summerof2010 said:
Lol! I freaked the hell out when I read that without your edit; I couldn't figure out how that happened. Yes, that was a typo there and it was supposed to be 39.94. I'm not sure I really agree with you though. I would need to see some data suggesting the degree to which rounding error affects the ultimate price -- if it were small enough, VAT pricing (I think I'm talking about that correctly) would still be practical in small consumer sales. And I can tell you right now that the price being particularly high or low has much less to do with the round off error than how "round" the price is. The error in for something that's like $5079.00 would be less than than something that's like $2.38.
Mathematical theory FTW!
I have a very simple example for you:
A product costs $9.95. I want 2 of them. For convenience let's say tax is 10%.
If taxed at the total price you get 10%*(9.95+9.95) = 10%*19.90 = $1.99 tax.
If taxed on each item you get 10%*9.95 = 0.995, which is rounded to $1.00 (so $2 total)
So one way you get one penny more than the other. You will only get at most 1 penny difference for each item after the first (although that becomes increasngly unlikely as you add more items).
Not much on an individual basis but when you consider how many transactions are made per day in even one county that can add up.