Guide be made explicitly pointing out ShadyDOM's deviating from how Shadow DOM handles the traversal of aria-annotated children & the children's behavior; along with how to leverage WCT to test & verify its behavior of such things compared to native Shadow DOM

[lozandier]

The Shadow DOM spec has traditionally outlined that user agents with assistive technology traverse & evaluate the document tree as rendered, but @cdata has verified that isn't the case w/ ShadyDOM

Accordingly, there seems to be value of a guide added to the guides section already outlined as an ongoing thing to have a guide that provides & claries the following things that would be especially useful in BDD environments that would make extensive use of Web Component Tester [WCT]:

• [ ] Show specs that show what would be normal behavior w/ the shadow global Polymer setting vs. the shadyDOM (default) settings of Polymer regarding this.
• [ ] Such specs & accompanying copy would simultaneously implicitly debunk a common misconception that Web components don't play nicely with assistive technologies.
• [ ] Demonstrate pragmatic organization of the specs of a Web component that accounts for its usage within an environment that uses either shady (default) or shadow be set as a global setting (with something likecontext, nested describe blocks, etc). Some often don't discover this, & this would reinforce its importance of a comprehensive test suite for components.

[ghost]

To help me prioritize and manage issues in the repo, I'll close this issue with no docs changes in 7 days. To prevent this, please comment within 7 days.

[arthurevans]

@katejeffreys do you have a11y docs on your plate? I think this would be relevant there.

[arthurevans]

This would make sense to cover in a future a11y doc, but would be hard to introduce elsewhere.