[odf-discuss] procedure at ISO Ballot Resolution Meeting

[Alex Brown]

Marbux,

Just found this while googling for OOXML stuff. Some clarifications below ...

marbux wrote:
> I have mixed impressions of Brown. On the one hand, he instigated and led
> the BSI group that developed the comments for BSI

I did not "instigate" the group -- I was appointed as chair of the group by BSI committee
(IST/41), which had resolved to establish a technical panel for reviewing DIS 29500.

> and I looked at the final
> Great Britain comments last night.  They are very high quality, voluminous,
> and very tough on MOOXML, with a hard-core stance of "must replace giant
> chunks of MOOXML with corresponding ODF sections." I.e., harmonization.
> Moreover, there are many comments with proposed changes that almost
> certainly Microsoft could not implement without rebuilding at least MS Word
> from the ground up (I'm less familiar with the other Office major apps.)
>
> E.g., several BSI comments call for changes to match features with ODF that
> impact page breaks in Word. Page breaks in Word have been exceedingly
> fragile and buggy since Word 1.0. It's clear that Microsoft has been unable
> to stabilize them, most likely because of architectural screwups in in
> Word's life and the mountains of spaghetti code atop the features that
> impact page breaks that are likewise extremely buggy, e.g., footnotes and
> other subdocuments, tables that continue on a subsequent page, etc. See e.g.,
> detailed study of Word footnote and endnote bug reports in the Microsoft
> Knowledge Base here that include data loss bugs as well as formatting bugs.
>   <http://www.llrx.com/features/word.htm>, from which one can also infer
> that at least much of Word's page layout engine is still 16-bit code.
> (Compare bug reports with version numbers of Word.)  My impression is that
> page breaks are so brittle in Word that several of the changes BSI has
> requested simply can not be implemented.

I think it's fair to say that the implementability (or not) of a revised text was generally not
a consideration for those reviewing DIS 29500. The general view of standards (in the UK) is that
implementations follow standards, not vice versa, unless there are compelling arguments to do
otherwise.

> Another BSI comment calls for WordprocessingML tables to be expanded from 67
> columns to 4008 (IIRC) to match ODF for interop purposes. Word tables have
> been largely off-limits for major changes for years, I suspect in no small
> part because of the page break issues.
>
> But back to Brown. As I understand the situation, it was his decision in the
> BSI working group to: [i] limit participation and input to invited
> participants only (no way provided for others to submit comments);

The rules of admission were determined by BSI. In practice admission to the panel was by
application as much as invitation. Everybody who applied to join (29 people) was given a place.
The panel comprised reps from large and small companies, public sector organisations, user
groups, academia and government. No one interest was over-represented and (in my view anyway)
there was a good balance of opinions.

Throughout the review BSI was also running an open public comment process, and this received a
large number of submissions exhibiting a wide range of views. Any new *technical* comments
(there were a few) were forwarded to the panel and incoporated into the Wiki.

> and [ii]
> decreed that all non-technical comments raised at the contradiction stage
> could not be considered because of a JTC 1 ruling (that apparently never
> existed) holding that that comments raised in support of contradictions had
> been rejected and therefore were not eligible for consideration in the
> post-contradiction phases.

That is not a correct characterisation at all. The terms of reference for the panel were clearly
set by BSI to be purely technical
(http://www.xmlopen.org/ooxml-wiki/index.php/Terms-of-reference). Questions of contradication,
law, etc.  were handled at a different level within BSI (by a committee called ICT/-/1).

> In the his blog article under discussion, he announces that patent
> disclosure and IPR issues are off the table, as though standards bodies
> including JTC 1 do not routinely revise their work to work around patents.

What I write is that "the BRM can (effectively) only take decisions on the text which can be
implemented by the project editor. Discussion of legal and IPR issues at the BRM is out-of-scope
and would be pointless. Countries that have concerns in this area need to pursue them directly
with JTC 1, and not wait until the BRM."

In other words, they _are_ on the table -- but JTC 1's table, and not the BRM's.

> That position is also curiously at odds with the BSI position in its
> comments that all "proprietary" components of Ecma 376 must be stripped and
> replaced with either ODF-equivalents or where ODF has no equivalent other
> ISO or W3C standards. I.e., if patents are off-topic, why did Brown allow
> the BSI group to develop and submit comments that called for removal of
> proprietary technology?

Good standards build on other standards, and do not reference proprietary technologies unless
absolutely necessary. That's just basic good practice.

> Reading between the lines a bit, he seems to believe
> that Ecma's letter saying that neither it nor any Ecma TC member knows of
> any patents involved that are not licensable under RAND terms somehow trumps
> a comment (I've forgotten which NB submitted it but ran across it in
> skimming the comments last night) that asks for clarification whether the
> Microsoft Open Specification Promise is intended as the RAND terms
> mentioned, arguing (correctly and never rebutted by Microsoft since we
> briefed that issue in the EOOXML Objections document) that the OSP grants no
> rights to implement any Ecma 376 optional features.

Well, that certainly is "reading between the lines" because I am content to leave legal
determinations to lawyers, any haven't paid any attention to Ecma's letter.

> Brown's justification is that only technical issues can be resolved during
> the Ballot Resolution phase.

I don't know what this "phase" is, exactly. To repeat, I wrote that only technical issues could
be resolved at the ballot resolution _meeting_. The result of that meeting can at most be
editing instructions for the text itself which the editor must obey. The meeting cannot instruct
him to reveal (for example) MS patent information, because it's not in his power to do so, and
especially not by editing the text of an ICT standard!

To repeat: countries that have concerns in this area need to pursue them directly with JTC 1.

> But that is so plainly wrong both under the
> Agreement on Technical Barriers to Trade and JTC 1 Directives that I'm left
> scratching my head. Moreover, at the time Ecma submitted Ecma 376 to ISO,
> the Directives version then in effect straightforwardly required that all
> relevant patents be disclosed. Ecma violated the Directives then in effect
> when it submitted only a letter saying that patented technology was
> available for licensing under RAND terms. It's about on a par with the JTC 1
> secretariat deciding to allow Ecma 376 to stay on the fast track despite the
> contradictions, apparently relying on a new draft Directive that had not yet
> even been adopted. Under the Directives then in effect, Microsoft lost the
> contradiction battle and Ecma 376 was required to be rejected, and was
> required to be sent back to Ecma to decide whether it wished to resubmit it
> on a slow track. But none of the NBs that raised contradictions, BSI
> included, objected to the procedural violation despite it being an
> appealable decision to the ISO/IEC Secretariats under the JTC 1 Directives.

You're guessing at what went on between NBs and JTC 1.

Again, IANAL, but my understanding is that the consensus view is that JTC 1 did operate in
accord with the Directives, and that an important distinction in them needs to be understood
between "addressing" and "resolving" issues. I hope to blog on this topic some time.

> It's this kind of rampant flouting of the established procedures, contrasted
> with the generally fine work BSI actually did under Brown's leadership, that
> leaves me scratching my head. I will note, hwoeer, that BSI did not raise
> any of the blatant anti-competitive aspects of Ecma 376 that are mandatory
> NB considerations under the Agreement on Technical Barriers to Trade and
> World Trade Organization Apellate Body decisions interpreting that treaty.
> But in sum, I still have no good picture of where the man is coming from and
> going toward.

I'm attempting to understand the process and apply the rules as best I can. While wondering
quite why I'm doing this...

- Alex.