Add BP on annotating non-git "subdatasets"

[mih]

There are two main challenges:

1. Where to put information on such a "subdataset", and in what format?
2. How to link such a "subdataset" with a DataLad dataset's tabby metadata record, such that it can be discovered by the catalog translator/builder?

My current thinking is this:

• we can use a tabby record for this without foreseeable issues
• linkage has to be made via an @id reference in the DataLad dataset's tabby record
• the simpliest way is to have a many table that lists all and enough identifying properties of such a "subdataset" that a corresponding override can turn it into a properly unique IRI
• the same override setup (or @id is used within the "subdataset" tabby record to establish the linkage
• the many table should also identify, where in the path namespace of the containing dataset such a "subdataset" shalle be represented/mounted

Example that would work with the setup in https://rdm.sfb1451.de/data-catalog-submission/

A "subdataset" concept @id is build from the crc-project @id plus /datasets/<name>, or the CRC project superdataset's UUID and a "mountpoint".

A "subdataset" version @id is built from the concept-id, plus a version label.

[mih]

#101 has an convention draft on this.

[mih]

This issue is now tracked as one aspect of dataset descriptions with tabby in #102