in070132



From: Kevin Wright [kwright68@gmail.com]
Sent: Tuesday, January 23, 2007 5:07 PM
To: Rajchel, Lisa; Garner, Jennifer; Desautels, Sara
Cc: bryden@iso.org; kwright68@gmail.com



Subject:  Kevin Wright Contribution in Response to JTC 1 N 8455 - 30 Day Review for Fast Track Ballot ECMA-376 | ISO/IEC DIS 29500 Office Open XML File Formats

---

Dear Sir or Madam,

I am sure that you will receive many quasi-form letters expressing concerns with the Ecma 376 Office Open XML standard, so let me differentiate my letter with some personal comments.

I work as a statistician for a Fortune 500 company.  From time to time, I am asked to speak on issues of Data Quality.  In my presentations I always include examples that illustrate how Excel is a poor choice for long-term data storage and for interoperability with other software.  For example, I use SAS, S-Plus, R, Minitab and other statistical software programs to access data from Excel.  Every single one of these software packages is unable to 100% reliably import data from Excel.

One trivial example: in Excel, the data type for a column is determined by the first 16 rows of data.  If the column contains mixed data (that is, some numeric and some character) the column may either be identified as numeric or as character.  Importantly, if the rows are re-ordered and imported again into the statistics software, there is no guarantee the data will be imported the same way.

I readily admit that I have not used Excel 2007, which may correct problems associated with my example, but Ecma 376 is full of specifications that exist solely for backward compatibility with earlier versions of Microsoft Office files (such as the infamous 1900 leap-year bug).

I have other examples of data-quality problems associated with Microsoft Excel and wish to re-iterate that it (as currently
documented) is NOT an acceptable long-term data storage format.

I hope you will agree that Ecma 376 is not ready for "fast-track"
approval.  At a minimum, Ecma 376 should remove all references to undocumented features of previous versions of Microsoft Office file formats.  There is simply no place in an international standard for such language.

Sincerely,

K. Wright