Many people have asked us why OPEN/Metis is not using UML as a notation.
Instead of using UML, OPEN/Metis proposes its own notation for a number of reasons. First of all, UML is not only a notation, but defines a set of modelling concepts such as class, attribute or association. This is usually known as the UML metamodel. The set of concepts defined by the metamodel of UML 2 is too large and not rich enough to support all the development needs addressed by OPEN/Metis. For example, UML does not include any concept similar to a Busy State, an Information Shape or a Control. The extension mechanisms of UML could probably help solve some of these issues, but the difference between the expressive power of UML and that of OPEN/Metis is so large that using UML is not cost-effective.
Furthermore, UML emphasizes diagramming over modelling, whereas OPEN/Metis assumes that modelling is the true aim of the software lifecycle while diagramming is just a means of depicting models. In addition, UML does not support notation forms other than diagrams, and OPEN/Metis needs other ways of depicting models such as sketches, structured text, free text and pseudocode, so UML would have to be supplemented with additional notation in order to cover all needs.
In addition, we believe that UML is flawed in several aspects (such as the semantics of whole/part relationships, the concept of stereotype and the lack of a clear distinction between attributes and associations), as described in a number of research papers published by different authors.
In any case, the notation described in the OPEN/Metis white paper is only a suggestion. Any notation capable of supporting the appropriate modelling concepts can be used instead. Furthermore, the OPEN/Metis metamodel can generate UML 2, so using it is still possible.