W3C

Accessibility Conformance Testing Teleconference

24 Jul 2018

Attendees

Present
Shadi, Wilco, Trevor, Moe, Alastair, Michael
Regrets

Chair
Wilco
Scribe
Shadi

Contents


Discussion with AG WG chairs - publishing rules through AG (process and where do they wind up)

<wilco> https://github.com/w3c/wcag/pull/381

WF: tried out how an ACT Rule would look like in the format/layout of a Failure Technique
... lots of overlaps but some differences too
... is that a direction we want to go?

AC: so the question is, do we want to incorporate all the rules and/or replace the Failure Techniques
... question is what is the volume

MC: wouldn't follow the format for the mere purpose of getting it integrated
... my question is what is the relationship between rules and techniques
... for example, is it a 1:1 mapping?

WF: agree, we also identified some of the questions
... do we want rules to replace Failures, or be stand-alone

MC: are rules failure-oriented?

WF: yes, the usually map to Failures

MC: WCAG is conformance-oriented
... so these rules would not determine conformance

WF: correct, they are to identify miss-conformance

MC: in that case, Failures are not comprehensive
... don't think it will ever be

WF: also not itending to be exhaustive but more comprehensive than currently
... much of this work right now is coming from an EC-funded project called WAI-Tools
... intending to develop 30 rules this year, 25 next, and 15 the year after

AC: historically there have been difficulties with identifying failures
... because they are then considered as absolute failures in any context
... if rules should pass the WG then the bar will be fairly high

WF: were not intending as replacement for Failures from the start
... but meanwhile accuracy seems quite high
... there are assumptions to contextualize the rules
... for example, the color contrast rule has the assumption that there is no CSS style change widget

AC: there is a Venn diagram of ACT Rules and Failures Techniques which may not be a complete fit
... not sure if that is a problem, though
... MaryJo was asking about process of publication
... as a sub-group of AGWG, then think could go into a queue for publication

MC: issue with reviews is that some may say "yes" and others may spot problems

AC: problem spotters would be better join the work

MC: problem spotters can't join all groups

<moekraft> Shadi: I want to say we have some criteria already. Rules being developed in Auto WCAG CG. Task force is focused on developing the spec. Could live beyond and become gate keeper. Certain criteria that rules need to be reviewed by 3 independent groups and two implementations

<moekraft> Shadi: We can preprocess for AWAG Spotters welcome to find bugs however if we do our work directly this should happen very rarely

WF: have a review process document
... needs updating but basically has a relatively high bar

AC: if they are coming in as pre-revewied ACT Rules than may be easier
... but if coming to replace Failures, then may be more tricky and may need more discussion

WF: would like to get it to replace Failure Techniques
... but understand that it may not be a perfect fit
... maybe revise the format for Failures so that we can fit the rules in?

SAZ: maybe broaden the question, are these rules considered useful for content authors?
... and if so, how would the group envisioning integrating them into the WCAG framework?

AC: would like to defer to AWK and MC for that, but do agree they would be helpful
... if they end up replacing some Failures then great
... but suspect may not be a 100% fit

WF: don't really like it to be half-half

MC: layers of guidance for WCAG
... think these may be a layer below the techniques
... maybe with cross-referencing
... trying to merge may be a huge project that is difficult to keep in synch

<alastairc> replace or add to?

SAZ: agree it is sort of a layer but not completely distinct

<wilco> https://auto-wcag.github.io/auto-wcag/pages/rules.html

SAZ: want to focus on integration in the WCAG framework

<wilco> https://auto-wcag.github.io/auto-wcag/rules/SC2-4-2-page-has-title.html

SAZ: this encourage development and use of the rules
... at the same time, more rules clarify the use of the guidelines

AC: would like to think about this

WF: absolutely - we have been learning over the past three years
... but what are the questions to take forward to the WG?
... I like the idea of integrating them into the Understanding at least

AC: this may be staright forward - not necessarily easy but straight forward
... would like to discuss this with AWK and MC about potential options
... like several of the ideas, like cross-linking

<wilco> https://w3c.github.io/wcag-act-rules/review-process.html

MK: still struggeling with the claim that ACT Rules lead directly to failures
... pretty strong disclaimer in WCAG

<moekraft> https://www.w3.org/WAI/standards-guidelines/wcag/faq/#techsnot

WF: there can always be a conforming alternative
... also if they are not used correctly
... they are also not exhaustive
... but rules do identify specific failures

SAZ: conforming alternative exception also apply to Failures

TPAC: Who is attending in person? Please fill out survey. https://www.w3.org/2002/09/wbs/93339/ACTAVAIL2018/

WF: please complete the survey, especially the question of TPAC

https://github.com/w3c/wcag/pull/381

https://github.com/w3c/wcag-act/issues/224

WF: like Anne's proposal

SAZ: me too. however, aggregation is currently a normative part of the spec, so it needs to be included

WF: how?

SAZ: tools that generated aggregated results, or that consume aggregated results

WF: WCAG-EM Report Tool could do that

<wilco> https://www.w3.org/TR/act-rules-format/#quality-accuracy

<wilco> https://www.w3.org/WAI/GL/task-forces/conformance-testing/wiki/Exit_Criteria_for_Rules_Format_Spec

Summary of Action Items

Summary of Resolutions

[End of minutes]