Report on the 2nd Workshop on Model Development and Validation - MoDeVa

摘要

The content of discussions led us to draw the following conclusions. First of all, the usage of formal tools to treat UML specifications is really meaningful when dealing with critical system specifications. This is due to the fact that potential users in the field of critical system design require having a simple, totally formally grounded semantics to a subpart of the UML. Using formal tools in a different context makes less sense. Secondly, in order to be compliant with the norm, defining a subpart of the UML to be mathematically grounded could be done using a profile approach. But profile themselves should be provided with a semantics. In the next edition of MoDeVa we propose to concentrate on these issues: What subpart of the UML should be considered in the field of formal treatment? Are there several subparts (possibly overlapping) of the UML to be considered depending on the system design domains considered? How this subpart(s) should be described? How to provide and describe a formal semantics in a way which would be acceptable for the OMG?