Re: terrible bug in Mathematica (//N), but why????
I tried to post on mathematica newsgroup with a cc to
here, but neither has come through yet. Here's a better answer.
As you know, Mathematica's command
expression //N does machine floating point evaluation of the expression
You probably want to do
expression //N[#,16]&
which does Mathematica's "significance arithmetic" , which might
give you a hint that something is going to be inaccurate.
So the "bug" is that //N or N[] in Mathematica should do
slow and careful arithmetic. Something like N[expression,MachineSpeedUnknownAccuracy]
could be an option.
RJF
Richard Fateman wrote:
Compute each part of
the first expression and look at the numbers you get. A careful
study of the results of floating point arithmetic may reveal why
the numbers come out the way they do.
symbio wrote:
Dr. Fateman,
Thanks for your response, but the problem I sent you is NOT using Sin,
it's using Sinh (hyperbolic sine is not periodic, rather exponential),
again I ask why does Mathematica not make this mistake for 44 or 42
but for 43?
.
