Victoria大学(カナダ)のDaniel German先生に"From Snippets to Identifiers"というタイトルでご講演いただきました。
From Snippets to Identifiers
Daniel M. German (the University of Victoria, Canada)
In this presentation I will describe recent work we have been doing in two different areas: The first is a qualitative empirical study of Stack Overflow idioms, with the goal to try to understand how the snippets that answer a programming problem are found within the answer of a question. Stack Overflow answers usually contain much more than just the actual source code that answers the question, such as examples of how to use the solution snippet, counterexamples, placeholders for code that is to be written, etc. The second is a quantitative empirical study of global identifiers in the PyPI ecosystem. In this study of 240k different PyPI products, we found 11 million different global identifiers (classes, method of functions). 76% of these identifiers were unique to one product, and 93% appeared in at most 3. This discovery allowed us to create an efficient method to identify the origin software product (within PyPI) of a file (or set of files).