Microsoft Research334 тыс
Опубликовано 17 августа 2016, 21:32
More and more developers rely on documents and code examples on the Web. However, as such documents often do not include sufficient enough information to fully understand the API usages, developers seek for usage examples through multiple search sessions. In this talk, I will first present how we automatically 'mash-up' API documents with usage examples extracted from large-scale open-source code repositories and discuss the quality of generated documents. I will then discuss how to mash-up usage example search directly into developers' editing sessions. As existing clone detection tools are not designed for instant detection scenarios, I will discuss our proposed techniques achieving 10-fold speed-up and enabling instant feedbacks from large-scale real-life repositories. This is joint work with HKUST and will be presented at AAAI 2010 and FSE 2010.
Свежие видео