Detection and Handling of Model Smells for MATLAB/Simulink models

Code smells in traditional software artifacts are common symptoms of the violation of fundamental design principles which negatively impact the quality of the resulting software product. Symptoms of code smells commonly occur in traditional software artifacts and cannot be directly mapped to model-based software artifacts. In this paper, we present a catalog for the detection and handling of model smells for MATLAB/Simulink, a widely used tool for model-based software development in the automotive domain. These model smells describe antipattern against universal quality requirements and have been collected in cooperation with an OEM from the automotive domain. To show the feasibility of detecting the proposed model smells, we realized a model smell detector and evaluated it within an industrial case study.

• Type: Journal
• Year: 2015
• Where: MODELS
• Authors: Thomas Gerlitz, Quang Minh Tran, Christian Dziobek
• Keywords: MODELS 2015, Natural Language Processing