Cliffhanger!

by Rick Jelliffe

[UPDATE] I thought I'd give some graphs for the results of the ballot-changes of DIS 29500 mark II. These are the results as at Wednesday, and I think they are the finals. (There is one non-P NB whose vote I am not sure of: I have shown it as abstain though it could be accept.)

Here is a graph of all the votes case, showing the change from the initial ballot until now (as far as it is publicly known). This is based on all the NBs who voted. (However, this is not the count that is used to determine success...)
Graphic with estimates of total votes, showing that the absolute number of accepts has risen to the mid sixty percents while the absolute number of rejects has lowered to the mid ten percents


At ISO/IEC JTC1, national standards bodies (called NBs) nominate what kind of participation they are interested, for each of the multiple subject-oriented Steering Committees (SCs). They can nominate in two classes: Participating Members (P-members) are supposed to maintain an active interest, attend meetings, and vote on all the standard drafts that come up. Observing Members (O-members) can vote, but they don't have any obligations to show up to SC meetings.

Here is a graph of all the votes case, showing the change from the initial ballot until now (as far as it is publicly known). This is based on the NBs who are O or P members (However, this is not the count that is used to determine success...)
Graphic with estimated  voting ratios for P and O NBs

Here is the vote when you just look at the P members (as far as it is known.) Note that "abstain" votes have a very particular meaning in ISO: it does not mean "reject" or "protest", it means that the voting body could not decide, or is happy let the consensus of other NBs determine. There is no shame or difficulty with an NB voting abstain. (At earlier stages of drafts, there are "No with comments" votes: these often are "conditional yes" votes, which can explain how a "reject" vote can become an "accept" vote. At the current stage, however, no means no.)
Graphic with estimated  votes for P-member NBs


Finally, now we have seen the big picture, we come to the real numbers that count. There is a negative test and a positive test. First, no more than 1/4 of all NBs who vote can be negative (ignoring abstains). This has not been reached (i.e. not enough rejections: the all-nation acceptances are over 75% on the following graph.)

Graphic with estimated  voting ratios accept to reject for all NBs, showing over 80% acceptance rate

Then there is a positive test: at least 2/3 of all P members who vote should vote for acceptance (ignoring abstains again). This has been reached (i.e. enough acceptances: the P-nation acceptances are over 66.7% on this graph.)

Graphic with estimated  voting ratios accept to reject of P-member NBs, showing over 70% acceptance rate

So OOXML has been accepted, seemingly by 24 to 8, which is enough of a margin to avoid "hanging chad" clawback games.

It is clear that most NBs think DIS 29500 mark II makes a credible and acceptable or useful standard, but there is a substantial and active minority that does not.