by mschwrtz Fri Jun 04, 2010 2:40 pm
That's an interesting approach sprparvathy. You are correct that S1 can be rewritten as, (x+y) (x-y) > 0. I don't know that I'd describe that move as simplifying, though. In fact, I'll chase down the answer based on your rephrasing, and I think that you'll find it pretty subtle. It will turn out that (x+y) (x-y) > 0 is sufficient to determine that |x|>|y|.
If (x+y) (x-y) > 0, then (x+y) and (x-y) have the same sign; they're either both positive or both negative.
There are three ways that this could be true: x and y could both be positive, and x greater than y; x and y could both be negative, and x less than y (that is further from 0 than y); x could be a positive and y negative, and x further from 0 than y. In every one of these cases, |x|>|y|.
Here's a way to justify that claim with algebra. Keep in mind the definition of absolute value,
If n>0, then |n|=n
If n<0, then |n|=-n
If both expressions are positive, then it's possible that x and y are both positive: x-y>0, so x>y, so |x|>|y|.
If both expressions are positive, then it's possible that x is positive and y is negative: x+y>0, so x>-y, so |x|>|y|.
If they're both negative, then x+y<0, so x<y, so |x|>|y|.
The best take-away here is that you should probably not approach this algebraically. Instead, think about absolute value as "distance from 0," at least for some problems. And if you do test values, explicitly consider various combinations of positive and negative values for x and y.
When you're testing values in an absolute value question, the relevantly different values are positive vs. negative, and big vs. small. Choose values accordingly.