|
Authored by: Anonymous on Thursday, July 19 2012 @ 08:07 PM EDT |
> A stored program computer must use erasable memory.
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.
> The structure mentioned in this argument is the
> structure of the machine, not some other kind of
> structure.
Programming a write-once program storage area embedded
in the processor chip, on the same die, does indeed change
the physical structure of the computer as the program
memory is a non-removable part of the computer.
The writing process physically alters the wiring diagram
of the computer and renders the computer capable of
performing only a single pre-programmed task.
The number of computers that are programmed in this way
(or by mask ROMs, same idea) vastly outnumber
general-purpose computers with rewritable programs.
Think digital wrist watches, microwave ovens, washing
machines, etc.
By basic logic, to prove a statement wrong, a single
counter-example is sufficient. Hence the arguments
in the article are wrong.[ Reply to This | Parent | # ]
|
|
Authored by: jesse on Friday, July 20 2012 @ 05:18 AM EDT |
I used to work with navigation systems.
These were based on the PDP-11.
Our standard implementation was to remove half the memory of the system and
replace it with a PROM that was loaded with the navigation program.
The processor worked just as designed. It could not detect the difference
between the PROM and RAM.
Most systems today have EEPROM - preloaded with an initialization and program
load function. Some of these are even divided into two parts - one part PROM
that cannot be updated, and one part EEPROM to receive updates.
The computer works as designed. It executes instructions taken from memory, and
it doesn't matter whether that memory is PROM, EPROM, or any other
implementation of memory.[ Reply to This | Parent | # ]
|
|
|
|
|