KMWorld 2024 Is Nov. 18-21 in Washington, DC. Register now for Super Early Bird Savings!

Semantic Web holds promise for KM

The Semantic Web is relevant to knowledge management because it has the potential to dramatically accelerate the speed with which information can be synthesized, by automating its aggregation and analysis. Information on the Web now is typically presented in HTML format, and while very beneficial in some respects, the format offers neither structure nor metadata that is useful for effective management. Without structure, elements of content cannot be related to each other, and without metadata, the nature of the elements themselves cannot be known. The discovery process is therefore human-centric and time-consuming.

The Semantic Web is designed to provide the following missing components: structure, through the use of XML tags; metadata descriptors, through the Resource Description Framework (RDF); and relationships, through the Web Ontology Language (OWL). Two years ago, the World Wide Web Consortium (W3C) finalized the standards for RDF and OWL. A variety of software tools and applications supporting the standards have been developed by leading software companies such as Adobe, Hewlett-Packard, IBM and Oracle.

No one is arguing that the prospect of tagging all the information on the Internet is a realistic one, or that a monolithic ontology covering all topics can be developed. However, within various professional communities such as life sciences, healthcare and financial services, developing ways of integrating information via the Semantic Web is an achievable target. Software tools that are already available include the following:

  • Adobe has developed the XMP platform as its brand of RDF, and most of Adobe's major products, including Acrobat, Illustrator and Photoshop, support the XMP platform now.
  • Hewlett-Packard's open source toolkit Jena is used to create applications for the Semantic Web. It has been used in a large percent of Semantic Web applications to date.
  • IBM is positioning itself as a major player in the semantic infrastructure domain, with its Web ontology development and management kits as well as semantic tools for Web services.
  • Oracle 10g Release 2 incorporates native support for RDF bundled with the Oracle Spatial option, along with rules and query language to integrate RDF with relational tables.

Although the challenges are significant, the case is very strong for knowledge management to make use of the Semantic Web. Searching unstructured data on the Web is a tedious and often unrewarding task, producing a large number of false hits and an even larger number of possible matches that need to be reviewed. Ambiguity intrinsic to language is one culprit ("bow" can refer to archery, a ship or gift wrapping), and lack of semantic structure is another. It is not possible with today's Web to search for articles authored by a particular individual, for example. Hits will include every mention of the person, whether as an author or not. On the Semantic Web, use of an ontology could disambiguate the word "bow," and "author" could become a metadata field used in a search.

Intelligent search can overcome some of the obstacles that stand between users and the information they need, particularly in the context of enterprise information that is already tagged. But search cannot move to the next step of integrating and aggregating data for interpretation. That's where even applications using structured data such as that in a relational database can benefit from use of Semantic Web standards.

KMWorld Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues