Analysing the Evolution of Exception Handling Anti-Patterns in Large-Scale Projects

Previous studies have shown that exception handling bad practices may impact the overall software quality. We believe that quality of exception handling code is directly affected by (i) an absence, or lack of awareness, of an explicit exception handling policy; and (ii) a silent rising and spreading of exception handling anti-patterns. To investigate such phenomenon, we conducted a case study in a large-scale Java Web system, trying to better understand the relationship between (i) and (ii). The study takes into account technical and human aspects. We surveyed 21 developers regarding their perception about exception handling in the system's institution. Next, we analyse the evolution of exception handling anti-patterns across 15 releases of the target system. Finally, we conduct a semi-structured interview with three senior software architects. Our finds beneficiated the system's institution by making it aware of these problems and enabling it to take actions towards to combat them.