Datamap / fragment selectors: Allow any Additional selectors other than w3 fragment selectors?

[kMutagene]

After the recent datasteward circle discussion, the issue of pointing inside a fasta file came up. As fasta is a plaintext file, the proposed solution was to use plaintext fragment selectors.

However, this ignores the additional formatting information provided by the fact that a file is a fasta file. There are many biological plain text formats. For fasta, it might still be feasible to select e.g. lines, but imagine trying to do that for types provided by something like pdb or genbank.

Maybe we can discuss if and how we might build terminology/selectors on top of the w3 selectors? Something like record=2 is way more expressive and easier to create than the line range of a fasta record. Also, to annotate via plaintext selectors, one has to understand the format specs which might not be something anybody wants ;). however, i see the potential problem with maintining things like this.

cc @Brilator @HLWeil @muehlhaus