Could you please correct the title of the posting? It's patently false and needlessly inflammatory (as evidenced by the multitude of reading-impaired commenters).

I'm having a hard time imagining how this could have happened. Does Excel use binary coded decimal for calculations or something like that?

Excel has always been an abysmal excuse for spreadsheet software, but who would expect anything better from a multi-billion dollar company...

eh, could have been a simple enough mistake, who knows, maybe somebody forgot to put a semi-colon or something somewhere.

I'd be more interested in knowing what causes the bug, and second, why does it spit out that particular number? (100,000)

First of all, not to sound too anal, but it's 77.1 * 850, not 7.1 * 850, that equals 65535. Fix this, please.

Secondly, to #16,
1.) there's no such thing as "binary computerlanguage". You can say "binary code" or "machine language" if you want.

2.) Early computers were, in fact, decimal. Switching to binary actually caused large performance gains, so getting large performance gains from switching *back* to decimal seems somewhat unlikely.

Lastly, to #24
Techincally, the maximum value of a 32-bit (unsigned) integer is
2^32-1 = 4,294,967,295. 4 GiB is 4,294,967,296 bytes. 4GB, on the other hand, is exactly 4,000,000,000 bytes. Isn't meaningless nitpicking fun?

and THIS is why i don't program any more :) a wasted degree, but no more debugging :)

UberGerbil: But the real question is why this wasn't caught in the code reviews before it ever got to testing, and why a change was made to the calculation engine without adequate design work.

Since the actual result is generated correctly and only part of the directly displayed result is provided incorrectly (for instance graphs of the data would be correct) I don't see the surprise this slipped past. The automated testing most likely used things like VBA and saw the real result, not the displayed result.

In my opinion, code review isn't that great a check and balance. Old OSS has new bugs and exploits still exposed even today. Given enough eyes all bugs are shallow I think is a patently untrue statement.

They better cut the price on MS Office......

http://www.joelonsoftware.com/items/2007/09/26b.html

From the text in the article and the comments, it appears that it's a problem with the display function and not the calculation itself. Unfortunately, they haven't detailed what the exact problem is that makes it display incorrectly, because I would be really interested to find out what it is.

That's pathetic. For the money they charge, they should have a mathematician come to your house correct this by hand.

Screw the bugs, the interface is a total kludge. And I'm not talking about the 'ribbon.'

If you go to Print a document, then Print Preview, your only options at that point are to print it, or cancel. You can't go back and change the options...

Oddly, you can do so in Word.

I also found an issue involving permanantly locking certain values that were password protected, and a minor quibble copying sheets.

Maybe they need to employ a tester or two ...

sounds like something caused by shoddy type casting