The Standards Blog

Showdown in Geneva: Majority of Proposed Dispositions Fail to be Approved by a Majority

OpenDocument and OOXML
I have now created a very extensive, indexed BRM Resource Page to hold the many links, press releases, delegate statements and other material that were originally found here.  You can find that extra materials here.


A rather incredible week in Geneva has just ended, bringing to a close the Herculean task assumed by the over 100 delegates from 32 countries that attended the BRM.  That challenge, of course, was how to productively resolve the more than 1,100 comments (after elimination of duplicates) registered by the 87 National Bodies that voted last summer with respect to a specification that itself exceeded 6,000 pages. 

I have spent the week in Geneva, and have spoken with many delegates from many delegations on a daily basis.  Each believed that a body that purports to issue "global open standards" should not impose an obligation of secrecy on how the standards that people must live with are approved on their behalf.  It would be fair to say that, notwithstanding all of the charges and counter charges that have been made leading up to the BRM regarding how National Body votes were taken last summer, how delegations have been selected, and how they have been instructed to act and vote at the BRM, there has been a good faith effort by all to try to achieve a successful result.  The same appears to have held true within delegations, even those that contained representatives of the most opposed parties.

There are two ways in which you may hear the results of the BRM summarized by those that issue statements and press releases in the days to come.  Perhaps inevitably, they are diametrically opposed, as has so often happened in the ODF - OOXML saga to date.  Those results are as follows:

98.4% of the OOXML Proposed Dispositions were approved by a three to two majority at the BRM, validating OOXML

The OOXML Proposed Dispositions were overwhelmingly rejected by the delegations in attendance at the BRM, indicating the inability of OOXML to be adequately addressed within the "Fast Track" process

[Paragraph updated]  In this blog entry, I will explain why the following is the best characterization, and help you read the various press releases and statements that may be made with the benefit of the appropriate context:

Only a very small percentage of the proposed dispositions were discussed in detail, amended and approved by the delegations in attendance at the BRM, indicating the inability of OOXML to be adequately addressed within the "Fast Track" process

It did not take long for the delegates to conclude that it would be impossible to discuss and resolve all of the proposed dispositions, notwithstanding efforts to streamline the process.  After several days, only about 20 to 30 dispositions had been thoroughly discussed and voted upon [updated: the Edited Meeting Notes appear to indicate that a total of 63 resolutions were discussed and individually voted upon].  Approximately 200 dispositions [Updated: the Edited Meeting Notes indicate that the actual number was 126] in the nature of minor editorial corrections (misplaced commas and the like) were also adopted.  Discussion increasingly turned as a result to seeking ways to streamline the process in order to reach a conclusion within the five days allowed.  Those efforts included instructing interested delegates to go off-line to discuss a resolution and come back with proposed compromise language.  However, these efforts proved insufficient to do more than nibble away at the huge number of dispositions remaining.

Acknowledging the impossibility of achieving the stated goal of a BRM (e.g, to carefully review each proposed disposition and reach consensus on an appropriate resolution), a proposal was made on Wednesday to approve all proposed resolutions in a single vote before the end of the BRM, thus nominally "resolving" each remaining proposed disposition without any discussion at all.  It was agreed that this was the only available option, and a written ballot with all of the c. 900 proposed dispositions that had never been discussed was accordingly issued on Thursday.  Each National Body delegation was requested to complete the ballot and return it on Friday.  The alternatives offered were as follows:

1.  Indicate "adopt," "disapprove" or "abstain" after each proposed disposition.

2.  Indicate such a vote on as many proposed dispositions as desired (or none), and vote "accept," "reject" or "abstain" on all of the rest.

It is significant to note that voting to accept all dispositions that were not discussed is a less obvious choice than might be assumed. In fact, few if any of the dispositions that were individually discussed and voted upon during the week were adopted without change.  In other words, adopting a proposed resolution without discussion could result in making OOXML worse, rather than better, because of dependencies.

On Friday, the ballots came back.  Some contained votes on a small number of dispositions and some adopted the default option for all of the listed dispositions.  The final tally (as recorded by participants, and subject to final confirmation) was as follows with respect to the "default" provision that on each vote covered all, or almost all, of the listed proposed dispositions:

                          P Countries Only                All Votes

Approve                              4                                       6

Disapprove                        4                                       4

Refuse to Register
  a Vote                               2                                       4

Abstain                             15                                     18

Total votes cast:            25                                     32


The appropriate rules to be applied to these results are as follows:

1.  Under Directive 9.1.4 under the standing rules of ISO/IEC JTC1, only the votes of "P" members are to be taken into account.  However, Alex Brown, the Convenor, decided in advance, notwithstanding the rules, to allow all attending delegations to vote.

2.  Only "approve" and "disapprove" votes are counted.

We can now turn to the two contentions that you will hear:

98.4% of the OOXML Proposed Dispositions were approved by a three to two majority at the BRM:  The argument is as follows:

 
-  Only "approve" and "disapprove" votes are to be counted.  The rules are the rules

  -  That said, ignore the standing JTC1 rule that only P votes count

  -  Ignore the protests and abstentions, regardless of the fact that, together with the "disapprove" votes, they represent more than 80% of the delegations

  -  Ignore the fact that only c. 20 [Updated:  20 - 3o] substitutions out of c. 900 substantive dispositions, were actually discussed

The OOXML Proposed Dispositions were overwhelmingly rejected by the delegations in attendance at the BRM

  The purpose of the adoption process is to ensure that a quality specification is approved

  -  The purpose of the process is also to achieve consensus on the final result, so that a finally approved specification is regarded as being appropriate, useful and desirable around the world

It is clear to me that the first conclusion is, at best, technically accurate, and even that conclusion assumes that the decision to allow O members to vote was justified.  The better conclusion is that despite the good faith efforts of all concerned and their willingness to see this process through to its conclusion, it has proven to be impossible for as large and poorly prepared a specification as this to be properly addressed via the "Fast Track" process. 

[Updated:  it would be inaccurate to characterize choosing "abstain" as a default position on the last c. 900 dispositions as a rejection of those dispositions upon which the NB in question did not specifically vote.  "No opinion" and acquiescence would be a fair characterization.  The significant conclusion to take away is that while the National Bodies had a chance to stand up and say what there major concerns were, there was no opportunity for the great majority of the dispositions involved to be discussed in detail, and to be amended as necessary before being approved or disapproved.]

There are a number of conclusions that can be drawn from this result:

1.  As many have contended, the Fast Track process was a totally inappropriate process for Microsoft and Ecma to have adopted for OOXML

2.  OOXML has not been adequately addressed within that process to be entitled to final adoption

3.  It would be inappropriate for the ISO/IEC members to approve the adoption of OOXML in the thirty day voting period ahead

Many, many, people around the world have tried very hard to make the OOXML adoption process work.  It is very unfortunate that they were put to this predictably unsuccessful result through the self-interest of a single vendor taking advantage of a permissive process that was never intended to be abused in this fashion.  It would be highly inappropriate to compound this error by approving a clearly unfinished specification in the voting period ahead. 

To paraphrase a former First Lady, it's time to "Just say No" to OOXML.


For further blog entries on ODF and OOXML, click here

sign up for a free subscription to Standards Today today!

Comments

Wasn't Frank one of only three votes in the final INCITS executive board that voted against the US approval of OOXML  (12 for approval and Frank together with IBM, and Oracle voting no)

It can't be surprising to anyone that he would react like this

My recollection is that Frank is a member of the Executive Committee of INCITS that voted on OOXML last summer.  I don't recollect at this point, though, how he voted, although it would not surprise me if he was one of the "no" votes.

I think that what is remarkable, though, is that the US delegation to the BRM - which included Microsoft employees and supporters as well as folks like Rob Weir - agreed to vote "no" as its default on the 900 dispositions.  They should really be given credit for this, and it's consistent with the uniform accounts of the meeting that say that everyone displayed an unusual degree of cooperation in trying to process OOXML under impossilbe conditions.  My only complaint is that those that adopted "abstain" as a default tried too hard.

My personal choice for how the BRM should have ended would have been with a vote that it was impossible to make any useful statement about the 900 batch of dispositions at all, and calling for further review to be taken before a vote.  I see in Rob Weir's account that something along these lines was considered earlier in the week, but (unfortunately) not adopted as a plan of action.

  -  Andy

Permalink
"9 Voting
9.1 General
9.1.4 In a meeting, except as otherwise specified in these directives,
questions are decided by a majority of the votes cast at the meeting
by *P-members* expressing either approval or disapproval."

"9.1.10 At all levels of voting, if more than 50% of the P-members
*have not voted*, the vote will have *failed*. Late votes shall not be
counted. No extensions shall be granted."

  --Dario

Dario, SC34 issued an FAQ about the conduct of the BRM on Nov 15, http://www.itscj.ipsj.or.jp/sc34/open/0932.htm - check para 6.8: "If votes are taken at the BRM, who votes? Those present."

