Macro processing is a strategy for platform-dependent code generation from a higher-level algorithm, with a venerable history [1]. Kim and his co-authors have reinvented a pale cousin of this technique to generate embedded processor code from an architectural description of an infusion pump.

They call their macro bodies “code snippets,” and expand them with a program that appears to know about specific aspects of their problem. There is no reference to the extensive literature on macro processing, and no indication of why the original, more general approach would not have been satisfactory in their application.

I view this paper as an unfortunate example of our failure to appreciate the past in our development of software: we tend to get caught up with the new, and regard the old as irrelevant.