|
Authored by: PolR on Tuesday, July 24 2012 @ 11:00 AM EDT |
"It takes more than... to make a machine"
This is the crux.
What more than moving atoms around is
needed to make a new kind of molecule or
even a new kind of
mousetrap?
. . .
Your last response, I think, boils
down to "but the analogy
between bits and atoms doesn't hold!"
I
thought I was merely pointing out that there are objects in this world which are
not machines. Or is it that I am confronted with an argument that everything is
a machine and that as long something is new it is necessarily a new
machine?
Water flowing in a river is moving atoms around. The turbulences
create new shapes every second. Lots of atoms are moved around. Where is the new
machine?
Or is it that I am confronted with an argument that the definition
of the word machine is unimportant and the courts are free to say it means
"smurf"? This is the complaint of this article, that we are confronted with a
law on what is an issue of fact.
I don't think I was trolled. I think the
trivialization argument is overly simplistic and easily refuted. If the claim is
a new machine is made, then we can answer "show me the machine." If there is no
machine in the trivialization then it doesn't apply.
Perhaps you should read
the article again. The argument is not "this is all about bits". The argument is
"show me the machine, I think you can't, nothing you can show me is a machine".
The argument is also "the machine you think you can show me doesn't exist, the
computer doesn't work like you think it works."
You say there is a policy
issue. This part I don't understand. Where is the policy issue that drives the
courts to call software a machine when it may just as well be patented as a
process? The patentability of software is not even at stake.
[ Reply to This | Parent | # ]
|
|
Authored by: PolR on Tuesday, July 24 2012 @ 12:00 PM EDT |
The argument is "show me the machine." I will recap a few salient
points from the article.
Bits are modified in a computer billions of times per second. If changing bits
in a computer make a new machine, then no computer can run software because as
soon as an instruction is executed, some bits are modified and this is no longer
the same computer. The clock ticks, a new computer is made. You press a keyboard
key, a new computer is made. Etc.
If you arbitrarily discriminate between bits, saying some make new machines and
some don't, what is the test? How do you know which bits make a new machine and
which bits don't?
You can't say executable instructions make new machines and data doesn't. I will
program a Python interpreter and these are the instructions. They will never
change. But the Python bytecode, this is non executable data, will be all
further programming from then on. Or in the alternative I will use a LISP
interpreter which generates and execute code on the fly. The machine will not
remain the same for the whole execution because the code itself is constantly
changing.
The argument is not that the bits are special, somehow different from atoms. It
is that the computer doesn't work in a manner where programming can reasonably
be construed as making a new machine. One common error is to assume that all
programs are machine executable instructions which never change during
execution. This is not the case.
[ Reply to This | Parent | # ]
|
|
|
|
|