Ben Langhinrichs

Photograph of Ben Langhinrichs
E-mail address - Ben Langhinrichs

Recent posts

Fri 27 Apr 2018

Our digital selves

Wed 18 Apr 2018

The long strange trip with Midas

Wed 4 Apr 2018

Pondering pre-anonymization and GDPR

May, 2018
  01 02 03 04 05
06 07 08 09 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30 31

Search the weblog

Genii Weblog

OOXML's Formula for Failure

Mon 9 Jul 2007, 11:06 PM

by Ben Langhinrichs
Rob Weir has an excellent essay on The Formula for Failure, which takes a closer look at Microsoft's Office Open XML (aka OOXML or Ecma standard 376) and finds a number of deficiencies.  While a few of these may be nitpicky, others are quite clearly flawed in critical ways.  His essay clearly refutes the argument that OOXML has a fully defined formula specification while ODF does not.  ODF is quite clearly further behind in defining the formula specifications for its spreadsheets, but it seems likely given their process that they will wind up ahead, especially since Microsoft and Ecma have seemed so resistant to making even very obvious modifications such as removing bit flags from the spec and replacing them with individual flags.

I highly recommend that anybody interested in the file formats take a look at Rob's latest work.

Update: The Open Malaysia blog has an excellent follow up called Mathematically Incorrect showing a very specific example, and how the slap dash approach by Microsoft and the Ecma standards people has led to serious inadequacies in the OOXML specs.  I recommend reading it in its entirety.

Copyright © 2007 Genii Software Ltd.

What has been said:

619.1. Charles Robinson
(07/10/2007 05:39 AM)

The Open Malaysia blog also has some good info:

I did some reading on OOXML over the weekend, and there is no complete implementation of the spec. I'm shocked and dismayed that a body like ECMA could so be so thoroughly swindled. I'm hoping that ISO isn't similarly blinded by the hand waving of the Microsoft magicians who have little more than smoke and mirrors.

619.2. Ben Langhinrichs
(07/10/2007 06:32 AM)

Thanks. I updated the post to show the excellent post the Open Malaysia wrote in response to Rob Weir's post, which shows the very specific example of the CEILING function, incorrectly implemented in Excel and OOXML, but carefully corrected in ODF with a flag for legacy (incorrect) functionality.

619.3. Ed Brill
(07/11/2007 01:57 PM)

Ben, thanks for posting this. I am glad to see others outside of those of us with dogs-in-the-hunt (or perceived so) writing on this topic. There is a heck of a lot of misinformation out there. I was interviewed by a reporter last week who really seemed to believe the MS story, until I asked him "where would we be today if there had been two competing HTML specifications?" Maybe that's the magnitude by which people should look at this.