Paragraphs 6.5 and 6.6 (rather confusingly, for a document titled to be about the BRM) talk about O or P status remaining fixed as at the September vote, but it's in a section titled "Voting" which seems to describe the ballot process as a whole. So for votes at the BRM, it seems P or O membership is as per September,  but as anyone at the meeting can vote, membership status is therefore irrelevant. I think...

>Dario, SC34 issued an FAQ about the conduct of the BRM on Nov 15, http://www.itscj.ipsj.or.jp/sc34/open/0932.htm - check para 6.8: "If votes are taken at the BRM, who votes? >Those present."
>
>Paragraphs 6.5 and 6.6 (rather confusingly, for a document titled to be about the BRM) talk about O or P status remaining fixed as at the September vote, but it's in a section titled >"Voting" which seems to describe the ballot process as a whole. So for votes at the BRM, it seems P or O membership is as per September,  but as anyone at the meeting can >vote, membership status is therefore irrelevant. I think...

I can't follow your argument.
Let's us check what 6.5 really says
"6.5  What voting status will NBs have?
    For voting at the BRM, and in the ongoing DIS 29500 ballot the voting status (either “P” member or “O” member) is fixed as per the result of the 2 September letter ballot."

This FAQ entry does without doubt indicate that P and O status matter to the BRM since their status is fixed as per the 2 Sep latter ballot.

6.8 Futher explains that all attending can vote...but if O and P status matter then the only logical conclusion is that the normal directives for O and P status apply. In the september ballot both O and P participants could vote...but the rules give these votes different meaning. Going by those rules the BRM letter ballot failed.

It could happen that ISO has an extra document that makes special rules for this particular BRM, but why have this then not having been published when people question the voting rules? That no objections was heard about this at the BRM could easily be explained by the reports of the board just counting the number of objections instead of hearding them before they proceeded.

Alexander Brown probably thinks he followed procedure when he equalled O and P status...the things that matter is if the paper work match his oppinion.

Permalink

The following was submitted at Jason Matusow blog , responding his post "The Open XML Ballot Resolution Meeting (BRM) Was An Unqualified Success" ( http://blogs.msdn.com/jasonmatusow/archive/2008/02/29/the-open-xml-ballot-resolution-meeting-brm-was-an-unqualified-success.aspx )

But never saw the light ( censored by Jason?, why?  )
-----

"The DIS 29500 ballot resolution meeting (BRM) finished up in Geneva today and was an unqualified success by any measure. "

please specify which are the measures that compose the "any measure" set

"A few things need to be kept top of mind as national bodies solidify their position within the next 30 days:"

thanks to worry about NB mind, let see what are the things ...

"A BRM is about technical work on the remaining open issues most important to participating national standards bodies"

first of all: BRM has gone... why are you telling NBs what is supposed they must know *before* the BRM. I believe that you are underestimating them.

This BRM was about discuss 3500 comments/observations detected during a few months of quick review of +6000 pages of a DIS, which were responsible of a "disapproval" of DIS 29500 in the september ballot.

"Many issues raised by national bodies can be addressed in advance of the BRM through the proposed dispositions of the submitting organization and, in some cases, discussions between the national bodies and the submitter.  This happened with Open  XML, where national bodies identified many issues that were addressed to their  satisfaction before the BRM started."

if this were the case, why were there 18 abstentions in the BRM vote tally? ( http://www.consortiuminfo.org/standardsblog/article.php?story=20080229055319727 )

"The BRM is an opportunity to discuss the remaining issues of importance to national bodies. "

according to reports ( Frank Farance, HOD of USA, Tim Bray, delegate from Canada and other mentioned in Andy post ), only 20 comments have been discussed. The amount of work required to address concerns with this proposed ISO standard was way beyond what such a large group could meaningfully address in the time allotted.

"   * The convener, Alex Brown, with the help of ISO/IEC and SC34 officers,  ran a very successful BRM."

please define "successful"

"   * During the meeting, each delegation was given the opportunity to identify those  issues most important to them thus defining the scope of work for the BRM.  "

let each delegation affirm this ( speaking or voting in one month ), thank you

"The BRM focused on building consensus on significant remaining technical issues and, in many cases, resulted in modifications to the proposed dispositions to refine and
 improve them."

let each delegation affirm this ( speaking or voting in one month ), thank you

"   * By the end of the BRM, national bodies were able to consider for their approval  each proposed disposition. 
The vast majority of those proposed dispositions were adopted, resulting in a better specification that will better meet the interests of national bodies and the broader community."

how many delegations voted for approve this "vast majority" of proposed dispositions? six over thirty two for example?  ( http://www.consortiuminfo.org/standardsblog/article.php?story=20080229055319727 )

"Extensive steps were taken within the last year to improve Open XML through  work with national standards bodies (and their participants) within the  inclusive framework of ISO/IEC  processes."

let each delegation affirm this ( speaking or voting in one month ), thank you

"   * The Project Editor and Ecma TC-45 reached out to NSBs for ongoing input to the  disposition process. Their valuable input was influential in the generation of the  dispositions and that is why such overwhelmingly positive consensus was achieved
 during the BRM. "

i would have preferred that ECMA TC 45 ( and Microsoft specially ) did their homework previous to submitting DIS 29500, this is the "spirit" of fast-tracking: to fast-review a mature draft standard , not to make rushed, massive and not reviewed changes.

"   * Conference calls, meetings, progress reports, early postings of dispositions, the  full report on Jan 14...all of this was done as the groundwork for the BRM. And this  was on top of the preceding 7 months of extensive technical engagement by the working  groups and committees within the NSBs."

i see this "progress reports" as an "artifact" of the work. Do you think that NBs around the world have *all* the time to work with you to correct this DIS?

What is needed is that you ( ECMA and Microsoft ) submit a "final" text , so NB can review it and finally disaprove ( or approve it ). Still today, this text is not available at http://www.ecma-international.org/dis29500_brm/

" Because of this, there were no "surprise" issue  to be dealt with."

may be the surprise were the +6000 pages of the initial draft and the +2000 page of the dispositions and the scarce time to review it ( see the following Tim Bray [ one of the fathers of XML ] post and search the word "frightening" :  http://www.tbray.org/ongoing/When/200x/2008/02/29/BRM-narrative )

" And without question, the specification was improved based on the diligent feedback of  NSBs from around the world. "

i would have preferred that ECMA TC 45 ( and Microsoft specially ) did their homework previous to submitting DIS 29500, this is the "spirit" of fast-tracking: to fast-review a mature draft standard , not to make rushed, massive and not reviewed changes.

"A BRM is successful if it produces technical improvements to the specification to  address comments raised during the ballot phase."

without rushing and with a reasonable timeframe, keeping the quality of the deliverables

"    * There were no "surprises" or "new comments" during this process. Every issue addressed was the result of the past 2+ years of work on the specification, and in  particular the past 5 months of intensive work leading into the BRM."

not agree, the last "draft" was published five days ago ( 22/02/2008 ) and still is provisional

"    * During the course of the meeting, much effort was put forward in order to come to consensus on those issues that were the most heavily discussed over the duration of  the past year.  The types of issues discussed during the BRM are represented here. "

did NB reach consensus in this heavy issues?

"   * The changes adopted (and denied) were based on consensus among all National  Bodies. This is exactly how BRMs are meant to function. "

how many delegations composed this "consensus" six over thirty two NBs for example?  ( http://www.consortiuminfo.org/standardsblog/article.php?story=20080229055319727 )

"I'll say it again - by any rational measure of ISO/IEC JTC 1 ballot resolution meetings, this one was a complete success."

do you think that repeating things can replace based argumentation?


"ISO/IEC standards are not only technically sound, but they should also be relevant to the marketplace."

not agree. In one month, NBs will have to answer ( with a vote ) the following question ( this was put it clear by Alex Brown, BRM convenor, on many blog posts ):

'Has DIS 29500 the technical merits to become an ISO international standard?'

"   * DIS 29500, as improved through the rigorous review of the past year and the  decisions made by delegations during the BRM, is a specific ation that meets both bars of technical quality and marketplace relevance."

too much PR here, i will no comment on this paragraph, sorry

"    * Independent implementations of the specification are already available on most  major operating systems platforms and in hundreds of applications. The statement that  Open XML is about a single vendor is specious and empirically false."

please provide % of conformance of each of this implementations

"    * Open XML has brought more attention to, and interest in, international  standardization than any specification in the history of the ICT industry."

could be true. Do you know *all* the specifications in the history of the ICT industry? i admire your knowledge

" The reason for this is simple - greater openness in all document formats (not just Open XML) is a good thing for everyone.
 There is general recognition that there will be broad adoption of this format around the world. Open XML delivers on that promise and is part of the rich ecosystem of open ??> document formats that are driving this issue forward."

too much PR here, i will no comment on this paragraph, sorry

"At the end of the day, customers should be able to choose the format(s) that best meet their needs and should not be told which technology to use.  Open XML, as improved through the hard work of national bodies over the past year, is an attractive
 alternative for them."

this has nothing to do with DIS 29500 fast-tracking process. I repeat:

In one month, NBs will have to answer ( with a vote ) the following question ( this was put it clear by Alex Brown, BRM convenor, on many blog posts ):

'Has DIS 29500 the technical merits to become an ISO international standard?'

