All Generated Schemas (etc.) Need Machine Processable Source Information #763
Description
Following a discussion on TEI-C Slack, I want to suggest that all generated outputs such as schemas and schematron take a consistent approach to recording machine processable digital provenance metadata. e.g. generated schemas and where applicable schematron should record the source of the schema itself, its name, version, and a URI at which it can be retrieved. This should not be stored in a comment and the maximal amount of derived information should be included (it is better to have it and not need it when storing provenance metadata).
@sydb points out that Relax NG can embed elements from other namespaces and so we could do something like this:
<tei:schemaRef type="RNG" key="tei_lite" n="4.9.0" url="https://tei-c.org/release/xml/tei/custom/schema/relaxng/tei_lite.rng"/>
<tei:schemaRef type="ODD" key="tei_lite" n="4.9.0" url="https://tei-c.org/release/xml/tei/custom/odd/tei_lite.odd"/> Or <tei:application> might also be a solution that could be explored.
I've put this as a stylesheet issue, which is where the change needs to be implemented, though presumably the decision about how to implement and to do so consistently across as many forms of TEI-C outputs as possible is not specific to the Stylesheets itself.
To be honest, I'm kinda surprised that our schema outputs don't already feature at least some solid indication of what customisation they come from! For example view-source:https://tei-c.org/release/xml/tei/custom/schema/relaxng/tei_lite.rng says in a comment (blech) that it is generated from ODD source and when, but not what schema it is, or where the source ODD is!