Currently, it is not mentioned that the random part of the trace ID (when the random flag is set) should be uniformly distributed. Compare https://w3c.github.io/trace-context/#trace-id, https://w3c.github.io/trace-context/#random-trace-id-flag, and https://w3c.github.io/trace-context/#randomness-of-trace-id. To be useful for sampling, it should be explicitly stated that the random part should follow a uniform distribution, meaning that all possible 2^56 values are equally likely. Otherwise, if we cannot rely on the uniform distribution, we would again need to hash the trace ID, which is exactly what we want to avoid with the random part.
Currently, it is not mentioned that the random part of the trace ID (when the random flag is set) should be uniformly distributed. Compare https://w3c.github.io/trace-context/#trace-id, https://w3c.github.io/trace-context/#random-trace-id-flag, and https://w3c.github.io/trace-context/#randomness-of-trace-id. To be useful for sampling, it should be explicitly stated that the random part should follow a uniform distribution, meaning that all possible 2^56 values are equally likely. Otherwise, if we cannot rely on the uniform distribution, we would again need to hash the trace ID, which is exactly what we want to avoid with the random part.
FYI @PeterF778 @jmacd