"There was an unprecedented number of delegations from national standards bodies that came to Geneva and participated in the BRM. I have the utmost respect for the contributions from all of the national bodies (P-members and O-members alike). "

Would Microsoft send a check to reward all the work of this national standards bodies ?

"The result of this week's discussions, by any reasonable measure, has greatly improved  the specification and produced a great result. "

let each delegation affirm this ( speaking or voting in one month ), thank you

"The BRM was a complete success "

please define success.

"congratulations to all who were involved with it."

thank you

   --Orlando Marcello

Permalink

Andy hi

To follow up on our exchange above ...

I must thank you for your kind words -- but this is not personal, or a question of friendliness (I’m sure you are a lovely chap), but a question of accuracy and ethics.

Your headline has now transitioned from being "not even wrong" to "wrong". If you want to fix it you should remove the words "fail to". However, since this is not then a very on-message headline for you I suggest maybe you should have something like "OOXML still in flux as clock ticks down" or "BRM performs emergency surgery on OOXML in desperate rescue attempt" or some such.

I think it is wrong for you to claim your original headline was some kind of necessary counterbalance to Jason Matusow's: his was predictably on-message (from the MS POV), yours was (and is) factually misleading.

Also, by my records Charles Schulz was not a BRM delegate as you categorize him.

You then raise several points about the adequacy of the Fast Track process. Fair enough; no comment. On the particular questions about in-meeting voting I can tell you:

  • Yes, it was a good idea to take votes (congratulations to the BRM on wisely choosing this route)
  • Yes, it was within my, and the meeting's, powers to allow it
  • Yes, what happened was fully in accord with the JTC 1 Directives (O-member voting and all)

OBVIOUSLY (given the red hot controversy here) voting procedure was discussed in minute detail, and decided, in consultation with ITTF before the BRM started. I (as somebody primarily used to SC participation, rather than fast tracking) had some un-learning to do, and I think some other commentators do too. If a country has a complaint it can appeal formally -- that (rather than wittering to the press) is the correct way to do it.

You then turn to the question of what votes "mean". Rather than venturing further into the mental maze you have made for yourself, let's take an actual example and work it through.

Consider a very simple and uncontroversial Response, no. 637. This concerned a comment raised by AFNOR (France) that noted an "extraneous square bracket at the end of [a] line". AFNOR's suggested solution was "remove the extraneous square bracket."

Ecma's proposed response to this was to edit the text to remove the square bracket.

Now, this is just a PROPOSED resolution; in order to make it into the DIS text (unchanged since Jan 2007) it HAS TO be blessed by the BRM. BRM delegations can take one of three positions on Ecma's proposed response. They can:
  • approve (broadly, "yes, we want Ecma to do this")
  • disapprove (broadly, "we do NOT want Ecma to do this")
  • abstain (broadly, "we don't know/care")

A key point here is that if no decision is taken, nothing happens. In other words, if no decision is taken the text stays in its original state.

Now imagine you are a NB considering voting on Response 637 -- Andy, what would you have voted for in this case? And would you have been happy not to be able to take any decision at all because of time constraints?

Now multiply this process by 800 and you'll know what voting was. For added realism, try staying awake all night, and wearing this: http://www.cafepress.com/freesociety.234283639 (thanks to the Portuguese delegation).

The voting form also had boxes for registering an overall position, purely as a labour-saving device. So if (for example) you as an NB approved 700 responses and disapproved 100, you could simply record an overall position of "approve" and then mark the 100 disapproved comments as "exceptions" to your overall vote. To save typing, many NBs (as Inigo Surguy has noted above) simply recorded an overall position of "abstain" and then recorded the exceptions to that position. Your reading of significance into the way this overall position was recorded is about as sensible as reading significance into whether NBs user uppercase of lowercase “X” characters on their voting form!

Finally, I deplore your mention of the Nuremburg Defence (“only following orders”) in relation to my role as convenor. Invoking the Holocaust (however indirectly) as a suitable parallel to a document format standardization project shows a total failure of perspective.

- Alex.


"OBVIOUSLY (given the red hot controversy here) voting procedure was discussed in minute detail, and decided, in consultation with ITTF before the BRM started."

Why weren't the ballot papers given to the NBs days before the BRM so that they could fill them in in a less rushed manner?

This would have allowed significantly more time to discuss specifics at the meeting.
(Since it has been reported that almost a day was spent discussing the mechanics of the ballot).

Alex, the reply I made to a previous post is more appropriate here so I'll restate it: "In all the different (political) meetings I've ever been in as either a participant or a chairman, it was customary for the chairman to draw up the agenda for a meeting according to the following rules: 1) schedule proposals on which the decision is a formality (to do away with stuff that can be implemented) and decide on them right away 2) then do a tally of the remaining proposals and draw up a schedule on what needs to be discussed 3) suggest (controversial) proposals where a large number of the participants indicate their urge to be heard to be taken off-line, either to resolve them by the end of the meeting or at a later date. 4) schedule proposals with little comments to be handled first. Alex, as far as I can tell the way you approached this BRM is the exact opposite of what I described above and (as you indicate yourself) in the end leaves a lot of room for interpreting the meaning of the votes on the comments not being discussed. With 20-20 hindsight, would you have chosen a different approach to the BRM, for example with the (universally) tried and tested method I described here?" The example you mention about the loose bracket AFNOR stumbled upon to me is a prime example of a proposal on which the decision is a formality. Would it have been too much work to lump the resulting dispositions together and be done with them at the beginning of the BRM? For that matter, how many BRMs have there been in the (recorded) history of ISO where more than 1000 dispositions needed to be addressed? Doesn't that (presumably) low number flag this BRM as something special, where special measures need to be taken in order to keep it manageable in order to send a CLEAR signal to all NBs who need to vote in the next days to come?

In fact, several of these approaches were taken.  A batch of c. 200 "easies" (typos and the like) were adopted as a group (that's where most of the 20% of resolved issues can be found).  Many issues were also taken off line just as you suggest.  Rob Weir's post describes this in some detail.

So while the final result may still not have been satisfactory in the end, in fact Alex did many of the things that you are suggesting.

  -  Andy

To do away beforehand with dispositions that common sense dictates to be non-controversial would have shed some light on the amount of time needed to resolve the remaining issues, and taking the most controversial dispositions off-line immediately would have freed up more time for the dispositions where little discussion was expected to be necessary.

From http://www.itscj.ipsj.or.jp/sc34/open/0933draft-rev2.htm I got the distinct impression the agenda wasn't set up with that prioritisation in mind (and that's what I'm suggesting):
"The convenor thanks those NBs who have informally indicated to him their highest-priority comments and responses. Taking the variety of these into account, and the partial nature of the response, the convenor has decided it would be counter-productive to prescribe an order of discussion in the agenda. Instead NBs will be invited to speak by the convenor in the meeting sessions, and the convenor will thus ensure NBs can enjoy fair representation over the course of the meeting by granting them permission to speak in alphabetical order, making as many passes over the delegations as time allows."

Of course that's a draft agenda, so until the audio recordings are released for public scrutiny we'll have a hard time finding out, won't we?

The abstain meaning is wrong.

Many NB use abstain to indicate we can't vote for this but we don't want to embarrass anyone over this either. 

Now that you understand that let's look at those votes again 6 for, 4 against, majority abstain.  Vote fails; didn't obtain consensus for the changes. 

BRM should report out that it worked a small fraction of the issues, managed to get a number of improvements voted on to make to DIS 29500.  However, on an overall vote the BRM could not recommend the results to be standardized.   Based on progress made at the meeting it would take at least a year to get to the point where a majority of the members would vote in approval.

Based on this, we would suggest that no additional vote be taken, and the issue referred to the maintainers/enhancers (OASIS Open Document Format for Office Applications (OpenDocument) TC) of ISO/IEC 26300 for consolidation in that standard. 

If MS had done this in the first place no one would be in this current predicament.  MS knew this but chose to attempt to take a short-cut through ECMA.  Well, that was a bad decision, now they have to go back do things the right way.  Too bad.]  Further, ECMA needs to be stripped of their fast path in ISO, until they fix their standards process so that it represents all sides (not just some big producers).  ECMA should have coordinated with OASIS on this standard, their failure to do so is reason enough to strip them in this case (we can cite that this is not the first time for ECMA to have serious problems with a standard at ISO).

Minor details aside, this is exactly what I was suggesting:
BRM should report out that it worked a small fraction of the issues, managed to get a number of improvements voted on to make to DIS 29500.  However, on an overall vote the BRM could not recommend the results to be standardized.   Based on progress made at the meeting it would take at least a year to get to the point where a majority of the members would vote in approval.

Based on this, we would suggest that no additional vote be taken, and the issue referred to the maintainers/enhancers (OASIS Open Document Format for Office Applications (OpenDocument) TC) of ISO/IEC 26300 for consolidation in that standard. 
This would have been, to my mind, much more useful and harder to be misrepresented.

