As the importance of enterprise computing systemscontinues to grow so does the need for sound but flexiblerepresentations of the information they manipulate. Thishas created a growing interest in informationrepresentation languages that are not only easy for domainand business experts to use but are also amenable tocomputer manipulation. Since traditional informationrepresentation languages have tended to focus either onhuman usability e.g. UML or machine processability e.g.OWL there is currently no language intended for humanuse that cleanly satisfies both requirements. In this paperwe discuss the different schools of thought on how to solvethis problem, and analyze the various concrete proposalsthat have been put forward. We then present our own viewson how best to meet this challenge.