I'm not an analytic number theorist, but my understanding is that the proof requires some values to be 0.5 + some positive epsilon to be correct, and for the sake of clarity, these somewhat arbitrary constants are chosen to get to the result.