  -  Andy

I don't see it being addressed anywhere, was a single resolution for the following agenda item voted on?  (In other words was a single closing resolution made and voted on to prepare the revised text according to the resolutions made during the BRM?  If not there is no new text and no vote should be taken - fast track is over, as BRM did not come to a resolution.)

The output of the Ballot Resolution Meeting shall be a single resolution that instructs the Project Editor to prepare a revised text. The resolution shall list the changes that the meeting has decided on.

I also asked Robert Weir and replied on his blog stating "There was no vote on that final resolution."

I think I'd like to meet the person who the meeting adjourned before that issue came up.  What was he thinking?!!  Without that issue being resolved, I don't think the BRM voted to issue a revised text.  As I see it, no revised text means no resolution was achieved at the BRM.  No resolution means exit fast track procedures and no voting is needed,  DIS29500 will have to go some other route to standardization.

"Finally, I deplore your mention of the Nuremburg Defence (“only following orders”) in relation to my role as convenor. Invoking the Holocaust (however indirectly) as a suitable parallel to a document format standardization project shows a total failure of perspective."

Yes, it does. It is important to be responsible.

I fully understand that the adoption of the ECMA dispositions is a matter of good pragmatism.

Permalink
Alex,

Thanks once again  for responding.  I think that from my point of view we're probably getting close to the "agree to  disagree" point (assuming that you're willing to cut me that much slack), so let me start with the central point where I don't think that one of use is likely to convince the other, as there may be some benefit to that.  I'll then go on to your individual points.

The main point at the 30,000 foot level is that I don't think that the implications of allowing the Fast Track process to be used on such a large and insufficiently prepared format  can be separated from the impact that this decision had on every stage of the process, and most especially on the BRM.   It was this decision that placed you and everyone else in the BRM in the impossible position that you found yourself in. 

Taking this down to ground level:  I have just read Rob Weir's piece, and its recital of "no time" decisions notwithstanding several objecting hands in the air, the "newbies" from NBs struggling with language issues and what was going on, the inability of NB's to advance objections (I have personally received a number of emails from delegates flagging further proposals that they were unable to get heard), the accounts I have personally received from delegations saying that it was impossible to go more than a small way through such a long list in so little time, and so on, all of which I heard first hand from people when I was in Geneva as well.

I cannot square this with any confidence that the decisions made under the ballot will be good technical decisions.  While I do not question the right of those in the room to have made the decision, I do not accept that this decision may not be questioned by whoever is troubled by it.  I have written before and I am sure will be writing again that I think that the integrity of the de jure standard setting process is taking a terrible beating as a result of the OOXML experience - and this in an organization that created the ISO 9000 series!

I am, of course, also particularly disturbed by the lack of transparency, which, absent stories like my own and the willingness of delegates to speak could lead people to have greater confidence in the results than I believe is warranted.  As you know, there will be tremendous pressure brought to bear on NBs all around the world.  Unless they have more public facts to point to, they will be defenseless to resist pressure to approve OOXML, whether they are as disturbed with quality issues or not.

Alex, I understand and respect your dedication to accomplish the job to which you were appointed.  I don't think that it would be fair for me to expect that you, as Convenor, should have attempted to declare a failure of the BRM to complete its task.  But I do think that the delegates collectively would have better served the process had they voted earlier in the process to declare that they could not complete their job.  That's where you and I will certainly differ, so on this point I think we need to  agree to disagree, as each of us is entitled to his opinion.

Now, on to your individual points.

Alex:  Your headline has now transitioned from being "not even wrong" to "wrong". If you want to fix it you should remove the words "fail to". However, since this is not then a very on-message headline for you I suggest maybe you should have something like "OOXML still in flux as clock ticks down" or "BRM performs emergency surgery on OOXML in desperate rescue attempt" or some such.

Andy:  If you and I were having this discussion before I posted the blog entry, I would have had no particular objection to "OOXML still in flux as clock ticks down," but at this point I think that it isn't honest for me to keep tinkering with a title that has already been read and linked to by so many people.  I think the text and comments together will give a clear picture.  I also think that a really accurate title that we could both agree on would have to be more like the following, but it would be a little long:
ODF Advocate contends that approval of most OOXML dispositions by a single ballot with 18 out of 32 NBs having too little time to perform an adequate view  and abstained with respect to anywhere from some to all of the more than 900 listed  proposed dispositions does not constitute  "Majority Approval" of those  dispositions for purposes of directing that they be  taken into account in preparing the final draft of OOXML for final March Vote.
That said, if we can agree on the above, I will add it in at the top of the text by way of an update and indicate our consensus on it.

Alex:  I think it is wrong for you to claim your original headline was some kind of necessary counterbalance to Jason Matusow's: his was predictably on-message (from the MS POV), yours was (and is) factually misleading.

Andy:  I think that we're spending too many words on the title, but that's just my opinion.  When someone reads the text, I believe that the meaning of the title becomes clear., and I doubt that many people will read the title at my blog and not the article.  Note also that the blog entry you are talking about, while based on data, also  clearly was as much an editorial as a fact piece, and this would hardly be a title that would be inappropriate on an editorial page where projecting message and point of view is the purpose of a title.  All  that said, I accept that you didn't like it, and I have already clarified it. 

Alex:  Also, by my records Charles Schulz was not a BRM delegate as you categorize him.

Andy:  Right you  are.  I corrected this not long before you posted your comment.

Alex:  You then raise several points about the adequacy of the Fast Track process. Fair enough; no comment. On the particular questions about in-meeting voting I can tell you:

  • Yes, it was a good idea to take votes (congratulations to the BRM on wisely choosing this route)
  • Yes, it was within my, and the meeting's, powers to allow it
  • Yes, what happened was fully in accord with the JTC 1 Directives (O-member voting and all)

OBVIOUSLY (given the red hot controversy here) voting procedure was discussed in minute detail, and decided, in consultation with ITTF before the BRM started. I (as somebody primarily used to SC participation, rather than fast tracking) had some un-learning to do, and I think some other commentators do too. If a country has a complaint it can appeal formally -- that (rather than wittering to the press) is the correct way to do it.

Andy:  I don't have any concerns over whether the 900 vote was within the rules or not.  Others may, but that's not the issue to me, so I'm fine with all of this.  As indicated earlier,  the big issue for me isn't whether deciding to vote on all 900 was a good idea.  Clearly, those in the room thought it was, and that's fine.  But it's also perfectly fine for those outside the meeting, as with any other political process, to state that they think that it was a bad decision. 

Alex:  You then turn to the question of what votes "mean". Rather than venturing further into the mental maze you have made for yourself, let's take an actual example and work it through.

Consider a very simple and uncontroversial Response, no. 637. This concerned a comment raised by AFNOR (France) that noted an "extraneous square bracket at the end of [a] line". AFNOR's suggested solution was "remove the extraneous square bracket."

Ecma's proposed response to this was to edit the text to remove the square bracket.

Now, this is just a PROPOSED resolution; in order to make it into the DIS text (unchanged since Jan 2007) it HAS TO be blessed by the BRM. BRM delegations can take one of three positions on Ecma's proposed response. They can:
  • approve (broadly, "yes, we want Ecma to do this")
  • disapprove (broadly, "we do NOT want Ecma to do this")
  • abstain (broadly, "we don't know/care")

A key point here is that if no decision is taken, nothing happens. In other words, if no decision is taken the text stays in its original state.

Now imagine you are a NB considering voting on Response 637 -- Andy, what would you have voted for in this case? And would you have been happy not to be able to take any decision at all because of time constraints?

Now multiply this process by 800 and you'll know what voting was. For added realism, try staying awake all night, and wearing this: http://www.cafepress.com/freesociety.234283639 (thanks to the Portuguese delegation).

The voting form also had boxes for registering an overall position, purely as a labour-saving device. So if (for example) you as an NB approved 700 responses and disapproved 100, you could simply record an overall position of "approve" and then mark the 100 disapproved comments as "exceptions" to your overall vote. To save typing, many NBs (as Inigo Surguy has noted above) simply recorded an overall position of "abstain" and then recorded the exceptions to that position. Your reading of significance into the way this overall position was recorded is about as sensible as reading significance into whether NBs user uppercase of lowercase “X” characters on their voting form!

Andy:  First, while I appreciate you going into this much detail (which is helpful), I think that we start going sideways when we get down to this level of detail.  I've had email from delegates who said that they didn't have close to enough time to get close to getting half way done, making these finer distinctions, in my view, often academic in the breach.  If the ballot had gone out a week before the BRM, I agree that all of this would be more relevant.

Alex:  Finally, I deplore your mention of the Nuremburg Defence (“only following orders”) in relation to my role as convenor. Invoking the Holocaust (however indirectly) as a suitable parallel to a document format standardization project shows a total failure of perspective.

