I think the important point is to have a formal, mathematically precise
semantics, let it be model-theoretic or operational. If needed, an
operational semantics can be turned into a model-theoretic one, by
considering expressions to be sentences which have a singleton model
class consisting of their operational semantics. (01)
Best, Till (02)
Am 06.11.2013 14:06, schrieb John F Sowa:
> On 11/6/2013 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
> The fUML document is badly organized, and it should be rewritten
> to show exactly what it is claiming and how it is specified.
>
> On the one hand, the authors claim to be specifying UML
> diagrams in Common Logic (with the CLIF dialect). On the
> other hand, they claim to be making UML diagrams "executable".
>
> But a declarative language, such as CLIF, isn't executable
> *unless* you use logic programming methods (e.g. Prolog or SQL)
> to add a "goal" or a "trigger". The SQL WHERE-clause, for
> example, is purely declarative. You can specify it by defining
> how any WHERE clause is translated to CL (or other version of FOL).
>
> But the SELECT verb makes the WHERE-clause a goal to be satisfied
> by a search procedure. For DB updates, it can be used as a trigger
> that raises an error condition when a constraint stated by the
> WHERE-clause is violated.
>
> There are two kinds of UML diagrams:
>
> 1. Declarative, such as the type hierarchies and the
> E-R diagrams.
>
> 2. Procedural, such as the activity diagrams, which are
> inspired by but not identical to Petri nets.
>
> For the declarative diagram types, it's possible to give
> a formal definition by specifying how each diagram type
> is translated to CLIF. That would define it as a dialect
> for a subset of CL, and it would inherit the CL model theory.
>
> For the procedural diagram types, the simplest way to specify
> the semantics is by preconditions and postconditions. Those
> can also be specified in CLIF.
>
> Recommendation: The logicians on this list should collaborate
> with the fUML authors to revise the fUML documents by the
> methods outlined above (or something similar).
>
> John
>
> _________________________________________________________________
> To Post: mailto:ontoiop-forum@xxxxxxxxxxxxxxxx
> Message Archives: http://ontolog.cim3.net/forum/ontoiop-forum/
> Config/Unsubscribe: http://ontolog.cim3.net/mailman/listinfo/ontoiop-forum/
> Community Files (open): http://interop.cim3.net/file/pub/OntoIOp/
> Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntoIOp (03)
_________________________________________________________________
To Post: mailto:ontoiop-forum@xxxxxxxxxxxxxxxx
Message Archives: http://ontolog.cim3.net/forum/ontoiop-forum/
Config/Unsubscribe: http://ontolog.cim3.net/mailman/listinfo/ontoiop-forum/
Community Files (open): http://interop.cim3.net/file/pub/OntoIOp/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntoIOp (04)
|