Open CECSpecialistI opened 2 years ago
Gordon has previously said: "there can only be one skos:prefLabel per language in a value vocabulary." We have already determined that we can not identify the language using $6. II.G.5.d of the Decisions index says: "Script and language of strings cannot be reliably determined from the MARC format in $6, and so are not mapped." So I do not think that multipel skos:prefLabels are an option.
@cspayne, @CECSpecialistI: The SKOS integrity condition is 'S14', stated here: https://www.w3.org/TR/skos-reference/#L1567
'A resource has no more than one value of skos:prefLabel per language tag.'
I think a script can be encoded in the xml language tag with the language, so in theory this should allow both the original and transliteration to be sources of prefLabel. However, if the script and language cannot be determined, the best options are:
Is the assumption correct?
@cspayne, @CECSpecialistI: The skos property skos:related is similar to rdand:P80113 (is equivalent to), but I don't think it applies to main/880 concepts. Basically, there is only one instance of skos:Concept, with the 'same' label in different scripts.
This issue looks like the Preferred vs. Alt Labels for concepts for 880 fields topic from our meeting today. It isn't assigned to anyone, so I'm going to assign to @cspayne - please reassign if I got that wrong and this should go to someone else!
We have decided how to treat 880s in all of the current contexts that we have encountered, so I have moved this to ready for transform.
It is complex code and will not be done until all other fields that have $6 subfields are coded
@GordonDunsire @CECSpecialistI How we are treating linked fields when we are mapping the 'main' field to a skos:concept? With nomens we are using rdand:P80113 (is equivalent to) "literal value of 880", is there something similar we can do for a concept? Do we mint a separate concept for the 880?