Andy:  Noted.  The phrase is a useful shorthand for saying that one puts the appointed task ahead of the consequences, and points to the rules to show why those consequences should not be taken into account.  Everyone in the course of their life is likely to be placed in a situation where they can choose to follow the rules, or to refuse to do so when they think that the situation at hand is a square peg to a round hole of rules.  Different people will make different decisions when they find themselves in such a situation.   It did not occur to me yesterday,  but there is another phrase, this time from the 1960's, that I could have used instead:  "If you're not part of the solution, you're part of the problem." 

In my view, the BRM delegates lost sight of the possible consequences, repeatedly saying, "This is not the venue where other than technical details can be discussed," and then even failed to take the consequences into account in the area of its appointed purview.  The result is that it lent its authority to the result by adopting a mechanism that was - in my opinion - not a responsible decision, although undisputably well-intentioned.  You are asking me to take responsibility for my decisions in writing my (equally well-intentioned) blog entry, and I am simply doing the same with respect to the decision of the delegates.

Once again,  I do  thank you  for taking the time to post your comments here, and I do think that it will help concerned individuals to have more insight into the decision making that went on inside the room.  While I respect that you did not like the way in which our dialogue came to happen, I hope you will agree that where we have ended up may have a useful purpose in supplementing public understanding of a stressful process.

Best regards,

  -  Andy

Permalink
I can sympathize with Alex, having been the chair of JEDEC's JC-16 committee for eight years.  JEDEC, please note, is one of the "industry consortia" that generally claim less rigorous standards than ISO does.  I can, therefore, offer some comparisons that may or may not be germane.

Quite a few of our processes are recognizably related, such as the four-way ballot options of "yes,yes with comments, no with comments, abstain."  In particular, I will focus on our post-ballot meeting protocol.  Rather than the ISO process where a passing letter ballot is automatically adopted, we always hold a post-ballot meeting to finalize the process and pass successful ballots to the Board of Directors for process review.  In that meeting, our rules require that all comments must be addressed.  Upon review of the voting and comments, the Committee votes on resolution of the ballot.  Simple editorial corrections (e.g. duplicate right bracket) may be adopted only by unanimous agreement that they are, in fact, only editorial.  The corrected ballot is then voted on to be passed to the BOD -- or not.  To be clear, a letter ballot that fails by the numbers is dead, but even one which passes may still fail in open committee based on the comments made during balloting.

Again, all comments made during balloting must be addressed.  Unanimity is, of course, not always possible; in fact, in my experience it is very rare.  However, even a "losing" position deserves a fair hearing.  Beyond fairness, sometimes the minority is just plain right and we are, after all, professionals.  I have seen, on more than one occasion, a ballot accepted by a ballot of 40:1 or more killed dead by one comment which the Committee agreed outweighed the rest.  In the end, our rules require that every comment requires a hearing and until it has its fair hearing and consideration the balloted material cannot proceed without consensus that the comment has been addressed.  For what it's worth, our usual committee vote asks if any are opposed and only if there is opposition is a count taken.

My understanding is that ISO rules have similar "must be considered" requirements.  If so,  we interpret them differently because although I have seen attempts to short-circuit the process in JEDEC committees they never succeeded.  The fact that ISO requires so much less of itself does not, in my opinion, speak at all well of it as an organization.

D. C. Sessions
lumbercartel.com is my domain; dcs is my account

Reference: JEDEC Manual of Procedure, JM-21: http://www.jedec.org/Home/manuals/JM21M.pdf

D.C.,

I really appreciate people with your experience leaving comments like this.  It's very difficult for people who haven't been part of the process to know what's reasonable and what isn't, taking all relevant factors into account, or even what normative practices are.  It's very helpful when people with solid experience in the trenches share their knowledge with everyone else.

  -  Andy

Permalink
If I understand it correctly, we have the following:
- The ballots for the mass vote saved for the last day were prepared ahead of time, Yes?
- Some people, but not the NB HoDs, knew of this plan and sanctioned the pre planned rule change, Yes?
- The voting choices created for this one-time event allowed "Not Yes" votes to be split three ways, Yes?
- The endgame was presented to the "participants", at the last possible moment, w/ no discussion, Yes?
- 800 + issues were "resolved" by "consensus" as the one way to vote "yes" got more votes then any individual way to vote "no"


Given the above (assuming its true) what information is available to Citizens to understand how the whole meeting was not well planned for for exactly this "80% issues resolved PR outcome"?

Who designed the ballot? When? Why? Who "approved"?

When early-in-the-week procedural questions were raised, or attempted to be raised, did those in charge of the rules know that the endgame was already defined?

What is their relationship to ECMA & Microsoft?

--tom

Tom, you're starting from the assumption that this was not the result of committee discussion.  Too many attendees have told us that they did discuss this approach and it was decided, by consensus, that it was the least evil of a bad situation.

There's enough grief going around here without adding gratuitous tin-foil hat stuff.

Tom, if you click on the delegate links collected at the end of the blog entry you'll find several accounts (Rob Weir's, as I recall, covers this in pretty good detail) that track the discussion on this point.  "Batch voting" was proposed by Alex Brown at his blog in advance as a possibility, but I have no information or reason to think it went beyond that.

  -  Andy

> "Batch voting" was proposed by Alex Brown at his blog in advance as a possibility, but I have no information or reason to think it went beyond that.

Andy, I suppose you are the same Andy that usually post under "Admin", please remember Alex has posted this admission in a comment on this very blog.
---------------------

You then raise several points about the adequacy of the Fast Track process. Fair enough; no comment. On the particular questions about in-meeting voting I can tell you:

  • Yes, it was a good idea to take votes (congratulations to the BRM on wisely choosing this route)
  • Yes, it was within my, and the meeting's, powers to allow it
  • Yes, what happened was fully in accord with the JTC 1 Directives (O-member voting and all)
OBVIOUSLY (given the red hot controversy here) voting procedure was discussed in minute detail, and decided, in consultation with ITTF before the BRM started.

Yes, I am the same "Andy" that has signed every "Admin" comment at this blog (a/k/a my blog, where I am also the Admin).  I'm not sure what your point is, though.  They chose the "route," not came up with the idea, which Alex had discussed earlier at his blog, where you can check it out.

  -  Andy

Thanks Andy.  I really like your blog, and this article is very informative.

I am the same anonymous poster as in the grand-parent post. My point was to support the original poster that suggested Alex Brown and ITTF could have been more forthcoming about the paper voting at the beginning of the BRM instead of waiting until Wednesday. They knew that was in the plan having discussed it in details and the lack of time issue was very predictable. Then the NBs delegation could have had more time to come up with some more palatable alternative. The way things were managed, the NBs were put in an awkward position. If they complain, they are being told "this was your choice, bear with it" to boot.

Yes, I see your point.  It does seem strange and unfortunate that more concrete planning wasn't done in advance - and particularly because if it had, the realization might have hit home that a one-week BRM was a hopeless exercise and not worth attempting at all, unless many more days, perhaps spread over some months, were dedicated to the effort instead.

  -  Andy

I am the same anonymous again.

Andy you must be a kind man to bring up such a charitable explanation.

My opinion  is the problem was obvious all along and there is no way Alex and ITTS could have miss it short of gross incompetence (which I don't believe). For some reason they think their duty is to make this pig fly no matter what and to achieve this goal, they pretend it is not a pig and it did fly just fine. The lack of minutes, recording and outside attendance serves no useful purpose other than making sure there is no record that can contradict this fiction. I think everything was intentional and went according to plan but for one thing. They didn't anticipate information leaking through bloggers.

I too have noticed that Andy will write based on making the most forgiving of assumptions, and I both appreciate giving people the benefit of the doubt, and giving the reader what they need to draw their own conclusions :-).

So, a thought experiment: Assuming the very best in all people involved, and lets assume no rule changes where made before the BRM.

Now, to this interested observer, the rules appear to allow for "No consensus reached" as an outcome, or:
 "The NB Consensus is that a one week BRM fine tuning process is
  insufficient by 45 times ( 900 / 20) for this proposal, and that is
  a clear indication that the details are not ready (regardless of
  issues of duplication), therefore Fast Track is inappropriate"

(apologies for any off base assumptions baked into the following questions)
Why was something like the above not given some sort of hearing/vote/discussion? 

Why, for a gathering of organizations that strive to be process-centric, not outcome-centric as a means of assuring quality work, was "Stop, pick a more appropriate process" not an option? 

-- Tom (OP)

Permalink

The whole thing just makes ISO look bad.  That any one company can so obviously skew the process to push such a big standard through a fast track process is bad enough.  That ISO would even let such a large standard with so many comments try to be resolved by a single week BRM is quite laughable.

Laughable but not funny.  It is simply sad that so much time and effort and public money is being wasted in dealing with this abuse of of the organisation.  One that is there presumably to serve the needs of ALL interested parties, not just a single company.

Will it even really matter in the end?  If the company pushes this through the system they may get a few more years out of their monopoly, but an ISO mark doesn't make a rubbish standard a golden one.  And neither is it any guarantee of success in the marketplace.

