Determining Thresholds for Detecting Enterprise Architecture Smells

Enterprise Architecture Smells (EA Smells) indicate possible weaknesses in an EA structure. Various EA metrics have been proposed as tools to detect EA smells, and the effectiveness of the detection solely lies on the applied metric thresholds, which determine whether the detected EA smell indicates a valid threat to EA qualities. Applying undervalued/overvalued thresholds will result in false-negatives/false-positives which significantly hinder the EA improvement efforts. A plethora of approaches to determining appropriate thresholds have been proposed in the field of Code Smells; however, since the research of EA Smells is still very young, no such an approach has been established for EA Smells. Therefore, this thesis topic aims to investigate the applicability of related approaches in the context of EA and to identify the necessary adaptations.

This topic requires knowledge about software/enterprise architecture, software quality, and object-oriented programming as well as programming skills to develop an application prototype.



  1. Thesis on the catalog of EA smells
  2. Introduction to code smell
  3. Catalog of code smells
  4. Related work on metric thresholds derivation studies