Description
Many engineers’ first encounter with a static type-checker is being told off by it: you roll your eyes and mypy shouts ERROR. Really, why invest your precious time in mypy at all? This is my conversion story, from disgruntled engineer to mypy believer. I’ll share with you the good gospel of its benefits, strategies to increase buy-in from your team and tips for how to add mypy into your workflow.
In a dynamically typed language like Python, it is worth asking, “Why use mypy and invest precious engineering time to add and check types?” My time as an engineer at Nylas has seen an engineering org-wide conversion to mypy, for good reason. I'll share how we were able to gradually increase type coverage in our entire codebase from 0% to 80%.
Warming to the change took effort, but the substantial benefits to the quality and sustainability of our codebase have made me into a mypy believer. Even so, becoming a believer is only the first step. Next, you have to inspire your fellow engineers to join your mypy quest. How can you convince your team to want to put in this extra work? I'll present strategies for expanding buy-in from your fellow engineers and processes to ensure that your type coverage does not regress. My story may just make believers out of you and your team.