David Wood, Zepheira
Persistent URLs (PURLs) have operated on the Internet since 1995, and were re-architected and described at SemTech in 2008. Part of the re-architecture was to add semantic capabilities to PURLs so that metadata descriptions of a resource may be computationally separated from resources themselves. The Open Source community at http://purlz.org maintains and extends the PURL code base.

Active PURLs extend the PURL concept. Active PURLs actively participate in the creation or gathering of metadata that describes their target URL. Using principles of Representational State Transfer (REST), implementation details are hidden behind a PURL’s URL. Any form of query mechanism may be used to implement an Active PURL. This talk, however, will introduce an Active PURL builder that uses the SPARQL query language for RDF.
Active PURLs using SPARQL may be used in a manner analogous to stored procedures for relational databases. SPARQL’s distributed query capabilities may be used to gather information from across the World Wide Web and present results at a single, persistent URL. Active PURLs with SPARQL thus treat the Web like a database. Examples are presented using FOAF (“What are my friends’ current telephone numbers?”), Dublin Core/RDFa (“Who authored updates to the Web pages I track?”) and GeoNames/SKOS (“Find known places in a postal code”, which demonostrates parameterization).