When I first learned OCaml (knowing only Python), I swore a lot at the type checker for refusing to compile code that I knew would work at run-time (e.g., a variable having multiple types, but on strictly disjoint execution paths). It seemed insane to me that people would want to subject themselves to this kind of bondage and discipline. And yet, many years later now, I have learned and internalized how type systems work, I have learned not to try and simply replicate Python patterns in OCaml, and as a result I use OCaml to write code quickly and without any headaches; and not only is the type checker no longer an obstacle, it's now a valuable assistant.

I believe that we are seeing the same kind of phenomenon in Rust: we are trying to replicate C or C++ code directly in Rust, and we get frustrated when our efforts are foiled by the borrow checker. It's going to take some time and some efforts before we learn and internalize how borrowing works in Rust and how to change the way we write programs so that it is no longer an obstacle. The Rust team looks very receptive to comments about taking away some pain points (e.g., non-lexical lifetimes), but we also need to accept that, for a little while, we are going to feel like we did when we were new programmers and were trying to make sense of these new constraints.