A metrics-driven inspection framework for model transformations

Maria Fernanda Granda, Otto Parra, Nelly Condori-Fernández

Research output: Contribution to ConferencePaperAcademic

11 Downloads (Pure)


[Context] Model transformations are key elements of Model-driven Engineering. They allow querying, synthesizing and transforming models into other models or code. [Problem] However, as with other software development artefacts, they are not free from anomalies and thus require both verification and validation techniques. [Objective] The objective of this study is to define a semi-automated framework for inspecting the correctness (notions of type and correspondence) of model transformations, by means of detecting and locating anomalies in the transformation rules. [Method] In order to compare the correctness of source and target models, we assume that operational behaviour can be compared by metrics applied on projections from the source model to the target (with deliberate loss of information), which should be preserved by the transformation. [Results] We demonstrate the applicability of our framework for inspecting the correctness of a model-to-model transformation required in a model-driven testing approach. The main result of the study highlights the advantages of metrics for detecting any missing, incorrect or unnecessary transformation rules that have an impact on the correctness of the model transformations. From the research perspective, the feedback produced by the implemented tool will be useful for future research.

Original languageEnglish
Number of pages14
Publication statusPublished - Apr 2019
Event22nd Ibero-American Conference on Software Engineering, CIbSE 2019 - La Habana, Cuba
Duration: 22 Apr 201926 Apr 2019


Conference22nd Ibero-American Conference on Software Engineering, CIbSE 2019
CityLa Habana


  • Correspondence correctness
  • Inspection
  • Metrics
  • Model transformations
  • Type-correctness


Dive into the research topics of 'A metrics-driven inspection framework for model transformations'. Together they form a unique fingerprint.

Cite this