DOD developers build on XML success

DARPA researchers seek smart data sharing

DARPA Agent Markup Language

Following on the heels of the Extensible Markup Language revolution, the Defense Advanced Research Projects Agency is working on the next generation of markup languages with its DARPA Agent Markup Language (DAML).

DAML has been in development since 2000, but has yet to make a major splash on the tech scene. DAML builds on XML and is designed to allow a higher level of interoperability among devices, Web sites and databases.

The language's full name is DARPA Agent Markup Language Ontology Inference Layer (DAML-OIL), which combines U.S. and European efforts. Most, however, refer to it only as DAML.

Michael Dean is the principal investigator for the DAML integration and transition effort and a principal engineer for BBN Technologies of Cambridge, Mass. He has been involved with DAML research almost since its inception.

Dean said that DAML uses XML as a transport vehicle for data, but has a much greater capacity for describing objects and the relationships between them. It is designed to link disparate data from different sources and determine the relationships among the data.

"A close analogy would be that DAML will do for data what hypertext did for text," Dean said. "It is hard to write a program that will extract true information from a collection of data and then perform an automatic function. But that's what DAML is designed to do."

Dean said the World Wide Web Consortium (W3C), DARPA, and U.S. and European Union researchers have all contributed equally to the DAML project.

DARPA is developing DAML as a technology with intelligence built into the language through the behavior of agents. Agents are programs that can dynamically identify and comprehend sources of information. DAML is designed to allow agents to interact autonomously.

DAML's potential consists of enabling computers to recognize how different and disparate forms of information are linked and draw conclusions based on those links.

For example, a computer checks weather reports and determines that a traveler is going to be flying into an area that is expecting inclement weather. The computer automatically checks for other flights, flights to nearby areas and flights at earlier and later times, and alerts the traveler to those options.

"This is like a database schema on steroids," Dean said, referring to the technology that defines fields and tables in a database and the relationships among the fields and tables.

Dean said progress is being made every day and some, albeit few, applications are using DAML.

"The development community involved in this is both large and open-source," Dean said. "One of our experts is a 15-year-old kid in Chicago."

Uche Ogbuji, a DAML expert who works for Foursight in Colorado, has been following the progress of the language for three years. He said the DAML project has evolved into what is being called the Ontology on the Web Language (OWL) and is in W3C's hands more than DARPA's.

"There really is no difference between DAML and OWL, but OWL is the evolution of DAML once the W3C took over," Ogbuji said. "The W3C has a well-defined process for standardizing a language and is therefore uniquely suited to taking on the role of leading the charge."

Ogbuji said that even if DAML and OWL never take off the way HTML and XML did, they will still find an important role in business.

"We're probably not going to wake up one day and suddenly be using DAML or OWL," he said. "It will be a slow adoption as industries see its usefulness. Rather than see this in the Web, chances are OWL will be adopted in a smaller, controlled domain, such as within a specific business or government agency."

The future of semantic languages — those that can autonomously define relationships among data — is still uncertain, Ogbuji said. But their potential has prompted government and industry groups to invest a lot of time and money into their development.

***

DAML Overview

* DAML stands for Defense Advanced Research Projects Agency Agent Markup Language.

* The DAML program began with a kickoff meeting in August 2000 in Boston.

* The program's purpose is to develop a language that will allow computers to determine the relationships among data and make autonomous decisions based on those relationships.

* DAML is a mix of work by the World Wide Web Consortium (W3C) in Cambridge, Mass.; DARPA; European Union researchers; and academic researchers in the United States.

* The DAML project recently evolved into the Ontology on the Web Language and is continuing development under W3C's watch.