Service Oriented Enterprise


Thursday, February 10, 2005

Semantic Web Services  

I've had several inquiries about semantic web services; I believe related to the recent predictions made by Graham Glass.

Personally, I'm a big fan. However, I'm thinking that the adoption curve of RDF, OWL and semantic ontologies is pretty far out. The use of first order predicate calculus applied against a structured knowledge base is an excellent idea (in peer to peer mode, when based on 'web of trust'), but again - this is quite a ways out as well. However, with Microsoft moving aggressively with XML in their products and now offering their "Information Bridge Framework", we will begin to see more emphasis on structured information being related by "exact match". By this I mean that the XML schema becomes a 'pivot point'. So, an XML schema that talks about "company X" may have several 'pivots': Company X's Partners, Products, Employees, etc. Thus the user will have the ability to 'navigate structured information'. With the inclusion of semantic ontologies, there will be less emphasis on having the pivot points being exact matches only. Eventually, the pivots will be navigable by synonym, hyponym or hypernym. This is extremely exciting capability, taking office productivity to a new level.

Most organizations tend to think of their web services and SOA efforts as "back office" functionality - a way to connect legacy systems. I am a firm believer that we will see a huge upside when we connect smart services to end user productivity tools. Like most people, I spend most of my day in a handful of tools: e-mail, word processing, spreadsheet, etc. Empowering users through better information is good business.

For those of you who've never queried a semantic ontology, give this a try:
http://www.cogsci.princeton.edu/cgi-bin/webwn
Make sure you look at hypernyms and hyponyms. If you're real interested check out the book "WordNet". I also recommend looking at the Cyc effort.

posted by jeff | 7:27 AM

archives