Conformance Issues and Decisions

The purpose of this page is to keep track of the little details of what we have worked out and what still needs to be addressed to have an overall conformance structure that works. Note the date last updated on the tab above so you know how current this information is.

Unorganized List

We want adjectival scoring as appropriate for each guideline. We haven't decided on the adjectives yet.
We like Rachael's proposal with passing basic tests and advanced tests
- Where do the basic tests vs. advanced tests go? The Method test tab?
- How many adjectives should there be? 5? 3? Most of the group wants 5, to be consistent with the Likert scale.
What should the declared scope be?
- Some in the group agreed that scope should be "logical subsections of the product" -- the status quo today.
- Others pointed out that modern sites have teams that often work on small sections of a dynamic page or product. If so, how do we include guidelines that apply to an entire screen or group of screens?
- It could be similar to the way that non-interference works today. Even if the cause of an accessibility problem is outside the scope, it is included in the scope. For example, declaring the scope as the crossword puzzle of a newspaper, still requires the ability to navigate to that scope. Like dependencies in software development. If you are testing a specific path through a series of pages, then the guideline wouldn't apply, even though it is on the same page, e.g. a heading H1 in the footer, shouldn't become part of the scope because it isn't in the task completion.
  - defining what tasks affect task completion. (discussed 28 April).
    - We need to give examples of what should be included and what should not.
    - Make it transparent instead of us defining what the scope is
  - what is some useful level of scope in a conformance claim - not just a button, but a collection of tasks
How to allow multiple ways through a path that are better for some disabilities rather than others. Form a subgroup with Peter Korn and John Kirkwood.

VF2F Results

Scope

We will look at ACT for atomic and composite rules
We will look at composite rules to look for holes where people could game the system

Representative Samples in Task Completion

Proposals for breaking down a task into paths
Looking at how different paths may be more appropriate for different disabilities - might work with adjectival.
Potential proposal: Define the paths and the units or widgets of code or interactions. Then define an intersection so that each path is tested, and each unit is tested, but not every path with every widget.
What should be level of granularity of reporting

Functional Needs and where they are included in WCAG3

appropriate grouping / hierarchy of user needs
how large and complex the list of functional needs should be.

How do companies track their issues over time

agreement on editorial controls not being robust over time, especially in a CMS with schools
If we use older subtests, how does that impact the score?
we don't want to include old test results in the current score
We don't want an expiration date and we can solve that with transparency.
questions on the granularity of the date tested

Options:

Date lasted tested in the claim
Date last tested decaying over time included in an overall score
The responsibility of the client to represent their claim truthfully.

Task Based vs. Traditional

If the measure of conformance is the task level, then we can do traditional testing within the task.
The investment in existing tools and how you use the output of the tools.
how (and where) to include non-interference?

Benchmarking - Jake's proposal

need examples
how do top tasks fit into
Happy flow or entire flow?
the scale of creating top tasks
creating a failure task flow

Benchmarking - Jake proposal of 2 June

Should there be separate currencies of a total? - DISCUSS