Modeling the Effect of Size on Defect Proneness for Open-Source Software

Software engineering is a decision intensive discipline. Do we really understand all the factors that can influence those decisions? Can we build models that reveal hidden patterns in software resource management, in development processes, or in software artifacts themselves? How well do these models predict? Can they be used without requiring domain expert intervention? Do the models lead to better decisions? How are we to validate these models? Is the model creation process repeatable? Are there better, faster, cheaper ways to build models? How effective are these models for identifying causal relations? The PROMISE workshop seeks to address these questions and others, and to deliver to the software engineering community useful, usable, verifiable models, and public datasets for building and evaluating new models. At present, the Promise repository contains more than 30 datasets.