Numerical evaluation of sequence looks suspicious



Re: ev(sin(%pi/5),float)=> 5 sin(0.2 %pi) etc.

I'm afraid this is the correct behavior, because the 'float' flag is defined
in a peculiar and unintuitive way.  For the expected result, use
float(sin(%pi/5)) or ev(..., numer).

The "ev" pseudo-function is another rich source of confusion, but let's not
get into that here.

              -s