At 13:04 -0500 9/27/10, Patrick R. Michaud wrote, and I snipped a lot: >In the case of a Float operation returning a Complex >result, the Complex PMC has some substantially different behaviors from >Floats -- stringification being one of the more immediate and obvious ones.
Just a potential user here but that scares me. Surely there is a type complex which is an array of two floats or doubles ready to participate in math operations without converting to strings, isn't there? Do those strings use binary to decimal conversion? MS Excel does that and it slows everything down to the point of being useless. But my preference would be to return an error or a well defined NaN unless the user deliberately requests complex by entering sqrt($X) with $X being previously defined as complex ($X = -1+i*0). Perhaps any previous use of type complex could trigger the complex result. Un-reported sqrt(-1) errors could be disastrous if they occur in the middle of a complicated calculation. That could generate very wrong answers if the complex part somehow disappears long after the line that didn't work as expected. -- --> From the U S of A, the only socialist country that refuses to admit it. <-- _______________________________________________ http://lists.parrot.org/mailman/listinfo/parrot-dev
