Requiring compression to be isentropic is just requiring "compression" to be a well-defined term with a specific quantitative result. Otherwise, it's a vague term that offers no quantitative prediction for how much heating it produces for a given change in density. That's not science.