![]() |
|||||
Thomas Baar |
|||||
|
|||
|
|
|||
| Senior researcher and lecturer in Computer Science, Swiss Federal Institute
of Technology in Lausanne (EPFL), 2003 -- 2007.
|
NewsThe Software Engineering Lab was closed on Sept. 30, 2007. I moved to Tech@Spree in Berlin, Germany. You can reach me now via email by using my new address: thomas.baar 'at' acm.org
MissionWhat is the purpose of software modeling? To entertain the management with funny pictures? To double the size of project teams? Or to have just another discipline lecturers can give courses on?
For some software developers it might be tempting to answer these rhetoric questions in the affirmative. Software modeling, however, has much more to offer. It gives developers the opportunity to describe a new (or an existing) system under a certain viewpoint, i.e. on
a well-defined level of abstraction. The abstraction emphasizes only very few of the system's
properties and structures whereas all others details are intentionally ignored. A model is thus better understandable and analyzable.
Once a model of a system has been completed, numerous questions such as
'Does the implemented system really behave the way the model says?' or
'How can one be sure that the developed model is implementable, otherwise
stated, how can one easily detect inconsistencies within the model?' arise.
So far, there is rather limited tool support for the software developer to
answer these questions. Ever worse, as long as there is no good tool support for showing the
conformance relationship between the model and the implementation of a system and as long as we do not have
tool support for the analysis of real-word models, software developers won't
be able to take full advantage of modeling techniques.
My group is currently developing the RoclET tool, an analysis tool for UML/OCL specifications. The ultimate goal is to give the user early feedback on semantical inconsistencies within OCL specifications. Furthermore, the tool supports model evolution, e.g. if the underlying class diagram is changed then the annotated OCL constraints are adapted on the new class diagram automatically.
Short VitaI graduated in Computer Science
from Humboldt-University Berlin in 1997. Until 1999 I worked
as a PhD student at the Institute of Pure Mathematics at Humboldt-University and was involved
in the ILF project aiming at developing
a platform to integrate several theorem provers based on different
logical calculi. From 1999 until 2003 I worked in the KeY project whose aim is the
integration of deduction-based verification techniques into industrial
software development. In July 2002 I defended my PhD thesis entitled Über die Semantikbeschreibung
OCL-artiger Sprachen (On the
semantic description for OCL-like languages). Since May 2003 I am
a Post-Doc researcher at EPFL.
In 2004 I was nominated as a lecturer by our faculty. Since October
2004, I head together with Jarle Hulaas the Software Engineering
Laboratory.
Teaching at EPFL Here is the list of
courses I gave in Lausanne so far. The main topic of the courses is
Software Engineering with an emphasize on formal methods.
Supervision of PhD studentsI'm very proud to have the opportunity to work with my two PhD students Slavisa Markovic and Frédéric Fondement.
Research interests
Publication
Events (involved as organizer, PC member, PC-chair, etc.)
|
|
| EPFL | IC | LGL Last modified |
||