  Michael

Permalink
I have written a number of comments at Groklaw, using the username of "E-man". I have no particular connection to ISO or Groklaw or any of the parties involved with the OOXML specification. Perhaps you have seen my comments at Groklaw, though.

As you are no doubt aware, Pamela Jones wrote an article largely based on your blog entry. I was very disappointed with your blog entry as it was originally written and wrote a number of comments at Groklaw that  reflect this. Since you are interested in receiving comments here, I will try to write directly to you this time. (Actually, I also commented anonymously about a math error that was quickly fixed.)

It should have been obvious that problems caused by what you wrote here were likely to be propagated far beyond your blog. Many people at Groklaw seem to have been confused about what happened at the BRM, and I lay at least some of the blame for this at your doorstep. At the same time you deserve credit for what you have helped them understand.

I note that you have made some changes to the article. Thank you for doing that. I think it is much improved. Unfortunately, I doubt corrections will as noticed or propagate as well as the original text. One relatively minor thing that might help would be to call attention to the fact that you changed the title. That change is not mentioned under the March 02 updates (I think that's when you made the change), yet it is certainly as significant, in my opinion, as the changes you listed.

Actually, I'm still not satisfied that the title is accurate; how about "Most OOXML Dispositions Not Even Discussed at BRM" if you don't like Dr. Brown's ideas?  While I am at it, I don't think the word "and" should be in the second sentence where you describe the the March 02 updates - that's just a matter of grammar, but it did confuse me for a while.

The remaining complaints that I still have about your blog entry, but which you have not yet addressed, involves your tabulation of "votes" and the text near your table. What you were doing, if my understanding is at all accurate, is using default values for the ballot as proxies for ~900 votes that were officially cast. (The tilde indicates the value is approximate.) That the default value was simply intended as a means to cast a large number of votes easily and that you were using the same default value as an estimate or approximation of how ~900 votes were officially cast is hardly clear from what you wrote. Certainly many of the comments at Groklaw indicated that the authors missed those points. I think you should have made much more clear that  the table was not a table of actual votes.

It appears that many people didn't understand your explanation of the default value for the votes. Compounding that, when you begin using the default value as a proxy or an estimate, you didn't really announce that fact, and as a result people became very confused and misunderstood what was going on, I think.

The labels you used in your table indicate that you tabulated results for actual votes, which is not the case. One column you have labled "All Votes". Actually, that is "All NBs" or "All Default Values". Likewise, you have a row labeled "Total votes cast". Actually, the total votes officially cast would be ~900 times those figures, given that there were ~900 dispositions on the ballot, so again, the label should be "Total NBs voting" or something like that.

You also describe the table's contents as: "The final tally (as recorded by participants, and subject to final confirmation)", as if it were something to be officially tallied as part of the BRM process. Can you blame people for interpreting it that way if you don't don't make it obvious that they shouldn't? Unless I'm very wrong about something, it was you who decided to draw conclusions based on the default values. No doubt others trying to get an early understanding of the voting did the same, but I very much doubt ISO has approved doing that.

I am also concerned, for a somewhat different reason, with this statement about the ballots: "On Friday, the ballots came back.  Some contained votes on a small number of dispositions and some adopted the default option for all of the listed dispositions." How accurately (and completely) does that describe how all NBs voted? It is important to know that in order to understand the usefulness of the default value as a proxy for the ~900 votes. It is particularly significant in this situation because the number of cases where "approve" was the default value was so close to the number of cases were "disapprove" was the default.

I realize that you talked to some delegates who were actually at the BRM. The question is: Were they telling you what all NBs had done, or were they telling you what almost all NBs had done?  I find it hard to believe that there weren't at least a few cases where the NBs arrived at the BRM already knowing how they were wanted to vote on any dispositions that weren't changed at the BRM, who liked (or disliked) a significant number of dispositions, but  took a different position on a significant number of other ones, and who then applied those votes when it was time to cast the ~900 votes.

My expectations seem to be confirmed by the data that someone has anonymously posted in your comments. That data supposedly indicates how the NBs voted on all votes, as I understand it. If  that information is accurate, even if it includes votes other than the ~900 that you are talking about, your statement certainly does not describe how Brazil and Switzerland cast their votes and it is questionable with respect to a couple of other NBs, as well. So, despite what you may have been told, I don't think that your statement describes the way all NBs voted.

I know it is difficult to describe how the vote was taken and how you are using the default values and how people should interpret the results. I've tried doing that myself, and am not terribly impressed with what I've managed to come up with, either. (There are things that I wrote in Groklaw's comments that I certainly would change if I were permitted to edit them.) Also, I know there is only so much you can do with respect to how people will interpret and use what you write, but in this case I think much of the misinterpretation was predictable. Frankly, some of it wasn't misinterpretation at all - your title did say (originally) that OOXML was rejected at the BRM. Naturally, my comments at Groklaw reflect you blog as it was written at the time.

I appreciate the new material you have added and the changes you have already made to what was originally written, I still think you can make it better, though. Please consider what I have said here.

Regards,
E-man

Wow ! I just read your comment and wonder whether you understand the conditions under which it was written.

Andy was trying to get as much valid information as possible as soon as possible because experience has apparently taught him that whenever Microsoft's interests are involved, the MS astro-turf machine will pre-prepare and pre-approve blog entries waiting only for some preliminary numbers from those MS employees that actually attended as delegates.

Witness the accuracy of his predictions by the way that Jason Matusow was able to blog misleadingly about the results of the BRM as early as Friday evening  (Friday Afternoon Pacific Time).  The only way he could get any numbers or results at all to use in presenting any semblance of authenticity (and that appears to be all he's got) is to have been fed BRM results either during the BRM or immediately after the BRM from those MS employees that attended.

I give Andy unrestrained kudos for his ability to put the original blog post together in a VERY short time to refute the FUD that Jason Matusow and other MS mouthpieces  had already published and in so doing he gave  those that choose to fight for sanity in standards some amount of truth about the real picture that is only now emerging from the BRMto use in countering that FUD.  You only have to look at how this blog has changed due to updated information that is intentionally sought out and actively elicited.  While I recognize (and agree) that your comments are valid in hindsight, even you did not have all the details correct on Friday when you posted your original information on Groklaw.  Even now, I have several questions outstanding such as "Jason M. says the  large number of dispositions passed with a clear majority.  I question that because I think most sane people would take the 'majority to mean "of those in the room that vote".  By that definition, the 'clear majority' abstained.  However Alex Brown chooses to read this otherwise - to include allowing O-Members to vote which may be a clear violation of JTC rules.  Of course if 'majority' is redefined from "of those casting yes/no votes only" to "of those in the room", the outcome of the BRM is MUCH MUCH different and then jives MUCH more closely with Andy's original blog post.

Unfortunately, while you take potshots at Andy's accuracy, Jason Matusow, Brian Jones (a US delegate to the BRM) and others are producing as much marketing malarky as they can before the SC34 results are officially released.  I for one am not aware of any other place on the web where accurate info about the BRM is being collected other than here.  It's even become clear that Alex Brown has done what he can (intentionally or otherwise) to assist Microsoft in maintaining their document lock-in by lending his credibility to BRM results being interpreted and and published as fact on the internet, in mainstream press, and in the minds of politicians and enterprise management in the way that MS wants them published without waiting for any official results to be tallied, for final drafts to be written, or for the expected appeals to be filed/resolved.  Meanwhile, those that care about the truth are hand-tied by lack of information other than that found here and what information continues to come in from other delegates.

All that having been said, I note (on my own nickel) that Andy has a typo in the article where he refers to BRM as BRT.  (text search would be the easy way to find it Andy).

I'd also like to thank Andy and Rob Weir for their efforts to get out first word to try to counter the MS BS machine that's already proclaiming complete victory.  I truly hope that BRM members file appeals against the votes that gave no time for review, that allowed O-Members in clear violation of the written JTC rules, and that defined an 'abstention' as a 'non-vote' and that apparently ignored a large number of abstentions in order to claim that most (all?) edits not discussed passed and were approved.  One thing I still don't understand though, is why those countries that voted 'yes with comments' were allowed to participate and raise issues that no doubt took up valuable time at the BRM when they'd already voted yes.  This should have disqualified the US, Poland, and I'm sure a few others that voted blanket yes and would have seriously changed the outcome of the BRM votes as well as the resulting DIS edits.

Perhaps Alex might address this question the next time he stops by ?

Thanks again Andy !  This is invaluable information - made even more valuable by being early and by increasing its accuracy !  ...and thanks ESPECIALLY for keeping it real and not succumbing to the misinformation and marketing spin that MS & their agents continue to try to insert into your blog.

-- Ed

Ed,

Thanks very much for the support.  Yes, it was a very hectic week.  I got no sleep on the way to Geneva (can't sleep on planes) and was working (keeping up with client work), presenting or in meetings every day from 7:00 AM to 11:00 (Thursday I took the train to a client meeting in France).  Friday night I was catching up and communicating with folks till 3:00 AM.

