|
Authored by: Anonymous on Thursday, July 19 2012 @ 08:17 PM EDT |
Almost none of the hundreds of billions of stored-program computers in use today
have a write-once storage area attached to them. And for the small subset that
do, its easy to consider that as a peripheral invention (and perhaps a
patentable one). It plays no part in the universal algorithm that the CPU
implements, for example.
[A specific example: Xbox 360 consoles contain a few dozen electrical fuses that
can be blown under software control. This is used to prevent hackers from
'downgrading' a console's operating system to an older buggier version.][ Reply to This | Parent | # ]
|
|
Authored by: Anonymous on Thursday, July 19 2012 @ 08:30 PM EDT |
>Having erasable memory for data storage is obviously irrelevant for the 'new
machine' argument.
No, it's very much relevant.
See, assuming non-self-modifying programs, ROM can be trivially replaced by
erasable memory--and nothing significant about the machine has been changed. It
would be absurd to treat the two cases differently. As PoIR says (and judges all
the way up to the Supreme Court are beginning to notice this) if you assume as a
legal fiction that a stored program creates a new machine when loaded, there is
no non-nonsensical place to draw the line. Does saving a file on a hard disk
create a new machine? What new machine does it create? Does loading that file
into memory into an un-executable segment create a new machine? Does starting
execution of a debugger capable of emulating that machine's instruction set
create a new machine? How about a debugger capable only of emulating some OTHER
machine's instruction set? None of these actions cause the computing machine to
"act" any different than it did before. Is it the act of setting the
PC register to a particular value that creates the new machine? Of course none
of these things do any such thing.
So how can all of these operations, done purely automatically
(logically/arithmetically) by the operating system itself, create a new machine?
It just doesn't make sense, it CAN'T make sense.
[ Reply to This | Parent | # ]
|
|
Authored by: PolR on Thursday, July 19 2012 @ 08:43 PM EDT |
True. However, the program executed by the computer
does not need
to be stored in erasable memory. Having
erasable memory for data storage is
obviously irrelevant
for the 'new machine' argument.
It depends on
how the computer is programmed. If it uses a functional programming language
which modifies its program as it executes the program must be in data
storage.
Besides operating systems for general purpose computers do store
program in RAM. They don't store programs in PROM and wouldn't work if they were
required to store programs in PROM. Your counterexample is a corner case rarely
encountered in practice, incompatible with how the vast majority of computers
are actually used and which doesn't take into account all possible ways of
programming computers.
By basic logic, to prove a statement wrong,
a single
counter-example is sufficient.
By basic logic this
depends on the statement. If I say my cat is white you can't refute this by
showing me a black cat. Besides I can return the fire because I have just given
counterexamples to your statement that data storage is irrelevant to the 'new
machine' argument.
Hence the arguments in the article are
wrong.
This defies basic logic. It is possible to have one wrong
statement in a mass of true statements. Here is another counterexample to one of
your points.
[ Reply to This | Parent | # ]
|
|
|
|
|