ISO has always worked on the assumption that no single interest will be powerful enough to game the system at the level of countries, while being desperate and brazen enough to want to. You can call this naive, but it worked for decades until now.

Jon Postel didn’t take spam into account when he designed SMTP; maybe he was naive, but it worked for over a decade. For that matter, it may have been naive of IETF to create a process that would lead to the internet’s daily running being based on hundreds of RFCs that were likely to never become STD. Companies have tried to take advantage of that, by claiming that RFC822; in fact, I remember IBM claiming that Notes didn’t have to follow the RFC822 standard because “ït’s not really a standard.”

But the IETF wasn’t discredited, and Postel even less so. In fact, the explosion of bad-faith business on the internet caused the better businesses to look more to IETF as the only ones who had any inkling how to solve the problems.

The same thing has to happen here–countries have to realize that ISO is the only real protection against Microsoft and against rubber-stamp standards bodies like ECMA.

Many people were fooled by Microsoft into believing OOXML was already and “open” “standard.” Some of these people created new projects, or even startup companies, to support OOXML on this basis. Presumably these people also believed that it was technically at least acceptable.*

So, imagine that you’re a CTO whose entire future is now bet on OOXML being a ubiquitous standard. You discover that it hasn’t been voted on yet, and appears likely to lose. Whether you love Microsoft or hate them and are planning a lawsuit doesn’t matter; your choice is clear: You shell out the $800 and delegate one of your employees to go vote yes. So, maybe Microsoft didn’t pack INCITS.

But you can’t make the same argument over ISO. No country has bet its technological future on OOXML. (Surely a much larger percentage of, say, India’s GNP comes from Microsoft technologies than Cote d’Ivorie, and yet India voted no.) The only possible explanation is that Microsoft deliberately packed the board.

* I have actually seen business plans for companies that declared that some Microsoft technology was a piece of crap, and by acknowleding this their new startup would be able to better compete with those who drank the Koolaid, but this is pretty rare, and I haven’t heard of any successes. Let’s just say that I almost invested in a company making Exchange appliances on this model, and I’m glad I didn’t.

Thus, the proposal fails.

Will ODF 1.2 and PDF be rejuvenated?

MS has launched a DDOS attack on ISO and the global quest for open standards. Isn’t it obvious?

I suppose the OOXML ballot has siphoned all the attention from the SC34 members lately. There was little time left to consider the issues in other ballots. The real test will be with ballots occurring post-OOXML.

It seems that the paralyzation just started:

See the result of this recent JTC1/SC34 ballot:

http://www.jtc1sc34.org/repository/0902.htm

Citing:

“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.”

Worrying … may be Microsoft could instruct his P-memberts to leave SC34 once this OOXML voting has finished. So, the “original” P-members could continue working ( this is the goal of SC34 isn’t it?: to do technical work and not to “overflow” ballots )

There is another scenario to look at.

Suppose over 1/3 of the P members don’t vote on a standard Microsoft doesn’t want approved. They effectively raise the bar to get the 50% P-member voting threshold.

I don’t know what is the usual participation of P-members to ballots but if there routinely is 1/6 of P-members that happen not to vote on a given ballot due to lack of interest or other reasons, then over 1/3 + 1/6 is over 50% which is the second of your magic numbers.

The key point here is there is no need to obstruct by actually voting no. Mere inaction will do.

In such scenario proving motive is hard. Non voting P-members don’t cast votes and don’t issue comments. They may even not hold any meetings at NB level and won’t leave a track record of meeting minutes. How do you tell a non-voting member doing obstruction from a non voting member that just lack interest in a particular ballot?

The only way to show motive that I can think of is to observe a trend in behavior over a series of ballots. Microsoft has the ability to jam such observations by generating noise in the behavior pattern. If they want to, they may themselves vote in the opposite direction of their interest, relying on the inaction of P countries to block the proposed standard. They also have the ability to call its army to vote somehow on insignificant (to them) standards and reserving the obstruction tactics to ballots that really matter. Both tactics will make behavior patterns harder to show and harder to attribute to Microsoft actions.

I don’t say this is Microsoft strategy and I don’t want to suggest they intend to do something like this. With the evidence we have, such statements would be conjecture at best. But as you said, they have the capability to pull out this sort of things and it is prudent to consider the consequences of such action.

If you want to look at it that way, there are three magic numbers:

* 1/3 of P-members can prevent any ballot from passing by actively voting No.

* 50% of P-members can prevent any ballot from passing by passively not voting. Think of it as a “pocket veto.”

* 2/3 of P-members can make any ballot be approved.

So if someone wanted to be the “gate keeper” and be able to extort a toll from every standard that passes through JTC1, they would need to at the very least line up 1/3 of P-members and have them vote No on every proposal that they did not favor.

I certainly wouldn’t recommend this to anyone, and I don’t want to suggest that this is Microsoft’s motive. But they have assembled an army sufficient to do this, and it is prudent to at least consider the consequences if they decided to use it. One strategy certainly could be, if you can’t beat them, then destroy ISO and the other systems that produce open standards. The first thing the ugly queen does is order all the mirrors in the kingdom to be broken.

