Customizable Workflow Support for Collaborative Ontology Development

As knowledge engineering moves to the Semantic Web, ontologies become dynamic products of collaborative development rather than artifacts produced in a closed environment of a single research group. However, the projects differ—sometimes significantly—in the way that the community members can contribute, the different roles they play, the mechanisms they use to carry out discussions and to achieve consensus. We are currently developing a flexible mechanism to support a wide range of collaborative workflows in the Protégé environment. In this paper, we describe our overall architecture that supports flexible collaborative workflows in Protégé. This architecture comprises an ontology for representing workflows for collaborative ontology development, a customizable ontology-development environment that our system generates based on a declarative description of a workflow, and a run-time integration with a workflow execution engine. 1 Overview of Workflows for Collaborative Ontology Development Collaborative ontology development has become an active area of research and practice. On most large projects today, ontology development is a collaborative effort, involving both ontology engineers and domain experts. The number of users participating in development ranges from a handful (e.g., the Foundational Model of Anatomy [7]), to a couple of dozens (e.g., the National Cancer Institute’s Thesaurus [9]), to the whole community contributing to the ontology in some way (e.g., the Gene Ontology [3]). With larger groups of users contributing to ontology content, many organizations define formal workflows for collaborative development, describing how project participants reach consensus on definitions, who can perform changes, who can comment on them, when ontology changes become public and so on. Some collaborative projects have been publishing and refining their workflows for years (e.g., the Gene Ontology). In other projects, ontology researchers are working actively with domain experts to make these workflows explicit and to provide tooling for supporting the specific workflows (e.g., ontology development for the UN Food and Agriculture Organization (FAO) in the NeOn project [4]). These workflows differ from project to project, sometimes significantly. A workflow for a specific project usually reflects that project’s organizational structure, the size and the openness of the community of contributors, the required level of rigor in quality control, the complexity of the representation, and other factors. We are currently working on providing comprehensive support for collaborative ontology development in the Protégé system.1 Protégé enables users to edit ontology in a 1 http://protege.stanford.edu distributed manner, to discuss their changes, to create proposals for new changes and to monitor and analyze changes. Integrating support for collaborative workflows in such a system would mean having the tool itself “lead” the user through the workflow steps. For example, the tool can enable or disable certain options, depending on the user’s role, indicate to the user the current stage of the workflow and the actions expected or required from this user at this stage, or enable a user to initiate new activities. Because workflows may differ significantly from project to project, developers must be able to custom-tailor the workflow support in terms of both the execution steps and the user interface. Our goal is to develop a framework that would support as wide a variety of workflows as possible. We envision that as a group of ontology developers defines the workflow process for a new ontology project, they will describe this process declaratively using our workflow ontology. This description may include the list of roles and corresponding privileges, the steps that a change must go through in order to be published, the way tasks get assigned and executed, and so on. Our tools will then use this description to generate a custom-tailored ontology-development environment. The team can then use this environment for their collaborative development, with a workflow engine controlling the execution steps. This paper makes the following contributions: – We develop an architecture for supporting customizable workflows for collaborative ontology development. This architecture integrates tools for ontology development (Protégé), declarative description of workflows (a workflow ontology), discussion support, and a workflow engine (Section 3). Our design is driven by a set of requirements (Section ??) that we have identified by studying a large number of projects that use collaboration in ontology development. – We present a prototype implementation of the architecture by • enabling generation of a custom-tailored ontology-deveopment environment from a set of workflowontology instances (Section 5); • mapping our workflow ontology to a workflow engine in JBoss (Section ??); • providing a Protégé-based implementation of specific activities in the workflow engine (Section ??).