I’ve always wondered why Rust creates a unique id per atom and keeps track of type information in a hashmap using that unique id as key (instead of storing the

I’ve always wondered why Rust creates a unique id per atom and keeps track of type information in a hashmap using that unique id as key (instead of storing the type information within the atoms themselves). I guess it is to keep things flat, or reuse the same structure for different passes. Other approaches like the nanopass approach creates different structs for each pass: https://legacy.cs.indiana.edu/~dyb/pubs/nano-jfp.pdf https://legacy.cs.indiana.edu/~dyb/pubs/nano-jfp.pdf