[Crm-sig] Qualifying predicates and objects (advice/criticism needed)

Dan Matei Dan at cimec.ro
Sun Jan 20 13:28:25 EET 2013


Still working on converting our ancient (i.e. 20th century) databases to something more current, I still have problems :-(

My records for the statements with literal objects have (of course) the fields:

•   subject id
•   predicate id
•   object (string/number/date/XMLLiteral)

Besides I have:

•   syntaxEncodingScheme (for the string/XMLLiteral-statements)
•   P91-has_unit (for the number-statements, in case they are dimensions :-)
•   P72-has_language (for the string-statements)

Moreover, in order to keep under control the proliferation of subproperties, I added:

•   predicate type id (of the ".1 P2" kind)
•   object type id (also of ".1 P2" kind)

In this image


I show (as an example) 3 alternative ways of avoiding the specialisation of P78-is_identified_by. 
I find merits for each but I'm in doubt.

What do you think is preferable ? Should I judge on a case-by-case basis ? Or something different would be better ? 

Is "object type id" useless ? What if I want to say that a string-object is a transliteration ?



Dan Matei
director, Direcția Cercetare, Evidență a Patrimoniului Cultural Mobil, Imaterial și Digital [Movable, Intangible and 
Digital Heritage Department] (aka CIMEC)
Institutul Național al Patrimoniului [National Heritage Institute], București [Bucharest, Romania]
tel. (+4)021 317 90 72; fax (+4)021 317 90 64, www.cimec.ro

More information about the Crm-sig mailing list