Ontologies are a popular research topic in various communities such as knowledge engineering, natural language processing, cooperative information systems, intelligent information integration, and knowledge management. They provide a shared and common understanding of a domain that can be communicated between people and across application systems. They have been developed in Artificial Intelligence to facilitate knowledge sharing and reuse. Recent articles covering various aspects of ontologies can be found in [Uschold & Grüninger, 1996] , [van Heijst et al., 1997] , [Studer et al., 1998] , [Benjamins et al., 1999 (a)] , [Gomez Perez & Benjamins, 1999] , [Fensel, 2000] . An ontology provides an explicit conceptualization (i.e., meta information) that describes the semantics of the data. They have a function similar to a database schema. Some differences are: 1
Currently computers are shifting from being single isolated devices to becoming entry points into a worldwide network of information exchange and business transactions. Given the exponential growth of on-line information available, an automatic processing of this information becomes necessary for keeping things maintainable and accessible. Automatic processing of information requires a machine-understandable representation of its semantics. Providing shared and common domain structures becomes essential and ontologies will therefore become a key asset in information exchange -- being used to describe the structure and semantics of information exchange. Such technologies will play a key role in areas such as knowledge management and electronic commerce, which are market areas with incredible growth potential.
In the area of information systems and intelligent information integration, we can distinguish different integration tasks that have to be solved in order to achieve a completely integrated access to information [Stuckenschmidt, submitted] :
As we already pointed out, ontologies are good candidates for providing the shared and common domain structures which are required for a truly semantic integration of information sources. The question then becomes: how will we describe such ontologies? A prerequisite for such a widespread use of ontologies for information integration and exchange is the achievement of a joint standard for describing ontologies. Take the area of databases as an example. The huge success of the relational model would never have been possible without the SQL standard that provided an implementation-independent way for storing and accessing data. Any approach that tries to achieve such a standard for the areas of ontologies has to answer these questions: What are the appropriate modeling primitives for representing ontologies? How can we define their semantics? and What is the appropriate syntax for representing ontologies? In the US, research funding agencies have already recognized the importance of such issues in setting up the DAML program 3 , which aims at a machine processable semantics of information sources which are accessible to agents.
In this paper we present a proposal for such a standard way of expressing ontologies based on using new web standards like XML schemas and RDF schemas: the Ontology Inference Layer . It is important to note that we intend OIL to be extensible, and that only the core language is described in this paper. This language has been designed so that:
It is envisaged that this core language will be extended in the future with sets of additional primitives covering areas such as concrete data types (e.g., numbers and strings) and extensional class definitions, with the proviso that full reasoning support may not be available for ontologies using such primitives. A further level of extension could include modelling primitives such as defaults, fuzzy/probabilistic definitions, additional collection types (e.g., bags and lists), and a more expressive axiomatic language. In a nutshell, we do not want to present the final version of OIL in this paper. We want to make a proposal opening the discussion process that may finally lead to a useful and well defined consensus amongst a large community making use of such an approach.
This paper is organized as follows. Section 2 provides the general background for the discussion on OIL (i.e., our position). Section 3 provides the language primitives of OIL and discusses technical support for OIL. We also sketch possible directions for extending OIL. Section 4 compares OIL with other ontology languages and web standards such as XML and RDF. Finally, a short summary is provided in Section 5 . The appendix provides syntax definitions of OIL in XML and RDF plus a formal semantics of OIL.
1. See [Klein et al., 2000] , for an elaborated comparison of database schemes and ontologies.
2. More precisely, ontologies are used for the semantic integration of information sources. The language we provide for ontology interchange is not for semantic integration of ontologies but for ontology interchange via reuse (i.e., reusing an ontology written in another language). We do not deal with the integration of heterogeneous ontologies in this paper.