Language Feature Studies

Programming languages are often designed by people by utilizing their existing expertise and experience. New languages often add features based on features seen in other (often popular) programming languages. I believe that programming language design can be more data-driven.

To that end, we utilize mining software repositories techniques to analyze hundreds of thousands of open-source projects so we can get a better handle on how developers actually use programming language features.

Such data can then be leveraged to help provide more data-driven inspirations for new programming language designs, or to help provide concrete evidence for the support for features that are assumed to be popular in existing languages.