Reducing Behavioural to Structural Properties of Programs with Procedures

Formal Methods for Components and Objects

2011

Self Cite

Abstract. Software product line engineering allows large software systems to be developed and adapted for varying customer needs. The products of a software product line can be described by means of a hierarchical variability model specifying the commonalities and variabilities between the artifacts of the individual products. The number of products generated by a hierarchical model is exponential in its size, which poses a serious challenge to software product line analysis and verification. For an analysis technique to scale, the effort has to be linear in the size of the model rather than linear in the number of products it generates. Hence, efficient product line verification is only possible if compositional verification techniques are applied that allow the analysis of products to be relativized on the properties of their variation points. In this paper, we propose simple hierarchical variability models (SHVM) with explicit variation points as a novel way to describe a set of products consisting of sets of methods. SHVMs provide a trade-off between expressiveness and a clean and simple model suitable for compositional verification. We generalize a previously developed compositional technique and tool set for the automatic verification of control-flow based temporal safety properties to product lines defined by SHVMs, and prove soundness of the generalization. The desired property relativization is achieved by introducing variation point specifications. We evaluate the proposed technique on a number of test cases.

Section: −−−−−−−→(V 3 ε)mentioning

confidence: 99%

Compositional Algorithmic Verification of Software Product Lines

Schaefer

Formal Methods for Components and Objects

2011

Self Cite

“…As already pointed out, the present work is based on a previously developed method and tool set for compositional verification of control flow safety properties [9,8]. Essentially, we provide a wrapper around the tool set to support the fully automated procedure-modular verification of programs annotated with temporal correctness assertions.…”

Section: Related Workmentioning

confidence: 99%

Procedure-modular verification of control flow safety properties

Proceedings of the 12th Workshop on Formal Techniques for Java-Like Programs

Huisman

2010

Self Cite

This paper describes a novel technique for fully automated procedure-modular verification of Java programs equipped with method-local and global assertions that specify safety properties of sequences of method invocations. Modularity of verification is achieved by relativizing the correctness of global properties on the local properties rather than on the implementations of methods, and is based on the construction of maximal models. Tool support is provided by means of ProMoVer, a tool that is essentially a wrapper around a previously developed tool set for compositional verification of control flow safety properties, where program data is abstracted away completely. We evaluate the technique on a small but realistic case study.

“…Thus, the maximal flow graph can be used as a representative of the class for the purpose of checking properties. Using maximal models for compositional verification was first proposed by Grumberg and Long [17] for finite-state systems, and was generalized for flow graphs by Gurov and others in [19,18].…”

Section: Definition 11 (Model)mentioning

confidence: 99%

Model Mining and Efficient Verification of Software Product Lines

Østvold

et al. 2015

SJC

Software product line modeling aims at capturing a set of software productsin an economic yet meaningful way. We introduce a class of variability modelsthat capture the sharing between the software artifacts forming the productsof a software product line (SPL) in a hierarchical fashion, in terms of commonalitiesand orthogonalities. Such models are useful when analyzing and verifying all productsof an SPL, since they provide a scheme for divide-and-conquer-style decompositionof the analysis or verification problem at hand. We define an abstract class of SPLsfor which variability models can be constructed that are optimal w.r.t. the chosen representation of sharing. We show how the constructed models can be fed into a previously developedalgorithmic technique for compositional verification of control-flow temporal safetyproperties, so that the properties to be verified are iteratively decomposed intosimpler ones over orthogonal parts of the SPL, and are not re-verified over the shared parts. We provide tool support for our technique, and evaluate our tool on a small but realistic SPL of cash desks.