I tried assuming the numerators were all positive:
ay > bx [1a] ay/x > b [1b] a > bx/y [1c]
cx > az [2a] cx/z > a [2b] c > az/x [2c]
bz > cy [3a] bz/y > c [3b] b > cy/z [3c]
Combining [1b] with [3c] gives
ay/x > cy/z
az > cx
which is a contradiction with [3a]
the others do the same
Assuming the numerators are all negative gives the same contradictions.
So the numerators must all be zero so the ratio x:y:z is the same as a:b:c

Posted by Jer
on 20121125 00:03:57 