Closed CoryCasanave closed 10 years ago
I believe Steve had expressed in interest in creating OCL to define GRA modeling constraints, re-assigning this issue to Steve.
I realize that writing OCL constraints in the normal way is not going to have any executable effect, because we are using InstanceSpecifications for all of the substantive modelling. As far as I know, current OCL tooling does not check InstanceSpecifications for constraint satisfaction. Plus Phase-1 processing is already a formal specification that will generate diagnostics where it doesn't like its input. So I decided not to pursue the OCL and instead use English to write constraints in the specifications of the annotation classes.
After discussion - put in OCL especially for realizations - make sure annotations realize the right kind of element and that the hierarchy is respected - ask Ed Willink about execution of InstanceSpecification constraints.
I've written many constraints about realizations, and also updated doc generation template to include them in the spec. According to Ed Willink their execution is not currently supported but is likely to be in a future release of Papyrus. So current status is purely documentation, but future tools might check them.
Tom to add OCL - related issue, report constraint violations