Programming with "Big Code" (Dagstuhl Seminar 15472)

This report documents the program and the outcomes of Dagstuhl Seminar 15472 "Programming with "Big Code"". "Big Code" is a term used to refer to the increasing availability of the millions of programs found in open source repositories such as GitHub, BitBucket, and others. With this availability, an opportunity appears in developing new kinds of statistical programming tools that learn and leverage the effort that went into building, debugging and testing the programs in "Big Code" in order to solve various important and interesting programming challenges. Developing such statistical tools however requires deep expertise across multiple areas of computer science including machine learning, natural language processing, programming languages and software engineering. Because of its highly inter-disciplinary nature, the seminar involved top experts from these fields who have worked on or are interested in the area. The seminar was successful in familiarizing the participants with recent developments in the area, bringing new understanding to different communities and outlining future research directions.