mc1098
commented
2 years ago On further research the crossref from BibTeX seems to allow for any valid entry to be a crossref target, though all the examples I find are more logically about pages in a book or use of inproceedings referencing proceedings.
This now seems less like a "view", was thinking pages in a book for this term mainly, and more of a valid entry providing a base for another entry to build off of.
Now it is more likely that this should be considered as a tree with the valid entry being the root which can only have a set of leaf entries that are only valid with all or part of the information from the root.
An important issue to resolve is how to identify an two entries that could have this hierarchy but are normalised in the current parsed format and where should this be resolved? by the Format implementation? or the ast?
Views would be a new bibliographic entry type that is based on a subsection of an existing entry, consider a page range in a volume in a series. This abstraction would be a deduplication of bibliographies that have multiple views.
Views can be represented well in formats that support something similar to
BibTeX'scrossreftag and for formats that don't support it then it should still be trivial for a view to act like a full entry.Outstanding questions
Should entries and their view be represented as a tree or flattened? (disposition: tree) Tree would allow for better management of all related entries, removing the root entry can remove all related entries easily whereas when flattened we'd have to go hunting for other related entries.
Which entries can have views and how to represent that