Hi,
Hans B Pufal <hansp_at_aconit.org> said:
> Jeffrey Sharp wrote:
>
> > I was thinking of the first machine that supported indirect addressing in
> > its instruction set, but we should also count machines where you could
> > kludge indirect addressing by modifying the machine code before executing
> > it.
>
> But surely ANY computer (other than Harvard architecture machines which
> seaprate code from data) can modify their own program. In that case I do
> not hesitatte to nominate the Manchester Baby machine, first operational
> program in June 1948. It did not have any hardwired indirect addressing
> but it could certainly modify its own program.
Indeed, and that was the normal way of indexing arrays on EDSAC (1949).
(I'm just writing some simple test programs for the emulator, so
I know that's the only way to do it!).
--
Cheers,
Stan Barr stanb_at_dial.pipex.com
The future was never like this!
Received on Thu Oct 03 2002 - 13:22:00 BST