Jennifer

ANSI has received the email below submitting comments on a matter presently out for consideration by the US TAG for JTC 1.

As Administrator of the US TAG, please pass along these comments for their consideration in developing the US position on the question of a Fast Track Ballot for ECMA-376¦ ISO/IEC DIS 29500 Office Open XML File Formats.

Thanks
Henrietta

Henrietta Scully
Program Manager
Standards Facilitation
hscully@abnsi.org

-----Original Message-----
From: Steven G. Johnson [mailto:stevenj@fftw.org]
Sent: Friday, January 19, 2007 1:33 AM
To: isot@ansi.org
Subject: please object to fast-track ISO standardization of Microsoft OOXML

Dear ANSI-ISO standards committee,

I am writing in regards to the ECMA-376 standard, commonly known as the
"Open Office XML" (OOXML) file formats, which have been submitted to the
ISO/IEC for fast-track processing.  It is my understanding that you (and
other P-member bodies) have until February 5, 2007 to submit your
objections to the ISO regarding this proposed "standard", and I am writing
to urge that you do so.

I am a professor of applied mathematics at MIT with a background in
computer science, and as a long-time user of many different operating
systems and software packages I have a great desire for standardized
interchange formats.  However, OOXML seems to be designed so that it can
only every be fully implemented by a single vendor, Microsoft.  It is
certainly inappropriate as a "fast-track" standard for ISO, as OOXML is
clearly too complex to be evaluated and revised on a short time scale.

The problems with OOXML are many, and no doubt you are familiar with many
of them.  I will summarize a few of the points here that are most glaring
to me:

* The OOXML specification is over 6000 pages long, and does not
   leverage existing standards like SVG or MathML; it is far too
   complex to evaluate in a short time period, and is probably
   too complex to ever be practically implemented by anyone
   other than Microsoft (who can leverage many years of work on
   MS Office).

* The ECMA standards process for OOXML was seriously flawed.  Its charter
was:
        "to produce a formal standard for office productivity applications
        within the Ecma International standards process which is fully
        compatible with the [Microsoft] Office Open XML Formats."
It is inappropriate for an international standard to be explicitly
designed exclusively for the needs and requirements of a single vendor.

* OOXML provides no benefit over the existing OpenDocument ISO standard,
   which was recently approved and provides similar functionality but
   with 1/10 the length, thanks to its leveraging of other ISO and W3C
   standards like SVG and MathML.

The main proposed benefit of OOXML is its ostensible compatibility with
legacy Microsoft binary formats, but this "compatibility" itself relies on
undocmented portions of the standard.  To wit:

* The OOXML standard includes a number of tags like:
   - lineWrapLikeWord6 (Emulate Word 6.0 Line Wrapping for East Asian Text)
   - mwSmallCaps (Emulate Word 5.x for Macintosh Small Caps Formatting)
   - useWord2002TableStyleRules (Emulate Word 2002 Table Style Rules)
   - useWord97LineBreakRules (Emulate Word 97 East Asian Line Breaking)
   - wpJustification (Emulate WordPerfect 6.x Paragraph Justification)
which themselves are undocumented in the OOXML standard.  In describing
these features, the OOXML standard states:
        "To faithfully replicate this behavior, applications must imitate
        the behavior of that application, which involves many possible
        behaviors and cannot be faithfully placed into narrative for this
        Office Open XML Standard. If applications wish to match this
        behavior, they must utilize and duplicate the output of those
        applications."
Such language is absurd in an international standard.  It makes no
difference whether these tags are only implemented for "compatibility"
with legacy documents -- they effectively give a single vendor (Microsoft)
the sole ability to properly format OOXML documents converted from those
legacy formats, and eviscerate Microsoft's claims of any advantage in
backward compatibility.

* The OOXML standard is so closely designed for Microsoft products that it
even incorporates and standardizes bugs in those products.  Most
egregiously, OOXML mandates that 1900 be incorrectly treated as a leap
year (section 3.17.4.1 page 2522) in order to match a bug in Microsoft
Excel.

Many other examples of such vendor-specific constructions, inappropriate
for an ISO standard (especially when the OpenDocument standard is
available and has been implemented by multiple vendors), have been
identified by others.  I'll refer you to the blogs of Rob Weir and Bob
Sutor at IBM for incisive analysis of OOXML:
        http://www.robweir.com/blog/labels/OOXML.html
        http://www-03.ibm.com/developerworks/blogs/page/BobSutor
A number of relevant links are also in the process of being collected at:
        http://www.grokdoc.net/index.php/EOOXML_objections

Thank you for your time.  I believe the reputation of the ISO process is
at stake here---if the OOXML standard is approved, especially on a "fast
track" without adequate time to review such a voluminous standard, it will
be clear that ISO has just become a rubber stamp to promote the interests
of any corporation with deep enough pockets, even a convicted monopolist.
Please don't let this happen.

Regards,
Steven G. Johnson