For each type of record, we should have a scheme similar to record metadata for scoring the provenance connections. i.e.: does a key finding cite something? Does a table have an array?
Probably the best way is for every record to categorize connections that are necessary to count as good (e.g. a chapter having references), and categorize those connections that may not be necessary but are good for provenance and boost a score (e.g. an array being derived from a dataset).
Amrutha should take the lead, but let's talk frequently so we stay on the same page.
Following Issue #538
For each type of record, we should have a scheme similar to record metadata for scoring the provenance connections. i.e.: does a key finding cite something? Does a table have an array?
Probably the best way is for every record to categorize connections that are necessary to count as good (e.g. a chapter having references), and categorize those connections that may not be necessary but are good for provenance and boost a score (e.g. an array being derived from a dataset).
Amrutha should take the lead, but let's talk frequently so we stay on the same page.