Necessarily on Friday I was dealing with information as it came in and trying to figure out what to make of it.  The only real mistake I made of significance - and one I regret - was interpreting the "abstain" votes as rejecting the process.  In fact, as the piece now reads, it was an effort by the delegates that chose this option to vote only on the dispositions they had a position on and say nothing about the rest.  That's not support for those dispositions, but it's not appropriate to characterize that as "rejection" either.  Only those that chose "disapprove" as a default option should properly be put in that category.

I was traveling home all day on Saturday and out of email contact for most of that time, so I was not able to make this change until I arrived home.  Since then, I've been in touch with members of many delegations, and tried to update the blog as much as necessary, while not agreeing that the vote taken can fairly be taken as a majority approval, from a legitimacy as compared  tio a procedural point of view, of the 900 dispositions that were put to a single ballot.  That is simply more than I think can validly be said, given accounts (such as that from Greece) that some of those involved had never had time to read them all, and few if any (especially among the smaller delegations) had much time to consider and vote  on them over night.

Turning to your question about attendance at the BRM:  my recollection of the rules is that those who had voted against OOXML over the summer had a duty to attend, while those that voted otherwise had the right to attend.  That's not as illogical as it might sound, since changes (as Frank Farance has noted) can make things worse as well as better.  So it seems fair to me that even though someone had voted to approve, under normal situations would have a right and a purpose to attending.  Under the current circumstances, of course, the motivations for those that voted "yes" last summer and "yes" again at the BRM would be of interest.

I think that it's particularly telling that the US delegation voted "no" as a default on the 900 dispositions at the BRM, notwithstanding the fact that the INCITS Executive Board members voted for the US last summer  to adopt OOXML without any improvement at all. 

  -  Andy

For the BRM to be successful at least 2/3 of the votes should be YES and no more than 1/4 of them NO.  I'm certain that it failed be the first part doesn't appear to be met (only 60% YES when only counting those who voted YES or NO, much less when counting abstains) and possibly the second part wasn't met either (40% NO when only counting those who voted YES or NO).

The BRM chair should recommend referring the issue to OASIS TC for ODF.  Further, he also needs to recommend that ISO look at stripping ECMA of its fast track ability.  Only NBs should be allowed fast track and no one else.


I understand your passion, but I find nothing wrong in the points raised by E-Man or in his raising of those points, even as a person who has developed a strong bias against OOXML.

You are correct that Andy did as best he could to get out early and fast. But, it should be early, fast and accurate.

You point out more than once that MS will have marketing out in full force; that's not surprising to anyone. But, "two wrongs don't make a right". You don't fight Microsoft falsehoods with opposing falsehoods. That's a slippery slope that we currently see ISO using...

You only get one out of two:

   1. Early = Fast.
   2. Accurate on day one.

Andy's blog appears to have become the center of this BRM hubub on the web.  Probably because of his sincere search for truth (which is not always readily forthcoming when commercial interests trump common sense) or perhaps because he actively searches out multiple sources of information (by inviting others to post their experiences here).

I was reacting to E-Man's apparent attack on Andy' intentions and results which he was vigorously pursuing while relying on the benefits of 20/20 hindsight.  Nothing more.  I don't disagree with his points, just his manner of delivery.  In a perfect world, his arguments might even be fully valid.  Unfortunately, the world is not perfect and we do the best we can.

There are things that are said in any early search for truth that are found to be inaccurate later.  But credit should be given when those inaccuracies are corrected as soon as they are found.  Andy gets credit.  Microsoft does not (since they don't accept corrections).

The BRM process just completed does not get credit.  The individual delegates do (except for those that intentionally gamed the system such as the MS reps that also functioned as HOD in order to blanket-vote yes to all dispositions).

I'm still waiting for the formal tally to be released to see what is actually formally accepted and what is left standing after any appeals get made.  I'm sure the NBs are also hesitant to commit to re-evaluating the spec since none but those in attendance even know what changes were vs were not approved and it sounds like even those in attendance may not have good information.

Question - how long after the BRM do the results get published so the NBs know what it is they are actually supposed to vote on between now and the end of March ?

--Ed

Ed,

You asked:
Question - how long after the BRM do the results get published so the NBs know what it is they are actually supposed to vote on between now and the end of March ?
I _believe_ but can't say for sure, given that this continues to be an unusual  situation, is that Alex Brown is to do a Report of the Meeting (that may not be the exact title) that would go just to the NBs - I assume all 87 that will be entitled to vote.  I don't know whether there is a set format, or whether that's up to the Convenor.  I am told that this report can, but is not required, to be made public, but that's from a committee member, not from ISO/IEC JTC1 itself.

I have no idea how long it will take for the dispositions to be turned into specification changes; perhaps Microsoft and Ecma might have already started on that before the BRM(?)  If not, it's hard for me to imagine that a full new draft could be produced much before the end of the 30 day voting period, much less reviewed.

So this is just supposition, but that would mean that the vote would be based primarily on (1) the proposed dispositions as originally distributed, (2) the small number of substantive dispositions that were revised during the meeting, (3) the 200 easy text revisions, and (4) the Report of the Meeting.

One intriguing question to ask is this:  what about all of the dispositions for which small ad hoc working groups were appointed that never had time to report back their recommendations for discussion?  This indicates that there was an issue with the disposition, and that a better fix might have been available, but no way to take advantage of that better alternative.

  -  Andy

Thank you for that information, Andy.

Do you have any feel for the impact of the various press and 'internet buzz' one way or another ?

My concerns are as follows:

   1. The MS marketing engine prematurely declares victory and does so loudly enough and often enough that (repeating a lie often enough makes it true) their version becomes 'conventional wisdom'.
   2. Meanwhile, official information is lacking so there is no way to counter this media blitz with true facts (because there aren't any).
   3. The changes & draft are *SO* large that whatever report with edits is inevitably delayed just due to the pure size of the DIS+changes and a need to verify the accuracy of the post-BRM draft.
   4. Every day the official report is postponed is one less day for NBs to review the revision.  Both the generation of the revision (or even a report of the revisions) and the review and assimilation of the revised draft need as much time as possible (and the clock is ticking).
   5. Should any NB appeals be filed due to the discussions either here or on Groklaw based on the JTC 1 not following its own rules, there could be a *huge* impact on the revised DIS (Take for example that the last vote gets invalidated and those edits are all ruled as 'fail' due to lack of consensus - the revisions would have to be backed out.  Should the O-Members get ruled as 'non-votes' the big-ballot vote results may change drastically with corresponding impacts on the revised draft that may have already been sent out to the NBs).  You see what I mean...
   6. After all the internet buzz and after MS has already announced its 'win' to their business partners, I suspect that economic pressures from those same business partners will not look kindly upon the JTC 1 results *finally* being published and indicating that due to irregularities, the BRM is a failure and the OOXML DIS has failed (for example).  Consider the case in Denmark that is already declaring OOXML an 'open approved format' for government documents.
   7. Will the NBs that did not attend the BRM have any time at all to review this monster spec before their vote is due ?  Will they know what to review ?  (Remember the mis-information spread to NBs that they *only* needed to review their own comments as preparation, then got a last-minute requirement to vote on everything...)
   8. On a different vein, is the presence of MS employees on so many delegations of concern to anyone but myself.  Do these people not have the decency or integrity to recuse themselves due to conflict of interest ?
   9. The more I think about it, the more I'm bothered by a statement made in either the OpenMalaysia or the Greek blog that Brian Jones (from Microsoft) made a last-minute (Friday afternoon?) proposal that the BRM be 'interoperable' with the ECMA recommendation.  This is the type of conflict of interest I meant in my previous item.  Such 'interoperability' would basically require that no change can be allowed that prevents Microsoft's ability to continue utilizing it's proprietary protocols & undocumented features.  This in turn prevents any hope of interoperability because it allows Microsoft to utilize it's Office 2007 market share to either demand royalties of any and all interoperable products for the use of their precious IP that is *not* covered under the OSP (referenced in the spec or reverse-engineered if macros, etc).  It would also allow MS to not bother to 'fix' MSO2k7 to make it open or interoperable with standards, but to take the 'deprecated' approach which is really not interoperable with any other implementation.
   10. I get the distinct impression that MS already has a multi-million-dollar advertising blitz primed and ready to go, just waiting for the final approval of their shiny, brand-spanking-new ISO stamp-of-approval.  I hope I'm wrong, but I fear I'm not.

--Ed

I did not mean to criticize his intentions. He had posted several comments were he had indicated he was interested in what flaws people were seeing in what he wrote. Since I was warning people elsewhere that his article was confusing and misleading, I thought perhaps the right thing to do was to comment here. Since he hasn't responded directly, I'm not sure if he appreciated the information or not, but it seems doubtful. Oh, well.

