Examining the Relationship between FindBugs Warnings and End User Ratings A Case Study On 10000 Android Apps

Authors -

Hammad, Khalid; Meiyappan, Nagappan and Hassan, Ahmed E.

Venue -

IEEE Software, Accepted Nov 2014

Related Tags -

• IEEE SOFTWARE
• Mobile Apps
• ESE in Ultra Large Repositories
• 2014

Abstract -

In the mobile app ecosystem, end user ratings of apps (a measure of end user perception) are extremely important to study as they are highly correlated with downloads and hence revenues. In this study we examine the relationship between the app ratings (and associated review-comments) from end users with the static analysis warnings (collected using FindBugs) from 10,000 free-todownload Android apps. In our case study, we find that specific categories of FindBugs warnings such as the Bad Practice, Internationalization, and Performance categories are found significantly more in low-rated apps. We also find that there exists a correspondence between these three categories of warnings and the complaints in the review-comments of end users. These findings provide evidence that certain categories of warnings from FindBugs have a strong relationship with the rating of an app and hence are closely related to the user experience. Thus app developers can use static analysis tools such as FindBugs to potentially identify the culprit bugs behind the issues that users complain about, before they release the app.

Preprint -

PDF

BibTex -

@article{Khalid2015,
 author = {Hammad, Khalid and Meiyappan, Nagappan and Ahmed E. Hassan},
 keyword = {Mobile Apps, ESE in Ultra Large Repositories},
 title = {Examining the Relationship between FindBugs Warnings and End User Ratings A Case Study On 10000 Android Apps},
 type = {journal},
 venue = {IEEE Software, Accepted Nov 2014}
}