The Standards Blog

If this were any other JTC1 Proposal, the OOXML Vote Would be Over Now

Monday, September 3rd, 2007 @ 11:08 AM EDT
Contributed by: Andy Updegrove
Views: 18,448

OpenDocument and OOXML

Updated:  I am now predicting that the OOXML vote has failed to pass.
Public announcements of how P members of ISO have voted on OOXML are now rolling in one at a time, and the trend thus far is meaningfully weighted towards "No with comments."    By my count, there are now four announced Yes votes, with comments (Germany, Poland, Switzerland and the United States), two abstentions (Australia and Sweden, the former due to a failure to achieve consensus, and the latter due to voting irregularities), and seven eight public No with comments votes: Brazil, China, Denmark, France, India, Ireland, and New Zealand. Updated: and Norway; tracking changes made below.   There is also a blog posting of a No with comments vote by Korea here. In addition, I expect at least Canada, Japan and the United Kingdom to announce "No with comments" today or tomorrow (that would take us up to 12), and that a number of additional countries will be revealed to have voted in a similar fashion when the official vote tally is announced by ISO in the next day or so. 

The reason I say that in any other case the vote would be over now is because of the 11 countries that upgraded their status from Observer to Participating member status in the last few weeks. Without those extra 11 P countries, it would only require 10 votes to make an overall vote to approve impossible under the ISO rules (i.e., one more than 1/3 of the former 30 P members, minus the two that have abstained).

But with those eleven new P countries, and assuming no further abstentions, it could now conceivably take a total of 14 no votes, calculated as follows: there are 41 total P members. If all vote, as they are supposed to do, then you would subtract the two abstentions, yielding 39, and two-thirds of 39 is 26. Hence, more than 13 no votes would be needed to block a yes vote. And only by blocking a yes vote can there be assured to be a meaningful effort to address the many comments that appear to be accompanying the votes – including the yes votes. (You can read more on how the results will be calculated here).
 

Interestingly, while 13 countries have publicly announced their votes, not one of the new 11 P members has thus far revealed how it has cast its vote. As I have pointed out in several recent blog entries, (the latest is here), it will be very interesting to see how these last-minute additions to the P membership cast their votes. Those new P members, once again, are Cote d'Ivoire, Cyprus, Ecuador, Jamaica, Lebanon, Malta, Pakistan, Trinidad and Tobago, Turkey, Uruguay and Venezuela.

Updated:  You can follow a visual estimate by the folks at <no>ooxml.org site here that I assume they are updating in real time.  As I write this, they're guessing 18 no votes, which would be sufficient to block an outright approval.  A spreadsheet I received over the weekend from another group that has been following things closely was forecasting 16 No votes (on receipt also sufficient), and since then, one vote they expected to be a Yes turned to a No, and another to an Abstain.

For further blog entries on ODF and OOXML, click here

subscribe to the free Consortium Standards Bulletin

Comments

Submitted by Anonymous on Tue, 12/13/2016 - 00:26

Permalink
It is still possible that some of those new P countries will vote no. We had been operating under the assumption that MS bought the new votes. We might be wrong. May be IBM bought a few as well, who knows ;-)

However, Brian Jones seems to be confidence to get 60%+. He must know something I don't, coz the incumbant vote is a rather strong "No".

If all the new countries voted "Yes", then ISO will have to seriously consider whether someone really stacked the voting process, as this is an unlikely result for 10+ new countries to come in with the same vote, particularly if incumbants voted "No" rather strongly.

"It is still possible that some of those new P countries will vote no. We had been operating under the assumption that MS bought the new votes. We might be wrong. May be IBM bought a few as well, who knows ;-)"

or may MS attempted to stack some of this countries but the people concerned about true open standards neutralized the unethical move  ( and had avoided another Portugal fiasco )

( this had happened in many NB: MS attempt of stacking -> community and concerned citizens response: i.e.: Australia, Sweden, etc. )

Your table looks consistent with what I've been able to glean, but I have to caution that what you're showing in some cases are pretty old links rather than final votes.  As the US vote and a number of others show, things can (and have) changed right up to the last minute.  At the moment, I'm counting 11 No's that are based either on public announcements or ones that I'm comfortable with my sources and the situation on the ground (12, if the story in my News Picks column is accurate).  Given that only 14 negative votes are needed, and likely fewer, since any additional abstentions will be deducted from the total by which the 2/3's will be calculated, things are looking pretty good.

On the question of P countries, I think that you can use this as the definitive link.  It will be very interesting to see whether any new countries at this page on Tuesday (the most recent addition was made last Thursday) .

  -  Andy

 

Submitted by Anonymous on Tue, 12/13/2016 - 00:26

Permalink

I would suspect that there are far more irregularities than shown on the map at NOOOXML.org. Finland, for example gave MS an abstention, like Sweden later did, despite having a government mandate for use of ISO/IEC 26300 and several ministries already putting the format into production.

When this is over and done with, the relationship between Ecma and other standards bodies needs to be re-examined. This mess came from their collective desk.

Furthermore, fast-tracking may have outlived its usefulness and might be reconsidered.

Lastly, ISO should take a page from IETF's book and require that any specification have at least two working, independent implementations (e.g. OOo and Koffice) before even being considered.

Submitted by Anonymous on Tue, 12/13/2016 - 00:26

Permalink
OOXML might qualify as having two independent implementations. Apple's new iWork '08 suite includes support for OOXML (and not ODF, for some strange reason). Granted, that support is probably constrained by the combined complexity and lack of necessary detail in the OOXML draft.  It's quite possible, even probable, the implementations do not conform with one another.  That's a testable, technical question.