Database Reference
In-Depth Information
5.7 POPULAR RDFS VOCABULARIES
11
In the topographic RDFS ontology example, we have coined (or, to use the Linked
Data term, we have “minted”) our own URI for the “
has _ name
” property, which
we can almost certainly expect to have been defined by someone else previously. It is
best practice to reuse common properties like “
has _ name
” rather author our own,
provided that we agree with the original author's understanding of that property. It is
less work and means that our data integrates better with other data published on the
Linked Data Web. In Chapter 8, we talk about how to go about finding out whether
someone else has already authored a URI for a property you want to use and how
to double-check that they are using it to mean exactly the same as you. For now, we
review some basic RDFS vocabularies that contain terms like “name” that you may
well wish to reuse.
Dublin Core
is an International Organization for Standardization (ISO) standard
(ISO 15836:2009) set of fifteen metadata elements such as title, creator, date, and
rights. The RDF version of Dublin Core is at
http://purl.org/dc/elements/1.1/
.
One
term of particular interest is “coverage,” which refers to the “spatial or temporal topic
of the resource, the spatial applicability of the resource, or the jurisdiction under
which the resource is relevant.” This can be useful in specifying the context in which
your ontology can be used.
SKOS
(Simple Knowledge Organization System) (Miles and Bechhofer, 2009)
was designed to provide a simple way to port existing knowledge representation sys-
tems (primarily thesauri and taxonomies), like those used in libraries, social tagging,
or other simple classification systems, to the Semantic Web. It is a W3C specification
and includes properties such as
skos:broader
and
skos:narrower
to describe
the relationship between one genre and its more specific species or between a whole
and its parts. Whereas RDF would distinguish between hierarchical (subclass of)
and mereological (part of) relationships, the SKOS terms are much less precise, and
hence SKOS may be more suitable for use where the exact relationship is not known.
Similarly, where RDF might coin a term “participates in” to indicate the relation-
ship between an event and a group of entities that typically participate in it, SKOS
uses the vaguer
skos:related
property, which can be used to indicate any sort
of association between two concepts. SKOS also offers some interesting properties
to be used with metadata. For example,
skos:scopeNote
indicates the limits of
where the concept can be used;
skos:definition
is used in the same way as
rdfs:comment
(to store a human-readable text description of the concept); and
skos:example
gives an example of how the concept should be used.
To map between different ontologies, SKOS uses
skos:exactMatch
and
skos:closeMatch
.
skos:exactMatch
is used to link classes that are equiv-
alent in meaning, for example, Car and Automobile, but unlike
owl:sameAs
,
which we discuss in Chapter 8, it is much vaguer and does not imply anything
about the two concepts sharing all their statements, as is required by the logic of
OWL.
skos:closeMatch
is even vaguer and just indicates that the two con-
cepts are similar in meaning.
skos:broadMatch
,
skos:narrowMatch
, and
skos:relatedMatch
are also available. A new ontology can reuse existing
concepts using the
skos:inScheme
prop er t y.
