Abstract. With the wider use of ontologies in the Semantic Web and as part of production systems, multiple scenarios for ontology maintenance and evolution are emerging. For example, successive ontology versions can be posted on the (Semantic) Web, with users discovering the new versions serendipitously; ontology-development in a collaborative environment can be synchronous or asynchronous; managers of projects may exercise quality control, examining changes from previous baseline versions and accepting or rejecting them before a new baseline is published, and so on. In this paper, we present different scenarios for ontology maintenance and evolution that we have encountered in our own projects and in those of our collaborators. We define several features that categorize these scenarios. For each scenario, we discuss the high-level tasks that an editing environment must support. We then present a unified comprehensive set of tools to support different scenarios in a single framework, allowing users to switch between different modes easily. Evolution of Ontology EvolutionAcceptance of ontologies as an integral part of knowledge-intensive applications has been growing steadily. The word ontology became a recognized substrate in fields outside the computer science, from bioinformatics to intelligence analysis. With such acceptance, came the use of ontologies in industrial systems and active publishing of ontologies on the (Semantic) Web. More and more often, developing an ontology is not a project undertaken by a single person or a small group of people in a research laboratory, but rather it is a large project with numerous participants, who are often geographically distributed, where the resulting ontologies are used in production environments with paying customers counting on robustness and reliability of the system.The Protégé ontology-development environment 1 has become a widely used tool for developing ontologies, with more than 50,000 registered users. The Protégé group works closely with some of the tool's users and we have a continuous stream of requests from them on the features that they would like to have supported in terms of managing and developing ontologies collaboratively. The configurations for collaborative development differ significantly however. For instance, Perot Systems 2 uses a client-server mode of Protégé with multiple users simultaneously accessing the same copy of the ontology on the server. The NCI Center for Bioinformatics, which develops the NCI The-1 http://protege.stanford.edu 2

No abstract