Checking Model Transformation Refinement

Abstract

Esta es la versión de autor de la comunicación de congreso publicada en: This is an author produced version of a paper published in: Abstract. Refinement is a central notion in computer science, meaning that some artefact S can be safely replaced by a refinement R, which preserves S's properties. Having available techniques and tools to check transformation refinement would enable (a) the reasoning on whether a transformation correctly implements some requirements, (b) whether a transformation implementation can be safely replaced by another one (e.g. when migrating from QVT-R to ATL), and (c) bring techniques from stepwise refinement for the engineering of model transformations.In this paper, we propose an automated methodology and tool support to check transformation refinement. Our procedure admits heterogeneous specification (e.g. PaMoMo, Tracts, OCL) and implementation languages (e.g. ATL, QVT), relying on their translation to OCL as a common representation formalism and on the use of model finding tools.

Specification and Verification of Graph-Based Model Transformation Properties

“…(3) Büttner et al [9] expressed properties in OCL and verified them using model finders. PaMoMo [11] is a graphical language used to express con-tracts and complex properties that manipulate contracts.…”

“…(1) Büttner et al [9] and Cabot et al [10] translated a transformation and its metamodels into a transformation model and used model finders (e.g., USE Validator) and constraint solvers (e.g., UMLtoCSP) to verify properties. Anastasakis et al [3] and Baresi and Spoltini [6] translated a transformation into an Alloy model and used the Alloy Analyser to verify the Alloy model within a scope.…”

“…We focus on properties expressed as input-output model relations since they have been highly investigated in the literature, using both textual (e.g., [9,3]) and graphical (e.g., [11,20]) property languages. After a thorough review of studies addressing the same research question, we found several limitations in the state of the art.…”

“…After a thorough review of studies addressing the same research question, we found several limitations in the state of the art. For example, several studies translate either the transformation (e.g., [9,10]) or the property (e.g., [11]) of interest to an intermediate format to facilitate verification, without proving the soundness of the translation. Secondly, other studies propose incomplete verification techniques that do not account for all possible transformation executions (e.g., [9]).…”

“…For example, several studies translate either the transformation (e.g., [9,10]) or the property (e.g., [11]) of interest to an intermediate format to facilitate verification, without proving the soundness of the translation. Secondly, other studies propose incomplete verification techniques that do not account for all possible transformation executions (e.g., [9]). Finally, a large number of studies proposed input-dependent verification techniques [2] (e.g., Henshin [4], AGG [20]) that prove properties for transformations only when run on a specific input.…”