Final Report Summary - STATOR (STATic analysis with ORiginal methods)
The STATOR project focused on the static analysis of software, that is, proving properties on software before running it. This is a very hard task: it is known since Turing that automated static analysis of programs is in general impossible.
We can however provide automated analyses that help in many cases; for instance we may prove automatically that certain bad conditions never occur, or that the worst case execution time of the program is less than a certain bound. In order to do that, we need to derive invariant properties (properties that can be shown by induction to hold at any point during any execution), on the program variables or on the state of the underlying architecture.
STATOR focused on bringing the state of the art of invariant inference forward.
It also contributed methods for formally proving that the implementation of an analyzer is correct (how good is it to check for the absence of bugs in software if the checker is buggy?).
STATOR contributed several new algorithms and abstractions, most notably for solving Horn clauses, computing over convex polyhedra, abstracting arrays, sets and maps, and analyzing cache properties.
We can however provide automated analyses that help in many cases; for instance we may prove automatically that certain bad conditions never occur, or that the worst case execution time of the program is less than a certain bound. In order to do that, we need to derive invariant properties (properties that can be shown by induction to hold at any point during any execution), on the program variables or on the state of the underlying architecture.
STATOR focused on bringing the state of the art of invariant inference forward.
It also contributed methods for formally proving that the implementation of an analyzer is correct (how good is it to check for the absence of bugs in software if the checker is buggy?).
STATOR contributed several new algorithms and abstractions, most notably for solving Horn clauses, computing over convex polyhedra, abstracting arrays, sets and maps, and analyzing cache properties.