The DARPA Agent Markup Language

James Hendler[1] and Deborah L. McGuinness[2]



The DARPA Agent Markup Language (DAML) program is a United States government sponsored endeavor aimed at providing the foundation for the next web evolution – the semantic web.  The program is funding critical research to develop languages, tools and techniques for making considerably more of the content on the web machine-understandable.  We believe that this will lead to the next major generation of web technology, and will enable considerably more “machine to machine” (agent-based) communication. The program includes participation from academic researchers, government agencies, software development companies, and industrial organizations such as the World Wide Web consortium (W3C).  The DAML project is also working closely with other efforts, including European Union funded Semantic Web projects (e.g. On-to-Knowledge[3] and Ibrow[4]), and the ongoing W3C RDF recommendation effort[5].  In the remainder of this report, we provide a short motivation, description, and status report about the program.



The modern information technology world is a dynamically changing environment with an exponentially increasing ability to create and publish data that rapidly swamps human abilities to process that data into information. Agent-based computing can potentially help us to recognize complex patterns in this widely distributed, heterogeneous, uncertain information environment. Unfortunately, this potential is hampered by the difficulty agents face in understanding and interacting with data that is either unprocessed or in natural languages. The inability of agents to understand the conceptual concepts on a web page, their difficulty in handling the semantics inherent in the outputs of a program, and the complexity of fusing information concept from the outputs of sensors, to name but a few problems, truly keep the "agent revolution" from occurring.


 One potential solution to this problem is for humans to, as it were, meet the computer half way. By using tools to provide mark-up annotations attached to data sources, information can be made available to the agents in new and exciting ways. Going beyond XML, the goal of this program is to develop a language aimed at representing semantic relations in machine-readable ways compatible with current and future Internet technologies. Further, prototype tools are being developed to show the potential of such markups to provide revolutionary capabilities that will change the way humans interact with information. Deployment of such tools to military and intelligence users, and showing the incredible dual use potential of such a technology, caps off the programs goals.


 To realize this solution, Internet markup languages must move beyond the implicit semantic agreements inherent in XML and community-specific controlled languages, and move towards making semantic entities and markup a primary goal.  To this end, DARPA is working with numerous partners and communities to create an eventual web-standard semantic language and to demonstrate the utility of such a language.  We are doing this through the development of an example language and sample tools and applications.  The language is current called DAML - the DARPA Agent Markup Language.  DAML is a language that will tie the information on a page to machine-readable semantics (specifications of term meanings stored in ontologies) and eventually provide a logical language embedded on the web. The language will allow for communities to extend simple ontologies for their own use, allowing the bottom-up design of meaning while allowing sharing of higher-level concepts. In addition, the language will provide mechanisms for the explicit representation of services, processes and business models, so as to allow non-explicit information (such as that encapsulated in programs or sensors) to be recognized and understood. Eventually, the language will also supply a mechanism by which logical statements and proofs can become first-class web entities, allowing a new set of capabilities in machine to machine communication.  This will enable the development of a wide new range of software tools for both industrial and government applications including diverse uses ranging from business-to-business e-commerce to government efforts in combating the spread of weapons of mass destruction.


DAML will provide a number of advantages over current markup approaches. It will enable semantic interoperability -instead of stopping at enabling only syntactic interoperability as is done in XML. Objects on the web can be marked (manually or automatically) to include descriptions of information they encode, descriptions of functions they provide, and/or descriptions of data they can produce. This will allow web pages, databases, programs, models, and sensors all to be linked together by agents that use DAML to recognize the concepts they are looking for. This will allow web-based information fusion from diverse sources to become a reality.


DARPA’S DAML program was kicked-off in the summer of 2000 and research is ongoing.  The language is being delivered in two portions – the first is an “ontology” language, DAML-ONT, with a later DAML-LOGIC to follow. DAML-ONT v0.5 was released on October 5, 2000, and an ongoing discussion of the language and logic issues, with the goal of reaching a more stable version, is being held on  The language is extends W3C’s Resource Description Framework [Lassila-Swick:1999, Lassila:1998] and its associated object-oriented type system [Brickley-Guha:2000]. It aims to add expressive power suited to agent and service interoperation. The DAML-ONT specification as well as supporting documents can be found on Also available on this web site are early versions of DAML support tools and pointers to various web resources marked up with DAML.  (An early release of DAML-LOGIC is expected to come out in the spring of 2001).


