Ahh.. My first thread =)
The xor function is often explained as 'exactly one of' in school, however as a programmer I prefer to look at it as a conditional inversion.
For less than three inputs this is just two views of the world but for three of more these two definitions yields different results.
So exactly one of them ;) must be the real definition.
A little searching reveals the following on Wikipedia.
Quote:The operation yields the result TRUE when one, and only one, of its operands is TRUE
Quote:For more than two inputs, xor can be applied to the first two inputs, and then the result can be xor'ed with each subsequent input
Quote:The operation yields the result TRUE when one, and only one, of its operands is TRUE
AFAIK XOR works on two inputs only (at a time). so that definition above is for two inputs on the XOR. For more than two inputs, xor can be applied to the first two inputs, and then the result can be xor'ed with each subsequent input.
Which bit are you confused with?
Quote:Original post by BosskIn Soviet Russia, you STFU WITH THOSE LAME JOKES!
XOR is a mathematical binary operator. That means that it works on two operands only. What the definition meant was, that if you want to use XOR on more than 2 operands, you have to do the XOR operation more than once.
Quote:Original post by doynax Ahh.. My first thread =)
The xor function is often explained as 'exactly one of' in school, however as a programmer I prefer to look at it as a conditional inversion. For less than three inputs this is just two views of the world but for three of more these two definitions yields different results. So exactly one of them ;) must be the real definition.
A little searching reveals the following on Wikipedia.
Quote:The operation yields the result TRUE when one, and only one, of its operands is TRUE
Quote:For more than two inputs, xor can be applied to the first two inputs, and then the result can be xor'ed with each subsequent input
So now I'm more confused than when I started..
you should not let multiple operations change anything. XOR is associative. an operation just needs an LHS and RHS. If looking at A ^ B ^ C ^ D is confusing, just break it in to (A ^ B) ^ (C ^ D) now it's just a plain old xor. XOR A with B, C with D, then the results of each.
This gives a general definition for XOR: It is true if there are an odd number of true values.
The XOR Operator is mainly used as binary Operator, and therefore your definition applies only to the binary version of XOR. If you want to use a XOR with more than two Operands then you would have to split it up.
e.g. XOR with three operands a, b, and c:
xor(a, b, c) = xor(xor(a, b), c)
It does not matter in which order you split up the XOR:
xor(xor(a, b), c) = xor(a, xor(b, c))
Surprisingly a much simpler definition of XOR, that works with as many Operands as you like would be:
XOR evaluates to 1 if an odd number of its operands evaluates to 1, otherwise it evaluates to 0.
I don't know why this definition is not the mathematical correct one, perhaps since it allows XORs with 1 or even 0 operands.
In digital electronics there exists gates (such as xor) which have several inputs and one output. Often these are internally implemented as described before altough the information of the exact implementation is not necessary if you use the definition of xor for multiple inputs. (odd -> true, even -> false)
