Awhuxlb.206
net.micro
utzoo!decvax!harpo!zeppo!whuxlb!mkg
Wed Apr 7 23:44:33 1982
IBM PC divide bug revisited
Having access to an IBM PC, I couldn't resist trying to make the
1/10 giving .001 bug happen. After many tries, I discovered that
as long as the division is done with variables, the result is ok.
To successfully cause the "bug", I had to do the division using
double precision literals (print .1#/10#). Unfortunately, I don't
think the bug is in the output routine as was earlier suspected.
The following program also fails!!
c#=.1#/10#
print c#
Maybe IBM will handle this the way they sometimes "fix" bugs on their
large systems... They'll change the manual and call it a "feature".
Oh yes, one other thing. Did you know that .1#/100# == .1#/1000#?
They both yield .0001.
Marsh Gosnell BTL Piscataway (...!whuxlb!pyuxr!mkg)
-----------------------------------------------------------------
gopher://quux.org/ conversion by John Goerzen <
[email protected]>
of
http://communication.ucsd.edu/A-News/
This Usenet Oldnews Archive
article may be copied and distributed freely, provided:
1. There is no money collected for the text(s) of the articles.
2. The following notice remains appended to each copy:
The Usenet Oldnews Archive: Compilation Copyright (C) 1981, 1996
Bruce Jones, Henry Spencer, David Wiseman.