Accessibility Conformance Testing Teleconference

<maryjom> https://w3c.github.io/wcag-act/act-rules-format.html

MaryJo: We did an ordering change and generated a new spec

https://w3c.github.io/wcag-act/act-rules-format.html#structure-outline

MaryJo: Accessibility Support has been moved up to Section 4.
... Any objections to this placement?

Pull request 143: Add note on unique identifiers (Issue 133) Define unique identifier

<maryjom> https://github.com/w3c/wcag-act/pull/143/files

<maryjom> https://github.com/w3c/wcag-act/issues/133

MaryJo: No objections to reorder. Change accepted. Issue 137: Spec sections ordering change can be closed.
... Let's move on to Issue 143. Add note on unique identifiers

Proposal from Wilco: Note: The unique identifier can be any type of text or number, such as a URI or a number in the database.

MaryJo: Romain's suggested rewording: The unique identifier can be any text string, such as for instance a URL or a database identifier.

Does it need to be universally unique or just in the scope of the ruleset?

Alistair: Presumably it's unique within the set. If we are talking about tools, we would want to have a unique rule, e.g. WAI Tool Rules
... You need the unique identifier for that set.
... Scope down until it is useful

?: Want to be able to compare results across rulesets. If separate ruleset has the same rule and we want to compare.

Alistair: We would have the rule id as meta in our own rule set. This is rule A and meant to be testing Rule X from testing rules toolset. Same as testing best practice X, e.g. W3C
... As long as they are unique within the rule set, we should be okay

MaryJo: Do we need any update to Romain's text?

Kasper: We should mention the scope

MaryJo: how do you propose we write it up?

Alistair: Two things, WAI Tool Rules, W3C pillar of rules. Within WAI Tools rules you would need unique id for each one.
... However, in organization, have unique id but have a referrer to X. If rule is testing another rule referred to, have URL to other test.

MaryJo: Do we need another optional field?

Are we mapping to SC?

Alistair: No just test level

Anne: Don't completely understand discussion. I think we discussing how to map your own and not the ACT rule. Shouldn't everything in X ruleset have a unique id?

<anne_thyme> https://w3c.github.io/wcag-act-rules/

Anne: X-r1 would this be a unique id?

Alistair: Each rule set has its own set of rules and each rule should have a unique id. What happens if you bring in another test suite and there is a mismatch.

Anne: If we have rules 1, 2, 3 and bring them into a rule set, we could have a clash

Alistair: Everything needs its own unique id within its own cluster

Anne: Use ACT rules format internally?

Alistair: One of the things that will be flushed out in reality
... Would be quite useful internally that we have rule X internally but it maps to WAI Tool Rule Y. And we want to conform to rules format. How do we map?

Anne: Isn't that based on implementation. We can add our own fields as long as it doesn't break the rules format.

MaryJo: Do we need to make changes?

<maryjom> Romain's suggested rewording: The unique identifier can be any text string, such as for instance a URL or a database identifier.

Alistair: Remove, for instance...

<maryjom> The unique identifier can be any text string, such as a URL or a database identifier.

Alistair: What's the difference between URL and URI

Romain: The most modern interpretation is WhatWG which deprecates URI

MaryJo: Everyone agree to update?

What about scope or cross reference?

Alistair: I would drop optional bit.

MaryJo: Can this stand as is?

+1

<anne_thyme> +1

<maryjom> +1

<rdeltour> +1

<Kasper> +1

Alistair: Uniqueid is what you put inside the test. We've never used URL's before for a unique id. We want to keep ids quite short. Since they are quite often used.
... I recommend using plain text

What about, The unique identifier can be any text string, such as a plain text, URL or a database identifier.

MaryJo: I think adding plain text helps

Are we using uniqueness by using the term any?

<agarrison> The identifier can be a unique

Alistair: The identifier can be any unique text string, such as plain text, URL or a database identifier.

<maryjom> +1

https://w3c.github.io/wcag-act/act-rules-format.html#structure-outline

This changes in context since the rule outline has unique identifier as an item

Romain: I think we are overthinking this one. It's just a note of what we mean by unique identifier.
... This is just informative.

<agarrison> The identifier can be any unique text value, such as plain text, URL or a database identifier.

<Michael_GK_Rasmussen> +1

<maryjom> +1

<Kasper> +1

<anne_thyme> +1

Here's an existing definition from the internet of things to compare: "A unique identifier (UID) is a numeric or alphanumeric string that is associated with a single entity within a given system. UIDs make it possible to address that entity, so that it can be accessed and interacted with."

+1

Reword informative prose in the Introduction and Scope sections - is this ready to merge?

<maryjom> https://github.com/w3c/wcag-act/pull/131/files?diff=split

MaryJo: Typo in last spelling of accessibility

Alistair: English doesn't flow well

<maryjom> Editorial change: It is intended to describe both manual....

This format is intended to enable a consistent interpretation of how to test for accessibility requirements so as to avoid conflicting results of accessibility tests. It is intended to be applicable to describe both manual accessibility tests as well as automated testing done through accessibility test tools (ATTs).

<maryjom> Editorial change: Fix spelling of "Accessibility"

This format is intended to enable a consistent interpretation of how to test for accessibility requirements so as to avoid conflicting results of accessibility tests. It is intended to describe both manual accessibility tests as well as automated testing done through accessibility test tools (ATTs).

Suggestions on 8.4 Test Subject - add state?

<maryjom> https://github.com/w3c/wcag-act/issues/128

Alistair: Web components and shadow DOM make this more complicated. You can describe it under test subject.

MaryJo: Maybe need to be more explicit. If there are states...

Alistair: But then you have to describe the states. 8.4 covers what the comment about.

MaryJo: Does SiteImprove agree?

Kasper: Big difference between single url and paragraph of description

Alistair: ... Yes. What is ATT?

MaryJo: Accessibility Test Tool

Kasper: Examples use URL as subject. Can we have an example that does not use URL? Make it clear that it is an open field.

MaryJo: It could be done so we have an example that shows that.

Alistair: I would take out "more complex action" Just use "action"
... I don't want folks think that putting in a single url is sufficient. We need to think of multiple tests across a page. This is not a complex action.
... Also, it can be that the initial page is very different from a page that is loaded seconds later. Action can be by the user or computer.

MaryJo: I will make note of what we suggest in issue. Once W3C is available, I will update the Availability Survey and send out,