Kent Johnson wrote: > *TEST FIRST* Don't optimize until you know it is too slow and you > have a test case that you can time to see if your 'optimizations' are > making it faster. Pardon my shouting :-) Kent