Assisting the Discovery and Reuse of Document-based Knowledge using Semantic Metadata

Due to the high importance computer-mediated communication and IT systems have in the execution of today’s business processes, enterprises are producing a rapidly growing amount of information as by-product of their business activities (Zadok et al. 2004, p. 2f). Lyman and Varian (2003, p. 4) estimate the annual growth of newly created information to be 30%. According to studies of Merrill Lynch and Ferris Research only 20% 40% of this information is stored in a structured, semantically described form in electronic databases or structured business applications (e.g. SAP R/3). The other 60% 80% are contained in the unstructured form of electronic documents1 (e.g. Microsoft Office files, emails, images, multimedia files, or web based content). This is even more dramatic as the annual growth rate of newly created unstructured content is considerably higher than that of structured information. (Ferris Research 2008; Blumberg and Atre 2003) The unstructured nature of the bulk of existing as well as newly generated information is cause for a whole host of inefficiencies and problems. In a single-user desktop environment unstructured documents are spread across local folder hierarchies and contained as attachments in email messages. This leads to potential redundancy and an increased required effort to locate document-based knowledge. Such problems multiply in a multi-user enterprise environment with numerous competing repositories, document management systems, shared network drives and local file systems for each distinctive user. On one hand a person’s ability to overlook the available number of information sources diminishes as the amount of documents in numerous different repositories grows. On the other hand a multi-