Formal Methods in Programming and Their Applications

Volume 735 of the series Lecture Notes in Computer Science pp 98-127


Galois connection based abstract interpretations for strictness analysis

  • Patrick CousotAffiliated withLIENS, École Normale Supérieure
  • , Radhia CousotAffiliated withLIX, École Polytechnique

* Final gross prices may vary according to local VAT.

Get Access


The abstract interpretation framework based upon the approximation of a fixpoint collecting semantics using Galois connections and widening/narrowing operators on complete lattices [CC77a, CC79b] has been considered difficult to apply to Mycroft's strictness analysis [Myc80, Myc81] for which denotational semantics was though to be more adequate (because non-termination has to be taken into account), see e.g. [AH87], page 25.

Considering a non-deterministic first-order language, we show, contrary to expectation, and using the classical Galois connection-based framework, that Mycroft strictness analysis algorithm is the abstract interpretation of a relational semantics (a big-steps operational semantics including non-termination which can be defined in GSOS either in rule-based or fixpoint style by induction on the syntax of programs [CC92])

An improved version of Johnsson's algorithm [Joh81] is obtained by a subsequent dependence-free abstraction of Mycroft's dependence-sensitive method.

Finally, a compromise between the precision of dependence-sensitive algorithms and the efficiency of dependence-free algorithms is suggested using widening operators.


Abstract interpretation Relational semantics Strictness analysis Galois connection Dependence-free and dependence-sensitive analysis Widening