Why is binary subtraction referred to as the invert-add-shift-add method?

Answer 1

This is because the given term refers to the method by which it works. You can subtract binary numbers in the way we're taught to do it in grade school, but computers do it using a different technique.

For example, let's say you're subtracting one number from another. We'll call them B and A respectively. The way this technique works is as follows:

1. Take the complement of the number being subtracted (see footnote)
2. Add it to the other number
3. Remove the first digit of the result, and add it to the last digit

So we'll start by picking a couple of binary values for A and B:

A = 10110101

B = 101001

They must have the same number of digits, so we'll pad the front of B with zeros:

B = 00101001

Now take its complement:

B' = 11010110

Now add them together:

10110101

+ 11010110

=110001011

And remove the leftmost digit of the result, adding it to the resulting number:

10001011 + 1 = 10001100

Giving us the result. If you want to double check that, just add the result to the number we subtracted in the first place:

00101001

+ 10001100

= 10110101

And you can see that this is our original value for A.

This is where that name comes from. First we "invert" the number we're subtracting by taking its complement. Next we "add" it to the first number. Then we "shift" the leftmost digit of the result over to the right end, and "add" it to the rightmost digit.

It is important to note that this is not unique to binary numbers. You can apply this to subtraction in any base. For example, let us subtract the decimal number 2343 from 14555:

We start by expanding 2343 to five digits by preceding it with zeros:

02343

Now we take its complement:

97656

And add that to the number from which we're subtracting:

14555

+ 97656

112211

Then we take the leftmost digit and add it to the remaining number:

= 12212

And again, we can double check our answer:

2343 + 12212 = 14555

Demonstrating that this technique works.

Note:

The complement of a number is another number in which each digit, when added to the corresponding digit in the original number, will give us the largest available single-digit value in the base we're working in. For example, in regular decimal notation, the complement of 12345 is 87654, because each corresponding pair of digits (8 + 1, 7 + 2, 6 + 3, etc.) adds up to 9, the largest single digit value in decimal. Similarly, in hexadecimal, the complement of of 12345 would be EDCBA, and the binary complement of "11011" would be "00100".