Open menzowindhouwer opened 2 years ago
Probably as part of the general FAIR Vocabularies plan in the SDR v2?
RELcat paper: Depending on the future size of ISOcat and RELcat the use of (almost) same as relationships might require special handling. These (loose) equivalence relationships might possibly lead to a combinatorial explosion. Some commercial triple stores already offer specific support to handle these kinds of large graphs:
In RELcat the same kind of approach could be followed: 1) either select one representative from the same-as clique or 2) give the same-as clique its own identifier. A possible complication there is the dynamic combination of relationship sets, which might require that these cliques need to be computed on the fly.
The references for Oracle and OWLIM needed to be refreshed: 1) Oracle: https://docs.oracle.com/cd/E11882_01/appdev.112/e25609.pdf section 2.2.8 2) OWLIM, now GraphDB: https://graphdb.ontotext.com/documentation/standard/sameas-optimisation.html
RR is an older idea (1, 2, 3, 4) that needs a fresh and actualized writeup