Hacker News new | past | comments | ask | show | jobs | submit login

I don't know about databases but if that decimal is the decimal I know of, it is not a floating point number.

In decimal you specify how many bits will be used for decimal & non-decimal parts. If your unsigned decimal has 1 bit for non-decimal and 1 bit for decimal, then it can only store: 0.0 0.5 1.0 1.5. 1 bit for non-decimal gives you either 0 or 1 and 1 bit for decimal gives you .0 or .5. In many ways a decimal is just an integer, except the printing & some math operations being different

In floating point, you are also storing "exponent" which represents where the point goes. So the point actually "floats" depending on value of this exponent while in fixed point the point is fixed (and in previous example, the point is always in between those two bits. The implementation is a bit harder to explain, you can check it here: https://www.doc.ic.ac.uk/~eedwards/compsys/float/

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact