadi: Here is one timeline that we have
(see Shadi's link)
... Here we said September to December for CR
<shadi> https://www.w3.org/2017/01/ag-charter
<shadi> October 2018: CR for ACTF 1.0
Shadi: The charter says CR in October 2018
... and October 2019 for recommendation, so we actually gave ourselves
quite a little time
Wilco: So that means that we need to go into
CR before TPAC.
... Do we need the exit criteria before we go into CR?
Shadi: Yes
Wilco: Then this is our next high priority
item
... Homework for everyone: Look at what they think should go into the
exit criteria
... I will send out email
Shadi: Actually in our plan we said March
2019 for recommendation
... We should have everything wrapped up pretty much by end of
September. It takes a bit to go through the process for CR