You can't really tell though. Maybe the dict is frozen but the values inside aren't. C++ tried to handle this with constness, but that has its own caveats that make some people argue against using it.

Indeed. So I don't really understand what this proposal tries to achieve. It even explicitly says that dict → frozendict will be O(n) shallow-copy, and the contention is only about O(n) part. So… yeah, I'm sure they are useful for some cases, but as Raymond has said — it doesn't seem to be especially useful, and I don't understand what people ITT are getting excited about.