758 resource identifier

[CECSpecialistI]

https://github.com/uwlib-cams/MARC2RDA/blob/main/Working%20Documents/7XX.csv

[cwarc]

https://docs.google.com/spreadsheets/d/1_E3wrUKGZaNpkPkPyylp2JFLnq8hNjzwunpGHFb2GL0/edit#gid=523701703

Hello Everyone,

Please see the first pass for field code 758 Resource Identifier (R). Could anyone take a look?

Thanks.

Best, Alice

[cspayne]

Mapping $4 and $1 seem pretty straightforward: if $4 is an RDA WEMI property, map it as such with the $1 URI as the object. $a and $0 seem trickier, because they are nomens associated with the object identified in $1. I don't see a way of determining what type of RDA entity that object is, which is how we would determine what property to use when mapping $a and $0.

[GordonDunsire]

@cspayne and colleagues: I agree that there is a straightforward transform for some of the instance data, based on the use of an RDA property.

If we assume that the property value is correct and valid, then the condition for an RDA property can be simplified to a generic function: position(subfield, 'rdaregistry.info') > 0. This tests if the IRI contains 'rdaregistry.info'. Anything more elaborate is a form of validation.

Complications :-(

1. 758 subfield $4 says: 'Code or URI that specifies the relationship from the entity described in the record to the entity referenced in the field.'

This implies that only one entity is described in 'the record'. (Unless I am misinterpreting US English :-)

Subfield $i says '... the resource described in field 758 and the resource described in the 1XX/245 of the record'.

This indicates that the 'resource' or entity is determined by 1XX/245.

In the example for subfield $a, there are two occurrences of the field, one for a manifestation and one for a work. This appears to contradict the implication that only one entity is being described.

The complication for the transform is:

Which entity is the subject of the 'record'/field 758/transformed statement:

{thisEntity} {rda relationship property} {relatedEntity} .

Is {thisEntity} 'the' instance of manifestation being described, or 'the' instance of work that is embodied by the manifestation? This is further complicated when whole-part manifestations and aggregates (multiple embodied works) are being described.

2. The domain of the RDA property recorded in subfield 4 is the class of entity that is the subject of the statement. The range of the property is the class of the related entity.

This is a strong declaration. If we assume that the data is valid, I think that this should have a high priority if there is a collision with the transform of other fields. That is, if there is doubt about the classes of instances that are generated by the transform overall, field 758 should be a deciding factor. I am thinking particularly of expressions rather the the work/manifestation associations implied by the manual.

3. Subfield $0 can be mapped to an RDA identifier recording method for the property that is used. So the same straightforward transform can include either or both of subfields $0 and $1.

If the RDA property is a canonical property, it can be transformed to a datatype property (for subfield $0) or an object property (for subfield $1) using a simple process:

If exists($0) replace($4, '/P', '/datatype/P') If exists($1) replace($4, '/P', '/object/P')

[The Registry and namespaces were designed for this purpose.]

4. If an RDA property is not used, then we may be able to develop an approach based on the work with relationship properties in other 7XX fields. That is, use a mapping from the other property to an RDA property.

Uncomplications :-)

Subfield $a seems straightforward. We know that it is a label for the related entity, but we don't know what kind and there is no way of automatically determining if a string is an authorized access point or title, etc. So the transform should use a general RDA appellation property for the entity of subfield $1:

{$1} {RDA appellation of entity} '$a' .

In the MARC manual example:

http://www.wikidata.org/entity/Q1613936 rdamd:P30277 'Bored of the Rings' . // property determined by the range of $4 http://id.loc.gov/authorities/names/no97079452 rdawd:P10329 'Tolkien, J. R. R. (John Ronald Reuel), 1892-1973. Lord of the rings' . // property determined by the range of $4

Is this in scope? Should we add RDA properties to existing instances, or leave them encapsulated for de-referencing? Similar (non-RDA) properties are attached to the existing instances ('label' in wikidata, the aap in LC authorities).

It is clearly in scope for instances that are minted in the transform, but such minting is triggered by other fields so is not directly relevant to this field.

[lake44me]

At the BIBFRAME Update Forum Jan. 22, Jeff Mixter of OCLC said they would start adding WorldCat Entity URIs to Worldcat records at the end of January (so presumably they are going in now) for Works. I asked in the Q/A what the WorldCat definition for "Work" was, and was told it aligned with BIBFRAME Work.

Since various people at LC have stated that a BIBFRAME Work is equivalent to RDA Expression with some RDA Work characteristics included (I'd have to hunt down references to this in print...), would this support taking the position that the entity being related most closely corresponds to an RDA Expression, and that what it is related to is the RDA Manifestation whose description is the focus of the record?

Jeff said they would use a 758 field but didn't show an example. While we can't assume everyone using 758 will follow OCLC's pattern in that use, as soon as this is documented, we should analyze it and the property asserted.

@GordonDunsire hate to say it but your use of the word instance befuddles me because I'm entangled with BIBFRAME now for another project and bf:instance is a thing. Hard to find words that haven't been used elsewhere with a different meaning these days!

[GordonDunsire]

@lake44me: I use the word 'instance' in its RDF sense, which is based on the vernacular use as 'an example or single occurrence of something': a single occurrence of a thing (i.e. of a class). RDF Schema (RDFS), since 2004, uses the term frequently and consistently: 'The members of a class are known as instances of the class'. I think this usage predates that of BIBFRAME. BF could easily have used the term "manifestation", which is used elsewhere with the same meaning, but chose to use a befuddling term instead. I don't speak BF at all, but I know a little RDF/LRM and I think that's more appropriate for an RDA project :-)

[cspayne]

@CECSpecialistI Could we see if someone is willing to take this mapping from me and complete it? I'm afraid I won't have time to.