API Deprecation: A Retrospective Analysis and Method for Detecting Deprecated API Usages in Code Examples on the Web
Abstract
Application Programming Interfaces (APIs) can get deprecated over time, to indicate that they are obsolete, dangerous, and/or to be eliminated in future versions. This thesis ad- dresses two problems involving API deprecation.
API deprecation has not been systematically studied so far. We empirically study how API deprecation has been used in 26 open source Java systems. We find that API dep- recation seems to be underused and not enough information is provided to help users via deprecation messages. More than 40% of all deprecated APIs were removed, most of which were removed during transitions to major releases.
A developer may waste valuable time consulting code examples on the web containing deprecated API usages. A version sensitive framework was proposed to detect deprecated API usages in code examples on the web and a prototype tool based on this framework was implemented. Our evaluation of this tool shows that it works fairly well.
Description
Keywords
Computer Science
Citation
Zhou, J. (2016). API Deprecation: A Retrospective Analysis and Method for Detecting Deprecated API Usages in Code Examples on the Web (Master's thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca. doi:10.11575/PRISM/24951