[om-list] The Semantic Web

[Mark Butler]

Hello everyone,

There is a considerable amount of research being conducted to add a semantic layer to the current world wide web that will provide all the benefits of a generalized knowledge representation system.

The current Internet infrastructure is an example of a classic tradeoff of performance in favor of interoperability.  Most of the higher level Internet protocols are all text based for a variety of reasons - interoperability, ease of implementation, ease of debugging, and so on.

You see a more extreme example of the same thing when you look at a typical XML file.  XML is the most inefficient data representation format conceivable, but it is rapidly becoming the metalingua franca for Internet data representation.

There are a number of researchers working on extending ontological structure to disparate XML data collections.  Such ontologies alone could make possible semantic queries like:  Search for all historical sites within 100 miles of Philadelphia.

There is also work on extending an XML based standard called RDF (Resource Description Framework) to handle arbitrary subject-object-verb relationships.

As often mentioned, XML is hardly ideal as an internal representation for any high performance system.  However, despite its quirks, it makes a great frame of reference for research as well as an effective export metaformat for data interchange between projects.

The LISP centric OpenCyc project has developed an XML schema and plans to release an open source version of the core ontology (6000 concepts, 60000 assertions) in an XML format any day now.

Some type of standard semantic XML interface to the One Model software would be very helpful.  See:

  http://www.semanticweb.org/
  http://www.opencyc.org/

- Mark