MCTest: towards an improvement of match algorithms for models

Abstract

Due to the increasing importance of Model-Driven Engineering and the changes experienced by software systems over their life cycle, the calculation, representation, and visualization of matches and differences between two different versions of the same model are becoming more necessary and useful. This work shows the need for improvement in the algorithms for calculating the relationships between models and presents a tool to test different implementations, thus reducing the effort required to measure, compare, or create new algorithms. To demonstrate the need for improvement and the framework developed, we have created different models that conform to the metamodel of a domain-specific language. Subsequently, we compared these models using the algorithms of the EMF Compare tool, part of the Eclipse Modeling Project, which is the framework of reference for Model-Driven Engineering. Thus, in the case study, our tool is used to measure the quality of the comparisons performed by EMF Compare.Keywords testing tools, reuse models, domain engineering, match algorithms Grammar of the domain-specific languageCreated models normally undergo evolution over time, which leads to new versions of those models. This is

Systematic review of matching techniques used in model-driven methodologies

“…We identified frameworks whose purpose is the evaluation of model matching/differencing/merging algorithms [62,152]. MCTest [62] evaluates Eclipse-based model matching algorithms. The framework is implemented in Java, highly configurable, and homogenizes results for every approach.…”

“…To sum it up, the two main problems we found regarding evaluation during our survey are (i) the technology dependence of benchmark data sets and evaluation tools (which are mostly Eclipse-based), and (ii) the manual verification of the result of the matching algorithm compared with the optimal result [44,50,62,152,161]. The latter is especially problematic in the case of large-scale models, for which we found a lack of solutions regarding evaluation.…”

“…-Lack of empirical studies Empirical user studies are rare, although they can be used to verify certain aspects of an algorithm (e.g., human effort required to use) that would otherwise be difficult to quantify. -Lack of technology-independent evaluation frameworks There are some existing evaluation frameworks for model matching algorithms, but they are usually technology dependent [62,152].…”

MDCI: Model-driven continuous integration

“…To meet the objectives, a prototype which saves all the current constraints that do not allow the union between these two recent tools of software engineering was build. Some of the main problems found were related to the selection of an appropriate development initiative [2], the quality of transformation models [3] or the correction in the detection of new models representing systems [4]. The separation of work in different blocks can provide solutions resulting in an iterative and incremental work from beginning to end.…”

Towards a Standard-based Domain-specific Platform to Solve Machine Learning-based Problems

“…There are a variety of research in MDE that serve to advance in the systematic use of DSLs. For example, Cueva et al work on bringing together the MDE approach and the Internet of Things field creating languages for automatic vehicle data capture [14] [15] or García.-Díaz et al work on improvement match algorithms for performing further operations with models [16].…”