You can't just go with a given size for the CoC.
But I would argue you can just go with CoC - as long as it is based on the variables you mention - ie film size, enlargment factor, and viewing distance (although viewing distance is kind of hard to specify). I'd even through in subject matter as a potential variable. There are likely others since perception is a complex thing. The point was once you have your preferred CoC that is what you need to balance against diffraction. Perhaps diffraction also first needs to be considered using the same variables. So I guess in the end I'm trying to figure out how far we need to stop down for sufficient depth of field, without stopping down too far, at which point any marginal depth of field (CoC) gains are more than offset by the effects of diffraction.