Anvik and Murphy focus on bug reports, which are traditionally underutilized by developers, as a tool to increase quality and decrease risks in software development. The authors’ bug report analysis tool--a recommender--shows definite promise in reducing bug report categorization and pre-grouping costs while achieving predictive results. However, additional work is needed to make the three presented recommenders valuable enough for practical use.

It is an international testing best practice to analyze bug reports for corrective development action. The authors demonstrate their recommender tool to be a more cost-effective bug report analysis approach than manual review.

In order to establish credibility for the paper’s recommender value premise and proof of cost reduction for a nontester professional audience, the authors must provide additional information. The paper’s value premise about cost reduction is stated numerous times; however, no benefits are presented to offset the recommender costs--the real value for practical use. In addition, I could not find a list of the Bugzilla bug report fields on which the authors based their recommender case study analysis.