On finding model smells based on code smells

Abstract

A smell in an artifact is a sign that the artifact may have a technical debt, meaning that it may contain the results of one or more sub-optimal design decisions. The debt makes it more difficult to understand, maintain, extend, and reuse the artifact. Technical debt could appear in any technical artifact. Currently the debt in source code is in the center of attention. Modeling is an important system development activity. Models can have technical debt and a sign of it is the presence of one or more model smells. This paper introduces a catalog of 46 model smells (https://github.com/erki77/model-smells ) that has been created based on a set of code smells. The cataloged smells are general but their examples are given based on system analysis models. The main takeaway is that most of the considered code smells are instances of generic problems that can also occur in models.