Calculation Flaw Spotted in Excel

Published: October 3, 2007

by Alex Woodie

Computers have become so powerful and sophisticated that we tend to take the seemingly simple things, such as addition and multiplication, for granted. However, a recently discovered calculation flaw in Excel 2007 serves as a reminder that computers and software programs aren't infallible gods of the logical, and that they're only as smart as their owners.

The multiplication flaw in Excel 2007 and Excel Services 2007, which was disclosed last week on Microsoft's team blog for Excel and Excel Services, only rears its ugly head when users try to multiply or divide a handful of certain numbers in a very narrow range, according to David Gainer, group program manager for Excel.

Gainer writes: "Of the 9.214*10^18 different floating point numbers . . . that Excel 2007 can store, there are 6 floating point numbers (using binary representation) between 65534.99999999995 and 65535, and 6 between 65535.99999999995 and 65536 that cause this problem." Users can't enter these numbers directly, since Excel rounds to the nearest 15th digit, but the problem will present itself if the results of a calculation are displayed in a cell.

Apparently, the problem isn't in how Excel performs the calculation (phew!) but in how the results are displayed in the spreadsheet (uh oh!). Rest assured, Gainer's team already has a fix for the program, and they're waiting until it's fully tested before releasing it to the public.