Why attempts to develop a thermodynamics of computation inevitably need to introduce ideas that can only be imported, in a question-begging way, from statistical mechanics, and have little or no justification in phenomenological (and fluctuation-free) thermodynamics.
University of Sydney