Genii Weblog

A sad after effect of the aggressive Open XML push

Tue 25 Sep 2007, 12:17 PM



by Ben Langhinrichs
In the recent aggressive push by Microsoft to get Open XML certified as an ISO standard, many countries were convinced to switch to P-countries in terms of voting.  The inevitable, but sad, after effect of this effort is shown with the failure of JTC 1/SC 34 N 870.  Did this standard fail due to negative votes or numerous comments?  No, of the eight countries voting, seven voted for the ballot, and even The Netherlands, which voted against it, had only two comments.  The reason why the ballot failed was explicitly stated:
Based on an insufficient response from the SC 34 participating membership, this ballot has automatically failed according to Directives 9.1.10 that states "if more than 50% of the P-members have not voted, the vote will have failed." Late votes are not counted and there are no extensions to the vote.
So, what would the results have been before this run up of P-countries?  As best as I can figure,  with 32 P-members now and 16 P-members joining fairly quickly to vote on Open XML, the vote would have been passed without those new members, since 50% of the P-members would have voted, and the vast majority voted for Approval.

Is it bad to have more voters?  Intuition tells us that it is not, but intuition is not always correct.  When there are more voters, but they don't vote, and when a quorum is required, those new voters can kill off a perfectly acceptable standard.  I know little to nothing about JTC 1/SC 34 N 870, but I do know that many other qualified and reasonable ballots may fail because of the surfeit of under participating P-members, and that is a real shame.

Update: I saw an argument on another blog that
that particular ballot failed because only 8 member voted which would never have sufficed in any scenario.
In case you say, "My, that sounds reasonable!", check out this ballot from one year earlier, JTC 1/SC 34 N 745.  It is interesting comparing these two.  Both had seven P-members voting for approval and one voting against, but the earlier ballot passed and this failed.  In case you need graphic evidence of what has happened in the meantime, look at the list of P-members from one year ago (country that did not vote in italics):
Canada
China
Italy
Japan
Korea, Republic of
Norway
Thailand
United Kingdom
USA

and then the list of P-members voting for this past ballot (countries that did not vote in italics):
Bulgaria
Brazil
Canada
Switzerland
Côte-d'Ivoire
China
Colombia
Cyprus
Czech Republic
Germany
Denmark
Finland
France
India
Italy
Japan
Kenya
Korea, Republic of
Kazakhstan
Lebanon
Malta
Netherlands
Norway
Pakistan
Poland
Romania
Sweden
Thailand
Trinidad and Tobago
United Kingdom
USA
Venezuela

Does this make it clear how much harder it is going to be to pass ballots now?

Copyright © 2007 Genii Software Ltd.

What has been said:


628.1. Richard Schwartz
(09/25/2007 08:14 PM)

It might be clear -- but I think you've italicized the entire comment, but the comment uses italics to separate the countries that did and did not vote! And I think the net result also includes an unclosed italic tag.


628.2. Ben Langhinrichs
(09/25/2007 08:19 PM)

Thanks, Rich, it is fixed now. I spelled "blockquote" as "blockqoute", and unsurprisingly, it didn't work.


628.3. hAl
(27-09-2007 03:11)

I would call this article serieus numbers abuse.

You as an interested party should wel be aware of the fact that the Opendocument ballot in may 2006 had 27 positive votes on it from P-members.

These number have been repeated quite a number of times in the last few weeks becuase of the ballot result on OOXML.

Now you are suggesting that the membership declined to 9 P-members only 4 month later in september 2006.

That is just rubbish and you know it.

If that were actually true than the opendocument vote would have been a realy serieus ballot stuffing example with members ship dropping to 1/3 just after that vote.

You are misrepresenting the secretariat document ballot results as they are not offical lists of P-membership. Please be a bit more accurate with writing such articles.


628.4. hAl
(27-09-2007 07:04)

I'm sorry. seeing the ZIP file I still conclude 27 P members however only 23 P-members voted for approval, one P-member abstained, two P-members did not respond, and the US was acting secretariat but (was and still is a P-members).

I was fooled by the 100% approval mentioned which does not include non responders or abstain votes and I did not know that the secretariat vote was not consider a P-member for voting purposes ?

And I still think you are misrepresenting the case as you use irrelevant voting results as a measurement for P-membership of JCT1/SC34.

You do not present actual member lists but conclude stuf on based on ballots without knowing the exact procedures for those particular ballots.

However you were aware of the ODF vote which at least has the closest resemblance to the OOXML vote and you did not show those figures but used very different figures of other votes. So you havew been looking for those completly non relavant non-related votes in a ton of auto generated secretariat documents whilst being aware that the most similar vote would be in a ballot vote on ODF which you were already aware of and which is much more relevant to the discussions around OOXML and more published.

An easy explanation for instance that in the secretariat document a ballot result is only generated for countries that have been entered in the voting result.


628.5. Ben Langhinrichs
(09/27/2007 07:25 AM)

hAl -

I just looked back a year. I didn't notice the ODF numbers from earlier than that. And in any case, you have no more evidence that these numbers I give are irrelevant than I do that they are relevant. Neither os us knows why the numbers vary.

And as for your "easy explanation", it doesn't make sense, since the countries that didn't vote are explicitly mentioned in the posts above. Somebody is going to have to answer the questions of "Why?", but you still don't have any evidence I am misrepresenting anything, although I am certain;y more wary about my data now.

- Ben


628.6. hAl
(27-09-2007 08:37)

That explanation stil makes sense because if "no response" has to be added as a result to show then it might be dependant on who actually add the results to the tooling. The secretariat page is clearly a generated page so there is a lot of thing that depend on tooling support. It is however not a P-member list and should not have been presented as such unless it states "Member List".

I can only find similar numbers info, concerning ballot votes that are failing relating to the number P-members, on anti ooxml sites. With such an article you place yourself in a similar category of information supply.


628.7. Ben Langhinrichs
(09/27/2007 08:54 AM)

hAl -

LOL. I am taking information directly from the JTC secretariat's website, and because you don't agree with it, I am anti-OOXML and my information is suspect? You have no solid basis for explaining away the specific links I made, but you think I am biased because I don't guess that the numbers they give are wrong? And you think I am biased? LOL.

- Ben


628.8. Ben Langhinrichs
(09/27/2007 09:01 AM)

hAl -

In the interest of fairness to your point of view, how do you explain the discrepancy between ballot JTC 1/SC 34 N 870 (to which I link above) which failed with a vote of 7 P-members voting to approve and one voting not to approve due to a lack of P-members voting, and the ballot JTC 1/SC 34 N 745 (to which I also link above) which was approved with a vote of 7 P-members voting to approve and one voting not to approve. Two separate ballots, same exact votes, yet the latter failed to P-member percentage and the former did not. How is that possible given your "explanations"? Until you can explain that away, my fundamental poitn remains. A ballot was approved with a specific vote a year ago. A ballot failed with the same exact specific vote this month, due to a failure to attain 50% P-member voting. Do you believe that somebody who is ant-OOXML fudged the numbers and messed with the official website?

- Ben


628.9. hAl
(28-09-2007 00:28)

Mayby it weas just a workgroup related vote ?

Not all ballot votes might require the need for all of JCT1/SC34 to vote.


628.10. Nathan T. Freeman
(09/28/2007 01:35 AM)

Oh, by the way, I seem to remember reading somewhere that continued P membership required votes on some minimum number of ballots in a year. Lack of participation demotes a member country.