FW: Formalising UML semantics (VOTE)



Date view Thread view Subject view Author view Attachment view

From: Andy Evans (andye@cs.york.ac.uk)
Date: Wed 11 Jun 2003 - 10:03:39 BST


------ Forwarded Message
From: Daniel Jackson <dnj@mit.edu>
Date: Tue, 10 Jun 2003 21:57:28 +0100
To: puml-list@cs.york.ac.uk
Subject: Re: Formalising UML semantics (VOTE)


*****  FOR THE ATTENTION OF THE MAILING LIST OWNER  *****

See X-Diagnostic: headers above for more details

*****  FOR THE ATTENTION OF THE MAILING LIST OWNER  *****

> If you made up your mind to formalise UML semantics and notations,
> which
> formal method would you prefer and why?.  Please all vote.

you need something with relations built in, since they're so
fundamental to data modelling. amongst established formal notations, Z
is the closest, but suffers from inflexibility in the typing of
schemas: you'll find you can't model UML classes as schemas because
they can't be subtyped.

i've found that the best approach is to use plain old first-order logic
with relational operators. for details, see http://alloy.mit.edu.

of course, a more important question is WHY you should formalize UML.
are you doing it just as an academic exercise? to sanction the
complexity of UML? or to move UML to something cleaner and simpler?

/daniel




------ End of Forwarded Message

Date view Thread view Subject view Author view Attachment view