The use of fixed point representation means that we can represent any weight using a regular integer, and just understand that a point is implied between bit 3 and bit 4 (remember, the least significant bit is bit 0). In other words, given an integer with a “native” value of x, it is implied that that its represented value x′ is actually x∕100002.
This also means that we can use the usual integral operators for addition, subtraction and comparison.