Designing a Metadata Model for Unstructured Document Management in Organizations

At present Document Management represents a critical problem for organizations, where a growing amount of information is produced in unstructured format. Metadata sets for Document Management addressing organizational needs are lacking, as most research efforts for Metadata specifications are focused on digital libraries and web communities requirements, rather than organizational ones. In this paper we propose a Metadata Model for Unstructured Document Management, called DMSML Document Management and Sharing Markup Language, which aims at modeling intrinsic properties of documents (e.g. title, author and keywords) as well as the relationships with the organizational context. Furthermore, while most metadata specifications don't rely on a well-developed data modeling methodology, our proposal is based on a three-layered data modeling approach, distinguishing a conceptual, logical and physical layer. This approach encourages understanding and easy adoption of metadata specifications among end users. At present we are developing a DMSML Framework prototype for validation purposes.