XML is not enough-Vision of a semantic web
A few decades back, three Jews were planning their future. The first said,’ I am going to settle down in London’. The second said,’ I am going to Canada’. The third took some time and said,’ I am going to Autstralia!’The first two asked,’ isn’t that far away?’ The third Jew asked back, ‘from where?’
The World Wide Web (WWW) has come a long way indeed. We used HTML to represent the information and view them via web browsers. Now we use XML to describe information more accurately by providing a ‘meaning’ to it by use of tags. However XML cannot describe relationships with respect to objects. In other words, it lacks what designers call ’ontologies’.It seems too much is never enough in the Information age. So where are we heading next after XML?
The vision is to provide a semantic grounding for the web where intelligent software agents will be able to process volumes of information found on the web. Well, software agents can be a something like persistent process/daemon, or a mobile code.OMG defines agent as “An Object that decides when to say go and when to say no”. What makes an agent truly intelligent is the fact that they can perform run-time integration via dynamic discovery and resource negotiation, which reiterates the fact that they are programs that operate at a high enough semantic level.So far software agents were used by the military, but their implications do not stop there. An agent application can be built on top of existing middleware (e.g., CORBA, EJB, and Jini) and can perform tasks like travel arrangements, meeting schedule coordination, enterprise workflow - e.g., sales, order processing, shipping, mobile computing and what not.
Let’s look at ‘Ontologies’ now. To put it simply, Ontologies are machine readable semantic specifications (might include terms, relations, and inference rules).OMG has intensively started working to extend the OMG Object Management Architecture (OMA) to better support agent technology.The World Wide Web Consortium is also active with the Semantic Web initiative which aims at the semantic grounding for the web which we were talking about.
Do we have something?
The Defense Advanced Research Projects Agency (DARPA) pulled seventeen research teams and 1 integration team consisting of experts in AI (Artificial intelligence), knowledge representation, logic and web technologies together with European Union IST Program and came out with an agent markup language which they called DARPA Agent Markup Language or DAML.
The DAML language is being developed as an extension to XML. The latest release of the language provides a rich set of constructs with which to create ontologies and to markup information so that it is machine readable and understandable. What’s new when I say “machine readable”? Even XML is machine readable. The difference here is DAML is aimed at “Resources (Like sensors, service, appliances)”, and not just web-pages. DAML have annotations linking their terms to ontologies across web pages, databases, legacy software, devices, and sensors and so on. If you tell a computer something new in DAML, it can give you new information, based entirely on the DAML standard itself. A set of DAML statements by itself (and the DAML spec) can allow us to interlink another DAML statement which is impossible with XML.Any DAML publication/manual can give you sufficient examples and cases to make understand this concept. But it does deliver what it says. (DAML is now actually supported by DARPA’s Information Exploitation Office).In short DAML has the capability to apply knowledge dynamically to find an answer rather that relying on predefined procedures. This will have extremely powerful implications in banking, finance and supply chain management domains.
Presently DAML tools range from ontology development and verification, web page annotation, distributed query processing and inference etc. DAML trial applications have been deployed in both US government and military domains. Currently an UML-Based Ontology Toolset (UBOT) is being developed by Lockheed Martin Advanced Technology Center.
- Ashwin Kumar N



