Suggestions for changes to section 3.8 of the DwC RDF Guide

[Paul J. Morris]

Suggestions:

(1) Changed language of 3.8 from "cannot" to "does not recommend",
something on the order of:

3.8 Darwin Core terms for which this guide does not recommend the use
of a dwciri: object property

(2) Retain the following terms in 3.8

dwc:relationshipOfResource - Out of scope, asserts class
relationship. dwc:relationshipAccordingTo - Out of scope, asserts
class relationship. dwc:previousIdentifications - Out of scope,
asserts class relationship.

dwc:dynamicProperties - Expected to contain JSON as a literal.

dwc:associatedOccurrences - Use dcterms:relation and rdf:type.
dwc:associatedMedia - Use dcterms:relation and rdf:type.
dwc:associatedSequences - Use dcterms:relation and rdf:type.
dwc:associatedTaxa - Use dcterms:relation and rdf:type.
dwc:associatedOrganisms - Use dcterms:relation and rdf:type.

(3) move the following terms:

dwc:measurementValue - move to 3.4, expected to be used with literal.
dwc:measurementAccuracy - move to 3.4, expected to be used with literal.
dwc:measurementDeterminedDate - move to 3.4, expected to be used with
literal. dwc:measurementRemarks - move to 3.4, expected to be used
with literal dwc:otherCatalogNumbers - move to 3.4, expected
to be used with literal, add note about owl:same as in comment there.

dwc:measurementUnit - move to 3.7, has dwciri analog
dwc:measurementType - move to 3.7, has dwciri analog, use controlled
vocabulary if available. dwc:measurementDeterminedBy - move to 3.7,
has dwciri analog, group with dwc:recordedBy dwc:measurementMethod -
move to 3.7, has dwciri analog

Looking at the text of the guide, I suspect that the rationale for
placement of the dwc:measurement___ terms comes from the statement in
2.3.1.5 "Any Darwin Core class IRI may be used as a value for rdf:type,
although it is questionable whether dwc:MeasurementOrFact and
dwc:ResourceRelationship instances make sense in the context of RDF. "
where a question is raised about the natural parent MeasurementOrFact.
Given "Any DarwinCore class IRI may be used as a value for rdf:type",
it feels like the guide should either take a position that explicitly
excludes MeasurementOrFact, or should (as I'm suggesting above), treat
the measurement___ terms in paralell with other dwc terms.


-Paul
--
Paul J. Morris
Biodiversity Informatics Manager
Harvard University Herbaria/Museum of Comparative Zoölogy
mo...@morris.net AA3SD PGP public key available