The goal of DAML-ONT is to capture the commonly used modeling primitives used in object oriented-modeling, frame systems, and conceptual schemas and include them in an integrated language for the web.  The language attempts to join the ease of modeling in frame systems, the ubiquity of the web, and the formal foundations of knowledge representation in description logics to provide a sound language for representing and reasoning with term meanings.  Currently DAML-ONT is being used to mark up project pages used in the DAML project and is being used to facilitate services and applications work.  In addition, a number of other communities (including the OIL project described in the accompanying report) are developing tools to translate markup to the DAML-ONT language and to provide pages in the DAML repository[6].


The next goal of the project is to create an early version of a logic language—DAML-LOGIC. At the time we are writing this article, DAML-LOGIC is expected to include both a language for expressing constraints in DAML-ONT and for adding inference rules to the language.  In addition, work continues on the evolution of DAML-ONT as we publish documents showing its intended meaning[7][Fikes-McGuinness:2000] and how it maps to current research in semantic web languages such as OIL[8][Bechhofer-et-al:2000], SHOE[9][Heflin-Hendler:2000], and KIF[10] as well as to emerging web languages like XML and RDF, and to ongoing efforts in agent-based systems particularly the FIPA[11] standardization effort.


We also are working hard to encourage usage from a broad spectrum of users – sometimes users will use the language as a unifying language for stating explicit information and not connect to deductive engines that look for logical implications of the implicit information.  Other times, users will connect to more or less complete inference engines that infer the logical completion of all of the statements.  This provides the possibility for a spectrum of users and does not require all users to have extensive computational resources connected to their systems.  Ultimately though, we believe the language will be used to express the meaning of information on web pages or in applications, and will lead to a web standard language for expressing semantic content.


Information concerning DAML-ONT and DAML-LOGIC is published on a publicly available web site:  Also, open discussion is encouraged on the World Wide Web mailing list www-rdf-logic.  Interested researchers and users are strongly encouraged to view the postings and participate in the evolution of the language and its use.


Acknowledgements:  The DAML program is funded by the US Defense Advanced Research Projects Agency.  The program is the result of many people’s work and the evolving list of major contributors is maintained on




[Bechhofer-et-al:2000]  Sean Bechhofer, Jeen Broekstra, Stefan Decker, Michael Erdmann, Dieter Fensel, Carole Goble, Frank van Harmelen, Ian Horrocks, Michel Klein, Deborah L. McGuinness, Enrico Motta, Peter Patel-Schneider, Steffen Staab, and Rudi Studer. ``An informal description of OIL-Core and Standard OIL: a layered proposal for DAML-O''. Posted at


[Brickley-Guha:2000] Dan Brickley & R.V. Guha, "Resource Description Framework (RDF) Schema Specification 1.0", W3C Candidate Recommendation, 27 March 2000, available as


[Fikes-McGuinness:2000]  Richard Fikes and Deborah L. McGuinness.  “An Axiomatic Semantics for DAML-ONT”.  November 13, 2000, available as


[Heflin-Hendler:2000]  Jeff Heflin and James Hendler.  “Semantic Interoperability on the Web”.  In: Proceedings of Extreme Markup Languages 2000. Graphic  Communications Association, 2000. pp. 111-120.


[Lassila:1998]  Ora Lassila, "Web Metadata: A Matter of Semantics", IEEE Internet Computing, July/August 1998.


[Lassila-Swick:1999] Ora Lassila & Ralph Swick, "Resource Description Framework (RDF) Model and Syntax Specification", W3C Recommendation, 22 February 1999, available as .


[1] Chief Scientist DARPA/ISO, 3701 N. Fairfax Dr., Arlington, VA 22203 ,

[2] Associate Director and Senior Research Scientist, Knowledge Systems Laboratory , Stanford University , Stanford, CA 94305,





[7] and .