|
CREDO project successfully concluded |
|
The CREDO project officially ended in December 2009 and underwent its final review in Brussels last March. However, Almende officiously celebrated the conclusion of the project this month with a Research Meeting hosted by Andries Stam. During this third internal meeting dedicated to CREDO, Stam presented some of the major achievements of this European collaborative project.
The original aim of the CREDO project was to develop a modeling and validation framework for evolving software systems. The project made use of three different modeling languages:
-
CREOL - an object-oriented language developed at the University of Oslo and CWI, used for modeling the behavior of components.
- REO – oriented on dataflow, developed at CWI and the Technical University of Dresden, used for modeling the interaction between components.
- Timed Automata, developed at Uppsala University and CWI.
Two case studies were defined for the project. The Scandinavian branch undertook the analysis of networked biosensors in the Norwegian Rikshospitalet, while Almende modeled the ASK communication platform.
“Our main challenge was learning which level of abstraction to use to deliver meaningful results,” Stam explains. “It’s a bit like those Russian matryoshka dolls. With these modeling languages you can begin at a very abstract level, defining the main processes of the ASK platform, and then zoom in on every single process to gradually refine the model.”
Making use of multiple levels of abstraction and the principle of compositionality, allows for analyzing and validating complex evolving systems. Stam: “We now have a very precise model of a large part of the ASK platform. We can analyze the current system, but it is also easy to input changes and test the outcome. The CREDO project delivered a very extensive and handy tool suite for the modeling, verification and analysis of such evolving systems.”
Besides these more important results, the project can also boast a record result: the project partners wrote the largest meaningful CREOL- and REO-models to date.
Being so pleased with the results of the project, Almende is already taking the first steps towards acquiring several follow-up projects. Stam: “We would like to start at least two more projects, one dedicated entirely to REO and one to CREOL. We found that REO is perfect for modeling and verifying change within a system. We would like to use CREOL to generate code for applications running on multi-core processors.” Links:
|
