On choosing abstractions, and the consequences of being expert friendly in DSLs: https://lexi-lambda.github.io/blog/2016/08/11/climbing-the-infinite-ladder-of-abstraction/