Information Retrieval and the Semantic Web

The Semantic Web has lived its infancy as a clearly delineated body of Web documents. That is, by and large researchers working on aspects of the Semantic Web knew where the appropriate ontologies resided and tracked them using explicit URLs. When the desired Semantic Web document was not at hand, one was more likely to use a telephone to find it than a search engine. This closed world assumption was natural when a handful of researchers were developing DAML 0.5 ontologies, but is untenable if the Semantic Web is to live up to its name. Yet simple support for search over Semantic Web documents, while valuable, represents only a small piece of the benefits that will accrue if search and inference are considered together. We believe that Semantic Web inference can improve traditional text search, and that text search can be used to facilitate or augment Semantic Web inference. Several difficulties, listed below, stand in the way of this vision.

Current Web search techniques are not directly suited to indexing and retrieval of semantic markup. Most search engines use words or word variants as indexing terms. When a document written using some flavor of SGML is indexed, the markup is simply ignored by many search engines. Because the Semantic Web is expressed entirely as markup, it is thus invisible to them. Even when search engines detect and index embedded markup, they do not process the markup in a way that allows the markup to be used during the search, or even in a way that can distinguish between markup and other text.

…..

There is no current standard for creating or manipulating documents that contain both HTML text and semantic markup. There are two prime candidates for such hybrid documents. First, semantic markup might be embedded directly in an HTML page. Unfortunately, while we call approaches like RDF and OWL semantic markup, they are typically used not as markup but rather as stand-alone knowledge representation languages that are not directly tied to text. Furthermore, embedding RDF-based markup in HTML is non-compliant with HTML standards up to and including HTML 4.0. This issue is currently under study by a W3C task force [23].

Download file here