Rust has succeeded in breaking ownership out of academia. Yet, while there is a thriving community of Rust programmers, mounting evidence shows that Rust’s ownership model still poses a serious barrier to adoption both for new and experienced developers. This is an instance of a more general problem: when you invent a fancy type system, how are you supposed to teach it to people? In this talk, I will describe our method of improving the pedagogy of Rust’s type system. Specifically, we (a) investigated why developers struggle with ownership, (b) developed a new way to visualize how Rust’s ownership analyzer “sees” a program, and (c) integrated this visualization into a new pedagogy of ownership that can explain phenomena like soundness vs. completeness and field-sensitivity to the typical Rust learner. We evaluated this approach through a large-scale field deployment into a popular online Rust textbook, finding that the pedagogy improved learning outcomes by roughly one letter grade on an exam of difficult ownership-related questions.

Will Crichton is an assistant professor of computer science at Brown University. His research combines programming language theory and cognitive psychology to develop a science of the human factors of programming. Will creates both tools and methodologies to help developers learn and use our most sophisticated programming languages. His research systems for learning Rust have been used by over 100,000 developers to date. Will received his PhD from Stanford University in 2022 advised by Pat Hanrahan and Maneesh Agrawala, and he was a postdoc with Shriram Krishnamurthi before starting as a professor at Brown.