Floating point overflow

This may be a dumb question.

I think that, for instance (* most-positive-double-float most-positive-double-float) should signal an error: it doesn't but returns a representation of double-float infinity.  I could not find a way to control this (or actually a reference to it at all), and I may well be wrong about what it should do (I'm not a floating-point expert but I realise it is hairy).

This is LW 6.1, 32bit on OSX.

Thanks

--tim