1280 and 1024 are being interpreted as integers and so integer division is used. You have to tell Blitz that one or both of the values are floating-point, before it will use floating-point division. To do this you need to put ".0" on the end of the number (as Fred shows).