I'm sorry, I think I missed your point. The paper is focused on a language that can do fully in-place FP for a broad class of programs, with the checks done at

I'm sorry, I think I missed your point. The paper is focused on a language that can do fully in-place FP for a broad class of programs, with the checks done at compile-time. I believe jq code is interpreted, so while it might be able to mutate some values in place, that doesn't mean the interpreter wouldn't allocate and release memory.