Database Reference
In-Depth Information
for publishing descriptions of a Linked Data set: Semantic Sitemaps (Cyganiak,
Delbru, and Tummarello, 2007) and the Vocabulary of Interlinked Datasets (VoID)
(Alexander et al., 2011).
7.6.1 S eMantIc S IteMapS
The Semantic Sitemaps protocol provides information to Web crawlers about the
pages available in a Web site. An XML document sitemap.xml is stored in the Web
site's root directory and encodes information about the URL, the site's location,
when it was last modified, and how often the site changes to help a search engine
optimize how frequently it needs to crawl the site. A Semantic Sitemap is an exten-
sion to the standard sitemap idea, adding information about the dataset such as a
URI and label for the whole dataset, as well as pointers to its SPARQL endpoint.
It also can include details of where the whole dataset can be downloaded from or
information about the shape of the RDF graph. The semantic search engine can go
to the dataset's URI to find an additional RDF description about the dataset. The
sitemap's “slicing” attribute specifies which shape graph is returned when a URI is
dereferenced. This is useful for the crawler to know in advance so that it knows how
much detail to expect. There are various ways to slice an RDF graph, depending
on how much detail about the URI should be returned. One common shape is the
concise bounded description (Stickler, 2005), which is a unit of specific knowledge
about a resource that can be used or exchanged by different software agents on the
Semantic Web. Given a particular node (the starting node) in a particular RDF graph
(the source graph), the concise bounded description subgraph is
All statements in the source graph where the subject of the statement is the
starting node.
For each of these statements that have a blank node as an object, recursively
add in all further statements in the source graph that have the blank node
as a subject.
If there are any reifications in the source graph, recursively include all con-
cise bounded descriptions beginning from the rdf:Statement node of
each reification.
The result of these steps is a subgraph where the object nodes are URI references,
literal values, or blank nodes that are not the subject of any statement in the graph.
This is also known as the “bnode closure” of a resource.
Since it is preferable in RDF not to worry about the direction of a property, an
alternative shape, the symmetric concise bounded description, has also been pro-
posed, which includes all statements on both the outbound and inbound arc paths of
the starting node.
7.6.2 v ocabUlarY of I nterlInkeD D ataSetS
While the Semantic Sitemap follows the standard Sitemap protocol and is
encoded in XML, VoID 27 offers more self-consistency as it is itself encoded in
Search WWH ::




Custom Search