RE: Transformation from UML to ASM



RE: Transformation from UML to ASM

From: Jrn Guy S <jgsuess_at_itee.uq.edu.au>
Date: Thu, 18 May 2006 10:12:35 +1000
Message-ID: <000301c67a0f$c79e8270$4e406682@itee.uq.edu.au>
Dear Tlili,

Your experience wuth UML semantics unfortunately is typical. I also second
Birgit's concern about the lack of QVT tools. 

If you are free in your choice of UML metamodel and instrumentation, I
suggest you use the UML 1.4.2.  The problems of this standard are - at least
- well researched and observable, and UML tools to manipulate instances are
available (MagigDraw, earlier versions of Poseidon UML). In this case your
consequent tool of choice for transformation would be the Dresden
transformation system.

If you are constrained to use UML 2.0, you should expect some difficulty,
due to the sheer size of the specification. To assess the impact of this
obstacle, read Arnor Solberg's paper "Navigating the Meta-Muddle". 
http://www.planetmde.org/wisme-2005/NavigatingTheMetaMuddle.pdf
It was motivated by Arnor's irritation over being side-tracked into digging
through the 1000 pages, rather than getting work done. He also offers an
approach to explore the spec based on views, which you might want to
consider.

For UML 2.0 tooling, I would suggest you use Eclipse EMF as a basis. It is a
simplified version of MOF and stands a decent changes of becoming a de-facto
standard. Also, there is a UML 2.0 sub-project which provides a ready-made
implementation of the meta-model. Finally, Tefkat, a declarative
transformation engine originally developed by the now defunct DSTC is
available as a plug-in to eclipse. Tefkat has very expressive declarative
syntax, and is based on a prolog engine.
http://sourceforge.net/projects/tefkat

If you absolutely have to use software based on standard QVT, the only
implementation I know so far is ModelMorf, by Tata Consultancies in India.
Its development route-map calls for a commercial version to be purchased for
any practical use. So you may want to consider raising some funds first.
http://www.tcs-trddc.com/modelmorf/index.htm

As for the novelty of the approach, your reading list will likely be long.
http://liinwww.ira.uka.de/ lists 215850 documents for the search string
"state machine uml". Also, a considerable number of UML tools have
implemented forward- and reverse-engineering schemes employing state
machines, mostly in the real-time domain. Thus, it would be helpful to work
out how your approach differs from existing ones.

Hope any of this helps you to arrive at a decision.

Jrn Guy S

Research Officer

Room 350, General Purpose South Building (building 78) Division of Systems
and Software Engineering School of Information Technology and Electrical
Engineering The University of Queensland Queensland 4072 AUSTRALIA

Phone: +61 7 3365 2883; Fax: +61 7 3365 4999

email: jgsuess@itee.uq.edu.au

-----Original Message-----
From: puml-list-request@cs.york.ac.uk
[mailto:puml-list-request@cs.york.ac.uk] On Behalf Of Nizar TLILI
Sent: Wednesday, 17 May 2006 21:51
To: puml-list@cs.york.ac.uk
Subject: Transformation from UML to ASM

Dear all,
I am making a model driven transformation from UML behavioral diagrams
to Abstract State Machines. I choosed to write it in QVT. I will use
the metamodel of ASM provided by Patrizia Scandurra.
I have some difficulties at the metamodel of UML, especially in
expressing the semantics of the metamodel and integrating them into my
transformation.
Can anyone help me with this? And do you think this approach is good,
and have it been treated already?
Thanks!

-- 
Nizar TLILI
em@ail       nizar.tlili@gmail.com
tel.      :      +216 98 997 823
adress :      45, rue Mahmoud Timour, cit des Orangers, Sidi Thabet
2020, Ariana, Tunisia.



To remove yourself from this list please mail
puml-list-request@cs.york.ac.uk
with a message containing the word "unsubscribe".
Received on Thu 18 May 2006 - 01:12:43 BST