I am doubtful about this, and perhaps I should explain why. I think the common assumption about DNA, that is somehow runs the show like a computer CPU is terribly distracting. From one point of view, a body is just a conglomeration of carnot cycles that are working desperately to restore themselves to equilibrium, and they do this through a negative feedback op amp with the DNA's ability to produce RNA as the gate of the feedback loop. DNA has inhibitors attached to it that prevent expression by clamping down on the DNA in one end, but can premit expression when the inhibitors catch, or--fail to catch, depending on their particular design--the molecules they were designed for on the other end.
All these cycles operate more or less independently, without a central master control. We are basically a citric acid cycle with some miscellaneous additional cycles as temporary doo-dads attached, most of which are to make up for the inefficiencies of the other doo-dads. What's why this is, was, and always will be a worm planet, with a light sprinkling of evanescent sports on the top.
It would be more accurate, in my opinion, to view DNA as off-line bulk storage. Not as critical, information-rich, structurally controlling content. DNA is inert crystals if you ask it to run the show. It does not run the show by communicating information as Shannon understood information, so it is, in my opinion, irrelevant in discussions of evolutionary fitness, to ask after the metrics of the information entropy it exhibits.
There is indeed a master control, that is why things are born and die. Just because your theory cannot explain it, does not mean it does not exist. You also forget something very important, an organism has many parts - but all of them must work together for the organisms to live and thrive. That so many different cells, genes within cells, processes, organs, do in fact work together and support each other's functions is totally amazing if not miraculous.
You are correct that it is more accurate to look at DNA as the software, but the distinction between the computational engine and the software isn't as clear as it is in Von Neumann architectures. Biology is essentially a purpose built computation engine and therefore doesn't need to distinguish between software and hardware; we are very comfortable with the general purpose computing paradigm, but it isn't an efficient model for single purpose applications. Software is just an abstraction of the hardware, and can always be described with complex hardware systems. DNA would be more analogous to punch cards for a Babbage engine.
Entropy IS irrelevant to the discussion of evolution and DNA for the most part. The only reason it gets dragged into these discussions at all is because people with very poor understandings of entropy think it IS relevant, mostly as a result of their poor understanding rather than as a matter of fact. Incidentally, while DNA is an example of Shannon coding, usefully discussing entropy metrics in this context has far more to do with Kolmogorov complexity which is a different branch of information theory than Shannon's work.