Making XPath Reach for the Web-Wide Links
Abstract:
Opting for link semantics in XML is almost like hyperlinks for HTML documents. XLink describes a standard way to add hyperlinks to an XML document. The current XPath technologies are restricted to follow ID/IDREF(S) links for intra-document navigation only. We investigate the Lightweight Directory Access Protocol (LDAP) that offers a rich collection of primitives to express links among distributed data collections in the network, and facilities to follow links when searching. In virtue of querying the underlying LDAP model by referrals, we developed an extended XPath processor that is capable of addressing links of any type (IDREF(S), XLink) embedded in XML data on the Web. Links may be inter-document, or even traverse across different local or remote servers. We describe the internal LDAP data representation and query model used by the processor for the storage and querying in XPath of XML documents based on links, and provide examples to illustrate them. We complement the discussion with experimental analyses that prove the efficiency of our query evaluation techniques. The latter confirms the relevance of our approach for applications that need to interact with interlinked XML document networks in an XPath-like fashion.