Re: aim of being unambigious/objective (was Re: ACT-R Meeting, tomorrow (November 28))

Hey Patrick,

Bummer you didn't make it. Yes, we alternate meeting times. The second
Thursday of every month is an EU afternoon meeting, the fourth Thursday of
the month is an EU morning meeting.

With regard to unambiguous / objective, this is something required by the
ACT Rules Format. They are intended to mean two different things:

- Unambiguous: It should only mean one thing. For example, if you say
something like element A is a child of element B, that can mean a child in
the DOM tree, the flat tree, or the accessibility tree. That's ambiguous.
Another way something can be ambiguous is by not providing proper
punctuations. "Element A or Element B if it has attribute Y" for example,
does the "if it has attribute Y" apply to element A?. That kind of stuff
isn't allowed in the applicability or in the expectation.

- Objective: This is about things that are not up for discussion.
Subjective things tend to be qualitative in nature. Something like "does A
describe B" has a qualitative aspect to it. How good is good enough?.
Another thing that makes things subjective is when you are looking at
intent. What is the purpose of a thing? How does a user perceive a thing?

In ACT, both the applicability and the expectations have to be unambiguous.
They shouldn't have multiple meanings. Only the applicability has to be
objective though. If you have something like "does A describe B". We can
discuss whether or not we think "A" describes "B", but we shouldn't have to
argue over what "A" and "B" are, and how many there are on a given page.

Generally, the biggest differences in test results come from disagreements
on what various terms in WCAG mean, and what they apply to. For example
does 2.4.6 headings and labels apply to rowheaders in a table? By insisting
that the applicability is objective, ACT rules ensure such questions get an
answer.

Hope that helps! Kasper and myself are working on a document to provide
better guidance on how to interpret these terms. There's also the problem
that something is never perfectly objective, or perfectly unambiguous.
We're not writing mathematical proofs here, so how do we draw that line?
There's definitely more discussion to be had on the subject. Happy to pull
you in if you're interested.

W

On Thu, Nov 28, 2019 at 11:33 AM Patrick H. Lauke <redux@splintered.co.uk>
wrote:

> On 28/11/2019 10:11, Wilco Fiers wrote:
> > Hey folks,
> > Here are the minutes from today's meeting:
> > https://www.w3.org/2019/11/28-act-r-minutes.html
>
> Oh, I had best intentions of joining the call for the first time today,
> but missed that it was in the morning rather than in the afternoon (UK
> time).
>
> One aspect that I wanted to maybe very broadly touch on is the idea of
> unambiguous applicability and expectations ... because as much as WCAG
> tried to be objective, there are many instances where it really comes
> down to subjectivity (to determine if a text alternative conveys the
> same purpose ... who can objectively say what the purpose is in all
> situations? or headings/labels being descriptive ... that's also
> subjective to an extent). While I think it's a laudable goal to always
> be unambiguous (assuming it's meant as "objective"), I'd question if
> that's always possible.
>
> Thoughts?
>
> P
> (also, is it best to raise this as an issue in GitHub, or is it ok to
> discuss things on mailing list as well)
> --
> Patrick H. Lauke
>
> www.splintered.co.uk | https://github.com/patrickhlauke
> http://flickr.com/photos/redux/ | http://redux.deviantart.com
> twitter: @patrick_h_lauke | skype: patrick_h_lauke
>
>

-- 
*Wilco Fiers*
Axe for Web product owner - Co-facilitator WCAG-ACT - Chair ACT-R