The increasing use of computational linguistics for semantic search and discovery tools requires much work on development and maintenance of associated ontologies. Related applications depend upon curated resources like dictionaries, gazetteers, etc. In order to scale these application models and leverage the respective communities of interest, a new set of tools is needed that facilitate community development and extension of these resources while retaining the curatorial model to ensure a reliable, high quality resource. We describe the requirements and principles for such a system, and present the CONCUR framework that addresses these needs. CONCUR defines a reputation model and a set of reusable infrastructure services to maintain the resource. The reputation model combines correctness as well as utility of participants' contributions, tracked over time and by sub-domain within the resource. We describe the architectural issues of the model, potential applications, and continuing research on the model.