Enhancing the Web for Knowledge Management
1. Introduction
In order to have a realistic, hospitable and practical new Semantic Web, certain challenges must be faced. The concepts developed as part of the Topic Maps standards offer some answers and inspirations.
2. Topic Maps: The HTML that is needed by the Semantic Web?
-
The Web has been successful because the HTML language is simple enough to use and understand. The Semantic Web needs a common language for representing units of knowledge and their interconnections. Topic Maps suggests an achievable first step in that direction.
-
Robust interchange of information on the Semantic Web requires a common understanding of managing addresses of information. The idea of a network of interconnected binding points indicating or constituting subjects, as developed by the Topic Maps standard, provides an element of an answer.
3. Outline
3.1. The problem
-
The Semantic Web promises to allow interdependent complexes of facts (assertions) to be amalgamated in such a way as to allow knowledge to be inferred from them.
-
The Web must face certain issues, and gain certain new features in order to make this practical.
3.2. Some issues to be faced
-
Semantic reliability of information addresses. There is no such thing as "dynamic information".
-
The meaning of information object identity, and the criteria for evaluating/establishing it.
-
In the current Web culture, the central focus is not yet electronic information management. The central focus must move from being publishing- and technology-centered toward being management- and information- centered. The central focus must change from primarily valuing technology to primarily valuing content.
-
The Semantic Web must provide a solid technical foundation for providing access to the full diversity of perspectives on Web-based information, while providing effective ways to avoid infoglut and preserve the productivity of Semantic Web users.
3.3. What the Topic Maps Paradigm has to offer
The Topic Maps Paradigm uniquely offers lessons and implicitly proposes some basic facilities, features, and requirements for the Semantic Web:
-
A neutral architecture/framework for encoding many kinds of semantic networks, from the simplest to the most sophisticated.
-
The vital distinction between subject-indicating and subject-constituting resources.
-
The modular distinction between the names, addresses, and other characteristcs of notions ("subjects"), on the one hand, and the subjects themselves, on the other hand.
-
The concept of scope as it can be applied to representations of relationships, and how it can be used to control infoglut.
-
Multilingual naming, and an end to the widespread confounding of the issues surrounding names, on the one hand, and information addresses, on the other hand.
-
The ability to control the merging ("federation") of knowledge resources, and the ability to represent merging instructions interchangeably, as a <topicMap>.
-
A mechanism for global knowledge federation via "published subjects". This mechanism facilitates the federation of knowledge resources of unlimited diversity.
-
The ability to federate other kinds of knowledge resources, including RDF, NewsML, UDDI, DOI, Dublin Core.