Antipatterns in software classification taxonomies

Empirical results in software engineering have long started to show that findings are unlikely be applicable all systems, or any domain: need evaluated specified contexts, and limited the type of systems they were extracted from. This is a known issue, requires establishment classification types. paper makes two contributions: first evaluate quality current classifications landscape. The second perform case study showing how create types using curated set systems. Our contributions existing, very likely even new, attempts deemed fail for one more issues, we named as ‘antipatterns’ tasks. We collected 7 these antipatterns emerge from both our study, existing classifications. These represent recurring issues classification, so discuss practical ways help researchers avoid pitfalls. It becomes clear must also face daunting task formulating taxonomy types, with objective establishing hierarchy categories classification.