It's okay to make a loosely based analogy of genetics using computer software. But it's invalid to turn that around by implying that the limitations of software apply to genetics as well. By definition a bitflip is a change in the value of a binary state. If the difference between DOS and Windows XP involved a change from binary to a quantum state computing, then it would be a difficult task indeed to find a bitflip path between the two. But such is not the case, the software is binary, and your tangental wordiness about how this is represented at the user interface is completely irrelevant.
In the world of genetics, mutations include crossovers, duplications, additions, and subtractions. Bitflip mutation programs include these as well. If you line up the binary code for DOS and Windows XP, you can bitflip your way from one to the other with great ease.
Your original point was that this is not true for a genetic code. Let's start with something very simple like two same sized-genomes from two different species. Can you imagine bitflip mutations that would transform one genome into the other?
Not if the constraint that each intermediate be a functioning life form/program is to be satisfied.