[odf-discuss] Re: countering Patrick Durusau's public statement favoring OOXML as an ISO standard

[Lars Noodén]

On Wed, Mar 26, 2008 at 11:39 AM, Carlos Moffat <
 carlos.lst at eldiabloenlosdetalles.net> wrote:

> ... how defective the process by
> which it was forced through ISO has been... 

That's the epicenter from which all the other problems cascade.
However, nothing can be forced without ISO's compliance in deviating
from protocol.  ISO is *not* the forum for development of new
specifications, which *everyone* even those in the MS movement see and
admits.

My 2¢ are as follows.  My ability to participate will be a bit
unpredictable.  The [xxxx] are where I am missing links.

-Lars

There are a lot of fallacies presented by the MSFT side and mistakes in
procedure, even by people who should know better like Dursau [1], that
it is clear that ISO must send back the proposal.   Time would be better
spent getting MS to join the rest of industry in backing the
OpenDocument Format which not only already went through ISO
certification a few years ago, but also was a finished draft.

First off, the International Organization for Standards (ISO) is not in
the business of creating or developing standards.  Only when the
specification is finished and ready for evaluation is it appropriate to
submit to ISO.  The bar is set even higher for specifications to be
considered on the Fast Track.

If MOOX gets another no vote, it goes off the Fast Track like it should
have according to ISO procedure the last time it got a no vote.   Then
it can still be considered on the regular track, which gives national
bodies the time they need to actually discuss and resolve the issues.
Perhaps Durusau, with his years of experience working with international
standards,  doesn't know that already and should be reminded.

The changes to-date on the fast track have been minimal.  So far nearly
all of the discussion in favor of DIS-29500 boil down to one argument:
MS wants the ISO stamp of approval.   However, procedural, licensing and
implementation questions do not matter if DIS-29500 never becomes widely
used, and wide-spred use is most unlikely - seeing as the specification
s not even fully defined yet, and not even MS uses it or even plans to
use it [3].  The chance is close to nill.  Meanwhile the standard it is
intended to compete against, namely the OpenDocument Format,  is
implemented in several dozen products and tools and is growing in use.
[xxxx]


The procedural irregularities, stacking of committees and other apparent
violations of procedure and ethics, for just this specification alone,
could fill a small book now, even if each incident is given only a brief
summary.  There appears to have been no country where such problem have
not manifested, not even the Nordic countries, which ones had a
reputation for proper procedure.  Further there has been identified a
statistical correlation between general levels of corruption and
favoring DIS-29500 [xxxx]  Further, these practices are deeply
entrenched in the backing organization's culture[4] and have been used
against other standards like again and again.

Size is a lesser problem, perhaps the least, but easiest to see and no
less insurmountable for all that.  The specification started out at 6000
pages and is now about 7500 pages.  Assuming a 48=week year, that
requires a conscientious 31 pages per day.  For a 5-month review period,
that would require a pace of 375 pages per day.  There are few people
who can manage even recreational reading at a rate 375 pages per day for
5 months.  Managing a technical specification at that rate, with its
interdependencies and cross-references, is highly unlikely.

On the technical side, a 7500-page specification is not ever going to be
something within reach of a 5 or 10 person shop.  Some of the larger
teams will have trouble with it as well.  For example, look at the
extreme difficulty one large vendor has had in implementing short and
relatively straight forward standards like Hype Text Markup Language,
Cascading Style Sheets, Portable Network Graphics, Java or Kerberos.
There would be no possibility of implementing a long, complex specification.

Furthermore, unlike OpenDocument, DIS-29500 does not re-use existing
standards, so that means nearly the entire 7500-pages are new ground.
In contrast, for OpenDocument, tools have existed for years to work with
its components, such as Scalable Vector Graphics.

On the licensing issues which disqualify it from being an open standard,
they seem to explicitly block MS key competitors, especially FOSS
projects, from using it.   That defeats any claim to the goal of
following OpenDocument as a universal format for document interchange.

OpenDocument is already the better format, both technically and in terms
of licensing.  Further, it has the backing of a 600+ member organization
of which even Microsoft is a member of.  The only thing further debate
does is to burn up valuable time, especially those few experts in
document standards, interoperability and digital archiving.

Even on a procedural level, sending the specification back to Ecma for
further work is the only option for ISO.

Regards,
-Lars

O
[1]	http://www.robweir.com/blog/2008/03/contra-durusau-part-1.html

[2] 	http://www.openmalaysiablog.com/2008/03/fear-that-shoul.html

[3]
http://www.techworld.com/storage/features/index.cfm?featureid=3685&pagtype=all

[4]	http://www.groklaw.net/article.php?story=20071023002351958