[e2e] OT: a different number base

John Day day at std.com
Wed Apr 2 08:42:35 PST 2003


At 11:06 -0500 4/2/03, David P. Reed wrote:
>PL/I had two number bases in its formal abstract semantic 
>definition, decimal and binary.   If you read the original "abstract 
>semantics" specs documents published by IBM in about 1969 (I don't 
>think I have mine anymore, I bet I could find one through friends 
>though) they REALLY tried very hard to make sure that the float 
>decimal and fixed decimal data types emulated the quirks of decimal 
>machines in terms of precision, etc.

Right, but didn't the 360 do both BCD and binary arithmetic?  PL/1 
was originally developed for the 360.

Also, what about the Burroughs 3500?  I don't know, but since it was 
architected to be a Cobol machine it might have done base 10 machine 
internally.

The B5500 was binary but had interesting take-offs on it.

Take care,
John




More information about the end2end-interest mailing list