From: "Gianna Reggio" <reggio@disi.unige.it>
To: "Andy Evans" <andye@cs.york.ac.uk>; "Suart kent" <S.J.H.Kent@ukc.ac.uk>
Cc: "Tony Clark" <a.n.clark@scm.brad.ac.uk>; "Steve Brodsky"
<sbrodsky@us.ibm.com>; "Steve
Cook" <sj_cook@uk.ibm.com>
Sent: Monday, November 20, 2000 3:11 PM
Subject: your MML proposal/towards dynamic core
Dear Andy and Stuart,
after reading your proposal ("A Feasibility Study in Rearchitecting
UML as a Family of Languages using a Precise OO Meta-Modeling Approach")
and the very recent draft paper by Perdita on Use Cases, we have immediately
realized that, out of our previous work on UML state machines and active
classes, we could easily devise, in your terminology, a dynamic.core package
for MML, based on labelled transition systems, which could fit quite well
into your schema and could be extended/specialized to define the dynamic
features of the UML families; indeed also Perdita feels and states that
an lts approach could be the right solution.
In developing such proposal
A. We came out with some remarks and some questions about the terminology you use, which is puzzling us (and, we believe, most people of the "classical semantics" community).
B. Moreover we discovered that some things would come out in a cleaner way, if you make some minor modifications of your static.core proposal.
So, before sending you and making public our proposal for the dynamic
core, we would like to discuss with you the above points, which you find
in the attached document (Postscript/PDF).
Best regards,
Egidio Astesiano and Gianna Reggio