Comme ça : 
Date: 07/14/99 at 12:03:05
From: Doctor Peterson
Subject: Re: Computer programming
Hi Tom,
Basically, two's complement numbers are a way to encode negative
numbers into ordinary binary, such that addition still works. It goes
like this, using a single-byte number for simplicity:
binary: 0 1 2 3 125 126 127 128 129 253 254 255
+----+----+----+-...--+----+----+----+----+--...--+---+---+
2's comp: 0 1 2 3 125 126 127 -128 -127 -3 -2 -1
This is most effective if you draw it in a circle, so that -1 is
followed by 0. Essentially, the second half of the number line (255
being the largest number that can be represented by 8 bits) is taken
over for negative numbers. Whereas normally (for unsigned numbers)
there is a break between 127 and 0, at which point in counting there
is a carry into the missing ninth bit and you "wrap around" like an
odometer, in signed numbers there is a break between 127 and -128,
where there is a carry into the "sign bit." Rotate the circle and it
looks like this:
binary: 128 129 253 254 255 0 1 2 3 125 126 127
+----+-...--+----+----+---+---+---+---+--...--+---+----+
2's comp: -128 -127 -3 -2 -1 0 1 2 3 125 126 127
In binary, the negative numbers are those whose leftmost bit, called
the sign bit, is 1; positive numbers look just like unsigned numbers:
00000000 = 0
00000001 = 1
...
01111111 = 127
10000000 = -128
10000001 = -127
...
11111110 = -2
11111111 = -1
The system is called two's complement because the negative of a number
is formed by complementing each bit (subtracting it from 1, the "one's
complement") and then adding 1 to the result:
11111111
2 = 00000010
--------
-2 = 11111101 + 1 = 11111110
Because of the way the numbers are assigned, adding one in binary
still adds one to the negative number; in fact, when you add one to
-1, you get zero (ignoring the carry). To add any two signed numbers,
you just add the binary values, and again ignore the carry, if any:
2 = 00000010
+ -3 = 11111101
---- --------
-1 = 11111111
-2 = 11111110
+ -3 = 11111101
---- --------
-5 = (1)11111011
There are other ways in which negative numbers can be represented, but
this has become very popular because of the properties I've pointed
out.
You can find more about this in our archives:
http://mathforum.org/library/drmath/view/55733.html
- Doctor Peterson, The Math Forum
Mais je ne comprends pas grand' chose à cette démonstration ! 
En France l'établissement?