I was wondering if it would help to have NORMA assign deontic modality when a FactType uses constraints that are not implemented in the DLL/DML generation. By setting deontic modality for such FTs, the modeler would have a visual reference to indicate that the FactType would not be enforced, unless the enforcement procedures were provided for outside of the NORMA tool. The downside is that the intent of the modeler may well be to show this factType as having alethic modality, considering enforcement a separate issue. For this reason, even if deontic modality was the default for these types, the modeler should have the option to change the modality.
Whether this method is used or not, some indication of non-implemented/enforced constraint types would be useful at the model level. Also, the ability to review all deontic (and/or non-implemented), constraints before logical schema generation would be very useful. As the error checking already in place helps to prevent violations of ORM theory from passing through the mapping process, such a review would give the modeler an opportunity to confirm or compensate for those parts of the model that may not be processed as expected.