I came across this interesting error while surfing through Google groups! On 23rd Sept 2007, this error was first reported by one of the group members.
1. When you try to multiply 850 by 77.1 excel displays the result to be 100,000 instead of the correct answer, i.e. 65,535!
2. It seems that *some* formula that involves *multiplication* and evaluates to 65,535 will act strangely. [
3. If you add 1 to a cell that has one of these incorrect results you get 100001, but deduct 1, and you get 65,534!
4. Displays correct result in Microsoft Excel 2000 and 2003. So it appears to be an issue infesting Excel 2007 alone!
5. Even SUMPRODUCT returns 100000!
Let me Analyze the Error:
Did you notice the peculiarity about the error? This happens for some multiplication formula that should evaluate to 65,535! Like,
So it appears to me that it the specialty of the number 65,535 that makes the error unique! Interestingly enough, 65,535 is the highest number which can be represented by an unsigned 16 bit binary and 65,536 requires 32 bits. In other words, the number, 65,535, happens to have two bytes worth of 1 digits in binary. I understand, Microsoft Excel 2007 might be using floating point for numbers, and hence it might not apply here, still I doubt this has got a lot to do with this error! Bugs are usually encountered at boundaries. I wonder how the test team at Microsoft, who were testing the multiplication feature, could miss out such an obvious looking test! Also I am interested to see how Microsoft is going to react to this bug!
Can you see some more interesting attributes of this error? Can you link this to some other testing perspective? Do let me know your viewpoints on this by leaving your comments.