On Thu, 28 Jun 2001, Gene Ehrich wrote:
> I remember back in the early days at IBM right before the 360 was
announced
that they were
talking in the industry of the possibility of a
disassembler. Most people poo-poo'd it as impossible. Datamation magazine
had a cartoon with a picture of a machine labeled disassembler with a
conveyer belt on each side. A worker was feeding cans of applesauce into
the machine and on the other side apples were coming out.
Maybe hindsight is always 20/20, but I don't see why machine code
disassembly would be viewed as an impossible task? I mean, weren't people
able to look at hex or octal dumps and translate them into the
corresponding mnemonics? This generally isn't a task that requires a high
IQ, it's basically a simple lookup operation. Hell, the problem is better
suited to computer based solutions than it is to people doing it.
People still *are* able to look at octal dumps; but since my arms got
shorter, all the numbers blend together.
So, since I don't think my arms will get longer any time soon,
I wrote a dissassembler. Using an unfinished simulator that
handles the instruction decoding and welding it into a dump
utility that translates from the CDC 60-bit words of Display
Code (6-bit character set) gave me a useful disassembler.
Technically, it's not a disassembler, rather, it's just an
opcode dump. I need to add a scripting ability that I can use
to specify data areas that should not get tramslated into opcodes.
Funny thing, when I first started doing this 25 years ago,
I did just take an octal dump and disassemble it by hand.
Regards,
-doug q