Speaker:Patrick Cousot
Time:2000-06-14 16:20:00
Place:Rm.1220, EE Bldg./KAIST


The 25 last years, computer hardware has seen its performances multiplied by 10^4 to 10^6 and the size of software has evolved in similar proportions. It follows that the main preoccupation of computer scientists is now (or should be) the production of reliable software, its maintenance and safe evolution year after year. Despite software engineering progress, bugs still too are numerous and can have catastrophic consequences, either very costly or inadmissible (for example for software embedded in transportation systems). In the future computer scientists might be liable for such bugs, which is not yet the case today. It is consequently absolutely necessary to widen the full set of methods and tools used to fight againts software bugs.

In this talk we rapidly overview the application of ``Abstract Interpretation'' to ``Program Static Analysis''. Program Static Analysis aims at automatically extraction of program properties by analysis of the program text without and prior to any execution. Such properties are formal descriptions of the program behavior which are therefore valid for all possible executions of the program. Abstract Interpretation formalizes program analyzers as computable approximations of the formal semantics of programming languages, with some controlled loss of information. These notions are introduced informally and extensively illustrated with historical examples and work done at the 'Ecole Normale Sup'erieure these last 10 years. Recent industrialization, indispensable future work and prospectives are quickly surveyed.


[ List ]