Agree needed varieties of metadata schema

[mvahowe]

In part related to #211 and #206

From the processing model:

Template

A template is not strictly an entry, since it does not have an entry id. Templates, by their nature, are incomplete. However, whatever structure is present should be consistent with a metadata-valid document. The template schema also expects additional fields to identify the template, such as an id, name and description. (These should be stored differently to entry ids etc to avoid confusion when the template is instantiated.)

Catalog Valid

This is the lowest level of validity, which is required for all entries. Fields that are used for presenting summary information about the entry must be present and valid.

Archivist Valid

This level of validity includes catalog validity. All fields that can reasonably be human-maintained must be present and valid. (This notably excludes ingredients and publication structure.)

Metadata Valid

This level of validity includes detail validity. The entire metadata document must be valid.

Burrito Valid

This level of validity includes metadata validity. The entire metadata document must be valid and, in addition, the metadata and ingredients must conform to any conventions listed in the metadata. (This level of validation is only viable for owned entries, or for the first revision of a delegated entry, since, in other cases, the ingredients may not be present locally.)

[mvahowe]

I think we can get back to this once we have recipes in derived variants. The next big thing in the schema is recipeSpecs, the first part of which is effectively embedding metadata templates in other metadata documents.

[mvahowe]

I'm going to get back to this very soon, having worked out how to create families of schema without massive duplication. The place to start is templates, since we need them as part of recipeSpecs.

[mvahowe]

I just added a very lax template option to the schema which is implemented in PR #177

[jag3773]

Need a clear set of use cases before trying to implement this.

[jag3773]

This came up again in our discussion related to #96 and #223 .