case execution time(WCET) of programs is to be calculated. Normally these
annotations must be given manually by the programmer.
By automating the generation of annotations the user is relieved from a task that
can be quite difficult and highly critical. If, for example, a programmer gives a too
optimistic annotation for a loop, a tool may calculate a too short WCET. This may
invalidate subsequent analysis, leading to that a real-time system may violate its
The method is based on the notion of abstract interpretation, and is illustrated
by showing the analysis of an example in Smalltalk.
This paper also discusses alternatives to the semantic analysis used in the
method. It is shown that syntactic analysis suffices in some cases, and that it can
be used to reduce the calculation cost for loops.