Assessing the use of slicing-based visualizing techniques on the understanding of large metamodels

Abstract

Context. Metamodels are cornerstones of various metamodeling activities. Such activities consist of, for instance, transforming models into code or comparing metamodels. These activities thus require a good understanding of a metamodel and/or its parts. Current metamodel editing tools are based on standard interactive visualization features, such as physical zooms. Objective. However, as soon as metamodels become large, navigating through large metamodels becomes a tedious task that hinders their understanding. So, a real need to support metamodel comprehension appears. Method. In this work we promote the use of model slicing techniques to build interactive visualization tools for metamodels. Model slicing is a model comprehension technique inspired by program slicing. We show how the use of Kompren, a domain-specific language for defining model slicers, can ease the development of such interactive visualization features. Results. We specifically make four main contributions. First, the proposed interactive visualization techniques permit users to focus on metamodel elements of interest, which aims at improving the understandability. Second, these proposed techniques are developed based on model slicing, a model comprehension technique that involves extracting a subset of model elements of interest. Third, we develop a metamodel visualizer, called Explen, embedding the proposed interactive visualization techniques. Fourth, we conducted experiments. showing that Explen significantly outperforms EcoreTools, in terms of time, correctness, and navigation effort, on metamodeling tasks. Conclusion. The results of the experiments, in favor of Explen, show that improving metamodel understanding can be done using slicing-based interactive navigation features.

An Automated Approach for Classifying Reverse-Engineered and Forward-Engineered UML Class Diagrams

“…Blouin et al [212] proposed an interactive visualisation tool for comprehending large meta-models. Their tool, Explen, used a model slicing technique to allow users focus on subset of model elements of interest.…”

“…It can help developers maintain and further develop the system. In particular, it can be utilized to improve the search, navigation, and exploration of software architecture design [211] [212].…”

Automatic production of end user documentation for DSLs

“…LD3 also explained that the approach may be useful for DSLs having a large metamodels. LD3, however, considered that designing a large metamodel for a DSL (e.g., UML, around 1.000 model elements [31]) in a bad practice since a DSL should remain small. RQ1 shows benefits regarding the correctness of the created ThingML models when using the approach while ThingML contains 55 model elements, much less than UML.…”

“…This is a usability issue of the UML tools. Slicing techniques have been proposed to ease the understanding of large UML models [31]. Such techniques could be used to apply the approach on small separate parts of the metamodel rather than on the whole metamodel in one shot.…”

Mastering Model Driven Engineering complexity by interactive visualization

“…The visualization approach also proposes good opportunities to support model understanding by providing adequate features to navigate inside an ecosystem of models. Some visualization tools were designed to visualize some model complexity aspects such as (Bihanic, Polacsek, 2012), UML tree-map visualization (Garcia et al, 2010), "Moose" (Ducasse et al, 2000), "Imsovision" (Maletic et al, 2001) and "Explen" (Blouin, Moha et al, 2015). The visualization techniques used are based on graphs.…”