Editorial: clean up small parts of sequential focus navigation

[domfarolino]

The sequential focus navigation infrastructure could likely use a revamping, but before this is accomplished, this PR attempts to clean up some low-hanging fruit by:

• Getting rid of confusing concepts like "home document" and "home sequential navigation focus order" which are definitions that appear mid-algorithm, and are likely not necessary in general
• Clean up implicit flag-like variables by making them more explicit enums
• Clarifying algorithm parameter types and return types
• Clean up wrapping
• Specifying that the "DOM" selection mechanism entails tree order
• Clean up recursive invocations of the sequential navigation search algorithm

---

/interaction.html ( diff )

[domfarolino]

Please see my question in the WHATWG matrix chat about this. I'm concerned that the old spec (and thus the spec after this PR) have no material difference between "DOM" and "sequential" selection mode. The note below the home sequential focus navigation order dfn says the only difference is that in the "DOM" selection mode, <var>starting point</var> is not in its own Document's sequential focus navigation order. But that shouldn't actually impact the assignment of <var>candidate</var>, since candidate will always be selected from the starting point's sequential focus navigation order anyways.

So I'm not sure what to do here. Either I am right, and there is no observable difference between the two mechanisms in this case, or it is the intention of the spec to have a difference between the two. I.e., maybe:

• In the "DOM" case, <var>candidate</var> should be pulled from the DOM in tree order relative to starting point (not necessarily coming from the starting point's Document's "sequential focus navigation order"
• In the "sequential" case, <var>candidate</var> should be pulled from the subset of the DOM that is included in the starting point's Document's "sequential focus navigation order"