Let A be the set of information understood by the OP in reply a and B be the set information understood by the OP in reply b.
I took A and B to be the information in a reply as understood by the OP and that a and b where the reply or the litteral content of a reply.
A, B, a and b are not code but sets of information. b-a is the content left in b after subtracting any duplicate content from a.
The exception I take with your methodology in this statment and in Re^4: In support of downvoting plagiarism is that you are must quantify the content of a and b into it's descreate elements. This would essentially result in Av and Bv (A and B from the point of view of a voter). Then remove those elements in Av from Bv and if there is no remaning elements in Bv then b was redundant. However, because a voter processes infomation in a different maner than another voter or the OP, it is posible that B - Bv != 0 and/or A - Av != 0. Because of that it is posible that even when Bv - Av == 0 that B - A > 0.
I don't understand what you are trying to do with these.
I was attempting to use the analogue of Perl's comparison opperators to state this. Many begining Perl programers fall into the trap of comparing $a == $b when $a = "foo" and $b = "bar" and being supprised when the result is true. For a program it is because Perl attempts to reinterpret $a and $b to derive a specific form of information from them, even though there literel content is quite different their interpreted information is 0 for both of them. For people it is because the way peoples brains are wired varies and one person may process the content of a reply into a dramaticaly different set of information than another person would.
The end result of my line of thought is that if there are two posts, an early post a and a significantly late post b. For any post p, if the set of people that find p useful is non-empty the post p is useful and may deserve a ++, but almost certanly does not deserve a --. If the set of people that find p useful is empty p could be seen as redundant and may deserve a --, but almost certanly does not deserve a ++. When I find b to be more enlightning than a I know that the set of people that find b useful is non-empty and therefor may deserve a ++. However, if I find that b is not more enlightning than a I cannot say for certan that the set of people that find b useful is non-empty, but only that I am not contained in that set. Because I cannot determin the state of the set of people that find b useful I cannot say with certanty that b deserves a --. Please take note that enlightning and useful are orthogonal in this sense, two exactly identical posts are equaly enlightning but only the earlyer of the two is useful, or a more enlightning post might be less useful to some if it is posted to a SOPW from 5 years ago.
Why do after saying all that I feel like I should say, "And therefore I clearly cannot choose the wine in front of you."?