Microsoft OOXML spec 'dangerously flawed'

Microsoft's Office Open XML (OOXML) document format specification is fatally flawed where it comes to spreadsheets, with many functions filled with careless errors, according to a critic.

The criticism comes in the midst of a pitched battle between different document formats vying to be the industry's choice for an "open standard" -- one that, in theory at least, is unconstrained by any one vendor and could be used equally well in competing applications.

The issue is important for businesses and other organizations because it affects the ability to keep documents accessible over time and to choose amongst competing office software packages. Currently most organizations use Microsoft Office partly because they are tied to Microsoft's proprietary document formats.

Microsoft's entry into the "open standards" arena is OOXML, which has been approved by a body called Ecma and has been submitted to ISO. Its chief rival is Open Document Format (ODF), based on the native file format of StarOffice/OpenOffice.org, backed by Sun, IBM and others, and already implemented in many office programs -- with the critical exception of Microsoft Office.

OOXML has been criticized in the past as encumbered with Microsoft intellectual property and as too complex to be effectively implemented by anybody but Microsoft -- the specification is 6,000 pages long, with 324 pages devoted to spreadsheet formulas and functions alone.

Now Rob Weir, a systems architect for IBM and a member of various ODF technical committees -- has alleged that even if third parties did manage to implement OOXML, many spreadsheet formulae wouldn't work properly.

Weir documented seven specific problems in a blog post, and said there are others. The problems relate to ambiguities and mistakes in trigonometric and financial functions, a function relating to setting workdays, and others, including the AVEDIV, CONFIDENCE and CONVERT functions.

In each case there is something left undefined, or something as simple as a typo, that Weir said would make the function work improperly if implemented by a third party as written.

Weir said the errors might be a side-effect of a rush to standardize the format. "OOXML's spreadsheet formula is worse than missing. It has incorrect formulas that, if implemented according to this standard, will bring important health, safety and environmental concerns, aside from the obvious financial risks of a spreadsheet that calculates incorrect results. This standard is seriously messed up," he wrote.

ODF supporters have good reason to find fault with Microsoft's spreadsheet standard -- since ODF doesn't yet define spreadsheets.

Weir said ODF developers are taking right approach by using specialist groups to review various aspects of the upcoming ODF spreadsheet specification. "Rather than rush, we're doing careful, methodical work," he wrote.

Microsoft did not immediately respond to requests for comment.