The code is linear but the transfer functions are not. DNA codes for proteins. Proteins have primary through quaternary structure. Multiple combinations of multiple variants of tertiary structure peptides with enzymatic properties and extraordinary numbers of cellular, tissue, and whole organism feedback controls means that the instruction set will be expressed differently from place to place and over time.
A single gene may code for parts of a dozen or more different functional proteins. Which proteins occur in which cells is a function of the history of the cell line and environmental influences. In other words, the code produces a facultative response, tolerates a broad range of conditions and insults, and does so with very high tolerance of variation in the code itself.
That's my point entirely, thanks. DNA programming is superior to Man's current level of computer programming.