Database Reference
In-Depth Information
catalogue of mapping products, that is, information that is related to the HTML con-
tent and hence can help with faceted browsing or semantic search. There is less of a
compelling use case for encoding GI data such as that in
Table 7.1
in RDFa as it is not
directly related to any human-readable information on an HTML Web page and can
only really be used and processed if it is converted from RDFa to RDF.
7.5.7 p
UblIShInG
a
l
InkeD
D
ata
v
IeW
on
a
r
elatIonal
D
atabaSe
or
t
rIple
S
tore
Since Merea Maps has a large volume of data that changes frequently and is not
directly related to particular HTML Web pages, the most obvious approach for
its RDF publication would be as a Linked Data view on an RDB or triple store.
If it wants to stick to just one workflow involving its existing GIS, a Linked Data
view of its RDB can be published using a tool such as Virtuoso or D2R Server, as
described in Section 7.5.3. If, however, Merea Maps wanted to store its RDF data in
a manner more suited to the RDF graph structure, it would be well advised to use a
“triple store,” a non-RDB optimized for storing Linked Data. Particular examples
of these are discussed in Section 7.10 on software tools. Many triple stores pro-
vide a Linked Data interface to make RDF descriptions of stored resources avail-
able via HTTP; however, for those that do not include a Linked Data interface as
standard (that is, they store the data using URIs that are not dereferenceable), there
are tools such as Pubby
25
and Djubby
26
that sit in front of the triple store's SPARQL
endpoint to serve Linked Data. When the HTTP request comes in as a URI for a
Linked Data resource, Pubby rewrites the request into a SPARQL DESCRIBE query
(see Chapter 8) by mapping the dereferenceable Linked Data URI onto the original
triple store's URI, and the query is then put to the triple store. Pubby also handles
303 redirects (see Section 7.3.1) and negotiates between the HTML, RDF/XML, and
Turtle descriptions of the resource. Djubby operates in a similar way but is written in
Python rather than Java and can be used with the Django Web framework.
7.6 DESCRIBING THE LINKED DATASET
Now that Merea Maps has been able to create and serve up its Linked Data, it needs
to consider how people can find this Linked Data and know what it is about. To
assist data discovery, any related HTML documents should have a
<link>
tag
inserted into the header to point to the RDF file. This helps make the RDF visible to
Web crawlers and Linked Data browsers and is known as the Autodiscovery pattern
(Dodds and Davis, 2012):
<link rel = “alternate” type = “application/rdf+xml” title =
“Topographic Object Data” href = “
http://mereamaps.gov.me/topo.rdf”
>
Merea Maps provides metadata alongside its data to assist third parties' under-
standing of what is offered. The purpose, scope, and competency questions, prov-
enance, licensing, and currency (when the dataset was last updated) are all important
elements of the metadata that should be provided with the Linked Data set to encour-
age not only reuse but also
accurate
reuse. There are currently two methods available