I suppose I was a little ticked off and embarrassed that I took his original title at face value. I went and posted a comment with the title "MS lost! OOXML rejected" at Groklaw. Once I understood the truth a little better (and remembered what the BRM was about), I posted a correction, but don't know who I mislead in the meantime. (It's still there if you care to look for it; I can't delete or edit my own comments.) That's very embarrassing to me. Maybe I should have known better than to trust Mr. Updegrove so much under the circumstances, but I did feel mislead, which is magnified by the embarrassment and concern that I mislead others. I, however, did the best I could to repair the damage that I had a part in causing. I thought Mr. Updegrove might want to do the same, if he was made aware of it.

What Mr. Updegrove is doing is a tricky thing. I'm no expert, but I know any flaws in what he writes have a cost. They will demoralize and confuse people on one side and the other side will take advantage of them for their own spin. OTOH, there's only 24 hours in a day, and only one Andy Updegrove. I was pointing out that people were getting confused (maybe he wasn't aware of that or why), but only he knows if he has the time or interest to fix things, especially since the new material is making the old less important. I'm fine with that.

Perhaps my tone came across a bit harsher than I intended. I certainly don't have any great communication skills, I know.

As if to prove I have no great communication skills, I just wrote the past tense (or past participle?) of mislead as "mislead". It should have been "misled." Sorry.

Sounds to me like you're suffering from some of the same shortcomings Andy suffers from:
   - Early confusion about what happened
   - Lack of full details on the day the BRM ended
   - Lack of complete understanding of the whole BRM/JTC1 process/environment as of the day the BRM ended.
   - Lack of first-hand details behind the big-ballot vote.  Lack of background and context for the big-ballot vote.
   - Lack of having been there
   - Desire to 'get it right the first time'
   - Lack of perfection
   - Personal Integrity
   - Sense of honor
   - Desire to set/keep the record straight both for posterity and as  an accurate reference for others
   - Desire to find out what REALLY happened - as opposed to reading spin after the fact.
   - Embarrassment when you inadvertently mislead people

I don't see where these shortcomings have impacted the Microsoft bloggers.  ...Must be some of the mental handicaps that we open-source supporters work under....

Like me, I suspect you and Andy both would have trouble working for Microsoft.  ;-)

-- Ed

@E-Man,

I apologize for not acknowledging your comment sooner. The main reasons are that I have been overwhelmed with other work (catching up after returning from Geneva) and because I think that I've already addressed all of your concerns in comments that I had already posted, or was in the process of posting, in response to Alex Brown, or in updates to the blog entry itself.  Also, I thought that Ed Winter had kindly stepped in to say most of what needed to be said.  That said, I'm sorry for the delay, as I do try to respond to all  comments that obviously are looking for a response.

More specifically, here are answers to your comments:

I do regret that there were some inaccurracies in the entry, and have said that in another response.  I've also explained in great detail why I'm sticking with the title in my dialogue with Alex.  I've also been quoted (at last count) in scores or articles in virtually all the technology press, and others such as the Washington Post, and in all of them the facts quoted are accurate and the message clear, and based on a press alert I sent out, rather than the blog entry, so I think that the accurate story will propagate. 

I didn't mention the change in title in the update summary, because it is in the comments to Alex, and also the substance of the change is included in another updated change in the text itself.

As to this:
The remaining complaints that I still have about your blog entry, but which you have not yet addressed, involves your tabulation of "votes" and the text near your table. What you were doing, if my understanding is at all accurate, is using default values for the ballot as proxies for ~900 votes that were officially cast. (The tilde indicates the value is approximate.) That the default value was simply intended as a means to cast a large number of votes easily and that you were using the same default value as an estimate or approximation of how ~900 votes were officially cast is hardly clear from what you wrote. Certainly many of the comments at Groklaw indicated that the authors missed those points. I think you should have made much more clear that  the table was not a table of actual votes.
I think that the table and text are clear.  I'm not a perfect writer,  I'm sure, but I'm not aware that other people had a problem with this.  I do the best I can, and people seem to find value in it overall.  I don't want to get into a discussion as to whether the table could have been clearer.  I have no editor and need to do the best I can.  I'm not aware that many people had a problem with it.
 Unless I'm very wrong about something, it was you who decided to draw conclusions based on the default values. No doubt others trying to get an early understanding of the voting did the same, but I very much doubt ISO has approved doing that.
Of course - and that is the entire point of the blog entry!  ISO/IEC closed the BRM crowing about how it had been such a big success, and how 98.4% of the dispositions had been approved.  It was crucial to give the facts on that so that everyone would not be misled, and so that Jason Matusow's account would not be the only one that journalists saw when they wrote the first articles.  If you do a Google search, you will see that most of the articles that have been written were written within 48 hours of the close of the BRM.  Absent my press alert, the story would already have been written that the BRM was a great success, and outside the blogosphere - such as in government circles - it would have been too late to correct the damage.  Sometimes you need to go to press with the information you have, before the worse damage is done.
I am also concerned, for a somewhat different reason, with this statement about the ballots: "On Friday, the ballots came back.  Some contained votes on a small number of dispositions and some adopted the default option for all of the listed dispositions." How accurately (and completely) does that describe how all NBs voted? It is important to know that in order to understand the usefulness of the default value as a proxy for the ~900 votes. It is particularly significant in this situation because the number of cases where "approve" was the default value was so close to the number of cases were "disapprove" was the default.
This is what I was told by those that saw the votes.  I can't do more than report what I've heard from sources that I believe to be accurate.

Finally, I think that you need to appreciate that I took a week out of my life and my job to go to Geneva (I worked non-stop without sleep on the flights each way), put in 18 hour days while I was there, talked to as many people as I could (which is a lot), and that later accounts have confirmed the great majority of what I wrote.  If you'd like to encourage people to continue to do things like that, you need to accept that sometimes non-professional reporters with no supporting resources are not going to be able to nail every fact - or, if they do, that it will be too late to matter.

  -  Andy

Don't worry about not nailing every facts rights. I once had an experienced journalist, now the director of a major newsroom, confess to me he couldn't read a single news article in the media on any topic he knew about without finding at least two errors of fact. By that measure your work is a huge scoop of outstanding quality.

Permalink
Hard to know just where to start.

Let me just say that I've had the pleasure/hell of chairing many sessions in my adult life on issues contentious and otherwise.

One thing, though, was blindingly clear before the BRM was held.  There as no way that a mere 5 days would even get a good start on doing what ISO/IEC say they wanted to do given the number of  comments in the September 2007 balloting.  So I'm not at all surprised that that is exactly what happened.

It isn't as if this wasn't known prior to the meeting and a longer session was rejected by ISO/IEC and the convenor Alex Brown.  All within the rules, I note.

Now we all move to the next step on this proposed "standard" with more bad feelings and recrimination that was present before it started.

It strikes me that the fast track process was put in place for quick adoption of proposed standards on which there was a considerable level of agreement so that compressed time lines for sessions like the BRM stood a chance of success.  That never existed here nor is it likely to given the proposal's present form.

ECMA may not be guilty of out and out abuse of the fast track process given the above though the effect is the same.  Perhaps worse as ISO/IEC has let the abuse go this far.  Probably also within the rules which, by now, are known to be utterly incapable of handling a proposal like this.

Nothing, I'm sad to say, will change the black eye ISO has inflicted on itself with this opaque process, lacking any and all transparency which, at the end of the day, has left more confusion than when it started.  The participants, NBs, Alex Brown and others present did their best with an impossible situation that would never have been allowed to happen by a competent standards organization.  That's how ISO looks now.

Incompetent.

ttfn

John

Permalink

Does the OOXML spcecification in its current state guarantee long-term access to data without legal or technical barriers?

Regards,

Permalink
Hi Andy (and Alex, I suppose)

I've read a lot of the accounts and your blog, but what I think fails to appear clearly is the basis of the entire mess of this process.

I think the clearest picture is achieved by combining http://www.iso.org/iso/pressrelease.htm?refid=Ref1114 and http://elot.ece.ntua.gr/te48/ooxml/brm-clarifications with your accounts of the procedure here (over the last year).

The cause of the fast-track is the relationship of ISO with ECMA, which I gather has been under strain longer than just this particular "standard". The failing of the first step of the fast track causes an abomination of a standards process, because it's trying to save spilt milk. The resulting BRM can only work with a very limited set of issues, which shouldn't cause a failure of the first fast-track vote in the first place.

As I understand it, the technical nature of the BRM will only try to resolve the issues itself, not resolve whether the proposed standard is fit for its job or the title of ISO standard. In fact, there is no vote on that, since ECMA is allowed to enter their "standards" in the fast-lane, but I digress. A tally of the "approve/disapprove" votes will determine whether the fast-track procedure has succeeded and the proposal is made a full ISO standard (with issues to be resolved, I presume). This tally is the apples and oranges bit, because the type of the majority of issues resolved (typos, grammar, etc.) are not indicative of the quality of the standard at all!

OOXML is a cuckoo's egg in ISO's nest!

I have no doubt the final outcome will be approval, because alternatively it will require an all-out subversion of the process to block it. Instead of taking the process serious, NB's should change their votes to an all disapprove one, to indicate their lack of acceptance of the standard, instead of the specific approval/disapproval of the technical issues.

Cheers

Simon Oosthoek (interested spectator)