Organize the presentation of the CDDL into skeleton and standard extension tables

[deeglaze]

This issue is getting at the heart of my call for simplification to really whittle down the presentation to separate form from function. It's very easy to get lost in the details otherwise.

This will separate out minimum necessary to express reference values from minimum for shared value of the authors of this spec. Endorsements are an extension that are not necessary to verify evidence from an attester, specifically because they are defined as a claim without evidence.

We've discussed but have not implemented the convention that every $$ extension should have its own IANA table to assign meaning to extensible maps. We should specify all extensible maps just as their repeated extension points and provide descriptions of the initial table definitions in a later section.

Extensible type choices are left indeterminate or as a very opinionated minimum selection since they could be parsed as anything. We can provide in a later section some common recommended choices that profile writers could be reasonably confident that their profile could be easily adopted by verifier implementations.