Database Reference
In-Depth Information
15. <void:uriLookupEndpoint rdf:resource = “ http://lookup.mereamaps.
gov.me/”/ >
16. <void:dataDump rdf:resource = “ http://mereamaps.gov.me/dump.rdf”/ >
17. <void:vocabulary rdf:resource = “ http://purl.org/dc/terms/”/ >
18. <void:vocabulary rdf:resource = “ http://mereamaps.gov.me/
admingeo”/ >
19. <void:exampleResource rdf:resource = “ http://data.mereamaps.
gov.me/000123”/ >
20. </void:Dataset>
21. </rdf:RDF>
There are four types of metadata specified by VoID: general metadata using the
Dublin Core vocabulary; metadata describing how to access the data; structural
metadata, which can be used for query and integration; and finally metadata denoting
links between the dataset in question and third-party ones on the Linked Data Web.
The general metadata includes information about any home page associated with the
dataset (line 7), contact e-mail address of the author or publisher can be supplied, and
Dublin Core metadata tags can be used to supply information such as licensing or sub-
ject of the dataset (lines 9 to 13). The void:feature can be used to specify which
RDF serialization format is used, for example, RDF/XML or Turtle. Access metadata
is used to describe methods of accessing the RDF triples in the dataset (note that the
dataset itself is not part of the VoID description). For example, the access metadata
can include how RDF data can be accessed via a SPARQL endpoint (line 14), a URI
lookup (line 15), as a dump of RDF data (line 16), or if the data publisher offers a
search facility into its data, this can also be described using a standard OpenSearch 28
description document via the term void:openSearchDescription .
The structural metadata describes the internal structure of the dataset and
provides additional information about the schema, which can be useful to
explore or query the dataset. For example, the vocabularies used by the dataset
(using void:vocabulary as in lines 17 and 18), the size of the dataset (using
void:entities , void:triples , void:classes , or void:properties to
express the number of entities, triples, classes, or properties, respectively), or exam-
ples of resources that are in the dataset. The void term void:uriRegexPattern
specifies the regular expression that matches the URIs of the resources in the data-
set. Another description that can be used is void:uriSpace, which specifies the
namespace to which all the resources belong. The structural metadata can in addi-
tion include information about how the dataset can be partitioned into subsets using
the void:subset property. Each part of the dataset is itself a void:Dataset ,
which can be independently annotated with any VoID annotation. A dataset may
need to be partitioned if different parts originate from different sources, the parts
have different publication dates, the parts can only be accessed via different methods
or at different locations (e.g., via different SPARQL endpoints or different RDF
dumps), or perhaps the most obvious reason for dividing a dataset, the parts are con-
cerned with different topics (i.e., they have different dcterms:subject values).
Alternatively, dataset segmentation can be carried out on a class or property basis.
Thus, one class-based partition (using void:classPartition ) would include
all the data that describes instances of a particular class, while a property-based
Search WWH ::




Custom Search