Obligatory reminder that in almost all cases, floating point binary type (i. e. Commonly known as float) should not be used to represent money. In particular, 64 bits only allow you to. Understanding that this is a normal behavior. The [fadd st, st (0) / sar i,1] trick fixes the round to nearest even number behaviour.
The author discusses ieee floating point representation that stores numbers in what amounts to scientific notation. He considers the sign bit, the logarithm function, function. Lets look at ways to manage and avoid these errors in javascript. One simple way to reduce floating point errors is by rounding off numbers to a fixed number of decimal places. Of course, the trick there is that js's min/max doesn't do strings. Other languages' min/max can handle arbitrary comparables, and comparables don't necessarily have an equivalent infinity. In this article, well look at what floating points.