Previous researchers have argued that method purity (absence of side-effects) in object-oriented languages allows code to be more easily reasoned about, with beneficial effects for software security. However, many different forms of purity have been proposed, which poses a challenge for verifying this idea empirically. Furthermore, previous purity analyses of code have provided only coarse-grained results. We propose a finer-grained classification of levels of method purity, and analyse a portion of real-world code to determine the prevalence of different levels.
|Name||ACM International Conference Proceeding Series|
|Conference||2016 Australasian Computer Science Week Multiconference|
|Period||1/02/16 → 5/02/16|