On or about 2/4/2009 9:41 PM, James Knott typed the following:
Many early computers, including some from IBM worked in decimal,
using a
modified hexadecimal system. One method that was used in an ancient
computer I used to work on was called "excess 6", IIRC. This method
used
four flip flops to hold a number, but if you read the binary it was
actually 6 more than the decimal value. This made it easier for the
stage to overflow and pass on the carry bit to the next stage.
Setting
one stage to 6, representing 0, was easier than detecting a binary
equivalent of 10 for the carry etc. IIRC, it was IBM's commercial
line
that used decimal, whereas the scientific stuff ran binary.
The early one-off machines, ASCC and SSEC, used decimal for scientific
work, as did IBM's first run at a supercomputer, the one-off NORC. But
the 701, IBM's first production model, went with binary, and they
continued to use it for science work until the System/360, which used
fixed-point binary for addressing, fixed-point decimal for accounting,
and floating-point hexadecimal for science. As I said before, current
models of the System z, the successor to the System/360, have IEEE
binary and decimal floating-point.
--
John W Kennedy
Having switched to a Mac in disgust at Microsoft's combination of
incompetence and criminality.
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
