Re: [Heinously Off-Topic] 2.0.30 - And undefined

Matthew Kirkwood (
Tue, 11 Mar 1997 11:01:28 +0000 (GMT)

On Mon, 10 Mar 1997, Neptho wrote:

> > >2 / 1 / 0 == 2 / 0 / 30
> > ^^^^^ ^^^
> > Actually, anything divided by zero is undefined; you can't divide anything
> > zero ways. So, in essence, 2/1/0 does not equal 2/0/30, as undefined is..
> > err, wait, I guess they are equal. Never mind.
> My my dear moriarty, are they the same undefined, or different levels of
> undefined?

People seem to have the wrong idea here. 1/0 does not equal some special
value called "undefined"; It _is_ undefined. ie it has no value; it is
the same sort of undefined (oops! an allusion; shouldn't use these) as
trying to take the first element from an empty list. There is no sensible
values so, for theoretical purposes, at least it has no value. In
practical terms, how things like this are handled are open to
implementors' choices. Intel have a divide by zero exception for
integers; IEEE FP as used by Intel give x/0 the special value NaN.

> I wrote the worlds most optimizing compiler;
> everything compiles down to a *single* NOP.

A *single* NOP? What a waste! Remove that NOP immediately. (In fact, I
copyright that idea, so _mine_ is the most optimising compiler :-)


PS. I find discussions like this quite fun, but perhaps I should
set up a list on our server for such off-topic stuff?
( :-)

Matthew Kirkwood  |  Mail:
LMH JCR,          |  Web:
Oxford OX2 6QA,   |  Microsoft: "Where do you want to go today?"
England.          |  Linux: "Where you want to go _now_?"