Closed wildit closed 4 years ago
Hi Tobias,
Just a few quick words for now. The better (I would even say the only) relation to be used is RiC-R024 includes (in RiC-O : https://www.ica.org/standards/RiC/ontology#includes), with domain Record Set and range Record or Record Set. It is defined as "Connects a Record Set to a Record or Record Set it aggregates or contains", which is exactly what an ISAD(G)/EAD hierarchy means. It is not a subproperty (in RiC-O) or narrower relation (in RiC-CM) of RiC-R002 has Part, since the definition and meaning are different. A Record Set may at some time lose (or aggregate) one or more included Record or Record Set, and still be defined the same. For instance, an open fonds may include, at some date of its history, a new series, and remain the same entity (defined as the records created or accumulated by some agent), or some records may be deleted within this fonds at some date, and the fonds will keep its nature and identity.
What do you think?
I understand the arguments on why to use "includes" / "includedIn". The problem is that for R024 'includes' only the range "RecordSet" is allowed, that means we can not nest records. This is a absolutely necessary functionality if we have eg. a "dossier" with different "items" in it. This is a nested/structured record and not a record set. Could we allow for R024 the range "RecordSet" and "Record"?
I have checked this, and RiC-O already defines rico:includes as having range a union of Record and Record Set classes, which means that you can use instances of rico:Record or rico:RecordSet as its object :
We want to model the tree of ISAD(G) in RiC-O. There seem basically two different ways to do that, RiC-R002 'hasPart' and RiC-R024 'includes'. What's the difference between these relations? What's the better relation to use when mapping ISAD(G)-hierarchies? And why?