Machines can be broken into two parts: the interface to the outside world and decision logic.
Decision logic can be hardwired or configurable. A configuration of the configurable
parts of a machine comprises the software. Software may be viewed as a program or
as data. This is an artificial distinction: data may be viewed as a program for an interpreter
(notepad.exe interprets a text file and generates an interactive graphical experience that
appears to represent the contents of the text file). Some data may be interpreted by
a hardwired logic unit.
All this was invented or discovered by WW2 with the exception of the fine details of the interface.
Use of a logic processing machine should not be considered patentable in any way.
In principle, once Turing's paper was written, one could enumerate all software for all machines.
Thus, since all the countable numbers were certainly discovered and logically formalised
in the 19th century, they should be considered already discovered.
As soon as an outside world interface involves decision logic, the decision logic should
be factored out and only that part that does not involve decision logic should be considered
a design and be patentable.
Copyright should not apply to what can be logically derived from a problem specification: it
should apply only to artistic expressions that are beyond derivation by a logical procedure.
Thus, what is truly art should be copyrightable (and there should be a common sense test
of whether someone could have produced something similar). Designs should be patentable,
but not decision logic. Decision logic should be considered already invented as of Turing's
paper and thus beyond the scope of copyright or patent.
That is the only logically sensible way to apply copyright and patents to technology. Trouble is,
laws are made by technologically naive people, presided upon by technologically naive judges and juries,
and essentially those who understand technology properly have no power of those who make laws
regulating it. This a dichotomy between understanding and power is why the system is as stupid
and broken as it is.
DNA should, accordingly, be considered software for a pre-existing interpreter that is over 20 years
old and thus equally non-patentable.