Additional cleanup for dynamic references

[jdesrosiers]

What kind of change does this PR introduce?

Clarification

Issue & Discussion References

• Closes Proposal: Remove initial resolution step for dynamic references #1140

Summary

This is mostly cleanup building on #1555, although I rewrote a couple of parts I though needed it.

The main objective is to achieve consistent and correct usage of terms around fragment identifiers. Here are the definitions I'm using.

• fragment identifier -- This is an identifier for a secondary resource (subschema) within the primary resource (schema resource).
• IRI fragment -- The fragment identifier that appears after the # character in an IRI.
• plain name -- A specific type of fragment identifier. $anchor creates a plain name fragment identifier.
• fragment-only IRI -- An IRI that starts with # and is followed by a fragment identifier.
• fragment -- This term by itself is ambiguous and imprecise. I replaced all usages with a more precise term from the above terms.

Here's a list of the other notable changes.

• Defining location-independent identifiers
  • I removed the language around constructing an IRI from the fragment identifier. All this section needs to do is define how plain name fragment identifiers are created. How to use a fragment identifier in any IRI is well established in RFC-3986/7 and doesn't need to be repeated here.
  • I tried to clarify that $dynamicAnchor does not create a fragment identifier. It's a special kind of identifier unique to the dynamic reference mechanism. Specifically, you can't use the identifier created by $dynamicAnchor in an IRI.
• Dynamic References with $dynamicRef
  • I provided a new introduction for dynamic references that I think make sense. I've always thought that description was awful and I think this an improvement.
• Example of recursive schema extension
  • I fixed a sentence that didn't make sense.

Everything else is minor grammar fixes or whitespace cleanup.

Does this PR introduce a breaking change?

No. (Actually, the issue this closes is a breaking change, but this PR doesn't break anything. This is just cleanup related to and around that change.)