"(/ 0 0.0) = 0" is a feature. "0" is exact 0, and "0." is "inexact 0". And something that's exactly zero divided by something that's not exactly zero is 0. There's also "-0." which is inexact negative 0. You've also got inexact infinities and inexact NaN's. If you divide a positive number by inexact negative 0, you get inexact negative infinity, and the other sign combinations work as you'd expect.
wow- I never would have expected that... It seems that anyone who has a computer program dividing by zero would want an exception raised or at least have an NaN generated. Generating "0" makes no sense- In my scenario, for instance, the function limit was towards 1, so generating a zero is mathematically seemingly completely nonsensical.... Thanks for straightening me out, kens :-)
In floating point arithmetic 0.0 isn't zero: it represent a range of numbers around zero, so it makes a little sense for 0/0.0 to return 0.0. I still think that raising an exception would be preferable.
