A fixed point representation is essentally an integer, but with an implied point between two fixed digit positions.
In some applications, it is okay to use a fixed point representation for numbers. For example, in the U.S., we frequently measure length by inches, and weight by pounds. Both are exceptionally well suited for fixed point representation.
If we need to represent weight quantities down to ounces, then this by itself is already a fixed point representation, only that the unit is a pound. Let me illustrate. The weight “5 pounds and 7 ounces” can be expressed as a fixed point number. Because there are 16 ounces in a pound, we can rewrite this as “1012 pounds and 1112 ounces”. Next, we convert the ounce units into pound units. There are 16 (100002) ounces in a pound, so the weight we want to represent is “1012 and 1112∕100002 pounds”.
The rule of division is the same regardless of the base. In other words, in binary, 1112∕102 = 11.12, 1112∕1002 = 1.112 and so on. This means that the weight can be represented as 101.01112 pounds.
If the required precision is limited to ounces, then we can fix the number of “binary places” to 4.