On 11/6/13 5:48 AM, Fabian Neuhaus wrote:
- fUML is -- to the best of my knowledge -- not equipped with a model theoretic semantics, but with an operational semantics
Please see http://www.omg.org/spec/FUML/1.1/, from p 351 on, where the
semantics is given in terms of CLIF.
I don't know how this semantics should be classified, but this is what
"the semantics of fUML" that is being referred to.
Tara, The draft RFP says A great diversity of logical languages with model-theoretic semantics is in use for these purposes:
[…] " • the modeling language UML [UML] (fUML [FUML] equips part of UML with a formal semantics)
I am not an expert on fUML, but what I understood from Conrad Bock , who participated in the fUML working group, the CL formulas are not providing a semantics for UML, but providing constraints for the operational semantics of UML. This is consistent with what the fUML spec it says:
"Base UML is expressive enough to define the execution
model and must be used when specializing the execution model through explicit variation. The base semantics specifies
when particular executions conform to a model defined in bUML. It does not generate executions. In particular, the base
semantics does not define a virtual machine to execute models directly. The base semantics is expressed in axioms of first
order logic. This has the advantage of being completely explicit, rather than using text to explain the behavior of a virtual
machine. This enables automatic determination of whether an execution conforms to the execution model."
Hence, I don't think that one can claim that UML (via fUML) has a model theoretic semantics.
Best Fabian
|