2002 (was:: Looking for pointers to old Control Data software)

From: Jgzabol_at_aol.com <(Jgzabol_at_aol.com)>
Date: Thu Jan 21 17:02:40 1999

Hello Hans,

you are certainly welcome to it - after all its only bits and not a single
atom,
so it can be copied easily. However, there is no documentation apart from
source code comments - it's a (hopefully) portable C program plus some
ASCII data files, so it should run on anything that has a C-compiler.

Since your address makes me believe that you are affiliated with Siemens
Munich,
and since I live in Munich too, maybe we just get together and discuss this ?
I will send you my coordinates in private email.

The simulator maps core and drum to RAM arrays, papertape input and output
to File I/O, and the console to keyboard/screen. It would be neat to interface
to
a real S2002 console - there are at least two in Munich, one at the Siemens
museum and one at the Deutsche Museum; maybe that would be a fun project
to try one of these to agree to that proposal ?????

The compiler I have is in fact the Alcor Triplex Main S2002, the result of a
thesis
(not mine) done at Mainz University in 1968, extending Algol by the data type
TRIPLEX which is an interval representation of a floating-point number:
  [lower-bound, standard-f.p., upper-bound]
Any operation on triplex numbers will produce the min/max values resulting
from possible roundoff, so that the error propagation of roundoff may be
studied. This is theoretically nice, but in practice you can invert a 10x10
matrix only if you are lucky and it is conditioned extremely well, otherwise
some pivot interval will include 0 resulting in a divide error. In those days
there were fears that one could not do any calculation involving millions
of f.p. operations because of roundoff - that turned out to be really
academic.

John G. Zabolitzky
Received on Thu Jan 21 1999 - 17:02:40 GMT

This archive was generated by hypermail 2.3.0 : Fri Oct 10 2014 - 23:32:08 BST