Then the next question is how much does Microsoft need a working JTC1? Or more to the point, how much do they need a *non* working JTC1?

They need it working fine to fast track things like XPS and Silverlight into ISO standards.

On the other hand they have been burned with ODF. They may not want more ISO standards they don’t control that could get in their way.

My understanding of your explanation of the P to O demotion process is it is designed to handle a single inactive P member in an otherwise well behaving group. It is not designed to handle the paralysis resulting from multiple inactive or obstructing P members. Am I correct?

Now that the techniques to hack ISO are known, denial of service attacks are possible if a sufficient number of P members only vote on matters of interest to Microsoft. They may not even need to vote No on other matters to stall JTC1 if it means JTC1 doesn’t meet the 50% voting threshold.

By the time someone figures what is going on and the demotion process is done, more new P members may join in and take the relay. This can go on for years before the process is fixed and JTC1 can function again.

I don’t say Microsoft will do this. I just say that the possibility is there, available for use to anybody with enough motivation and resources to use such tactics.

It is scary what you get when you apply what you learned about computer security to ISO standard processes, isn’t it?

Now there is procedure for a P-member to be demoted to an O-member if they miss votes and fail to respond when reminded of their voting obligations. The NB could then reapply for P-membership after 12 months. I do not see this rule being currently applied.

So in theory this problem is correctable, though a number of ballots could still fail while awaiting this procedure to kick in.

Maybe we need a new category of membership “JTC1 M-member” who are only counted for Microsoft/Ecma fast track ballots.

If you look at recent ballots you’ll see that there is a core membership that participate and vote on most proposals. This core membership previously constituted over 50% of P-members, but think this is no longer true. So we may have a problem on our hands.

The following link shows outstanding ballots fro JTC1/SC34:

http://www.jtc1sc34.org/document/secretariat_temp.html#calmtg

If I read this correctly combined with your post, is 50% participation required for this business to complete? There are six outstanding ballots between now and the end of the year.

I think we saw in Sweden what would be the easiest action. The real problem in Sweden was that MS had bought votes, ie, blatant corruption. Instead of going through the difficult and uncertain way of starting legal action, the Swedisch committee rejected the vote on a “technicality”: One member had voted twice.

I would think the most we can expect from ISO is that they throw MSOOXML from the fast-track on a technicality. Eg, the fact that MS’ online patent pledge falls far short of the required “IP” formalities (as I understood). Or the fact that MSOOXML clashes with another ISO standard (eg, the rediculous date problem). MSOOXML can still become an ISO standard, just not in fast track. Say, use the French proposal of making it a technical “standard” for three year.

Then ISO does not have to change its rules in a hurry, and does not have to challenge MS in court. They can even formulate it in ways that would save MS’ face.

Winter

I was looking for exactly the same numbers on JTC-1 evolution but I have different results.

From notes taken when ODF was ratified, I have 24 P members in May 2006 and another note with 30P members in January 2007

However, you say + 4 in 2006; + 12 in 2007 while I have + 6 and + 11

As your vote total of 43 P members is obviously wrong (the ISO web sites today identifies only 41 P countries), I think you may have a problem with your numbers.

Could you share your sources about P member evolution at JTC-1?

ODF will need two elements to appear as superior to OOXML. One is the spreadsheet parts included in the standard. This is already on the way…we must make sure that all voting members at the feb meeting understand this.

The second thing is that ODF need to have macro support in the long run. Most people probably agree that macros don’t serve much purpose except virus spreading…yet their absence mean that Microsoft have something critical. They can state that it impossible to convert old office documents to ODF since the macro part, that “might” be needed by their customers, can’t be converted. If ODF work included a branch that works on macro extensions the whole reason to have ooxml as standard becomes void.

Could an open source work group be started that examine how ODF might be extended with macro capability? If there was such group I think it would weaken microsofts position even if the group had not any official position. Even better if the ODF crew could start up work with macros themselves…they don’t need to make much progress or promise that the macros will be included in the end. Just the fact that ODF works with the problem would make Microsoft unable to claim that their new standard solves something that ODF lacks.

Finally this brings us to what we can do…surely the fact that Sweden changed its vote show the way to go. SIS in sweden used a formality, that by logics did not affect the outcome of the vote, to change the vote to what was morally right. The reason is most probably that the papers wrote about it. When nobody watch people are inclined to avoid to take the fight, they are just a single vote and it is very unpleasant to stand on your own.

Yet if main stream papers has highlighted how flawed ooxml is, what kind of countries that support the standard, that the final ODF standard will handle everything despite what Microsoft claims and how Microsoft tries to abuse the fast track procedure…the SB personal will have the situation that their credibility will be challenged if the don’t reject ooxml. They will be opposite the public opinion if they vote postive.

Make main stream papers write about what is happening and Microsoft will be doomed. If only blogs like this are source of information about what is wrong with ooxml most voting persons next year will probably never have heard the valid critcism because the public opinion does not require them to really investigate the criticism.

Ecma has a standard presentation on their web site describing what they do. In the section titled “What is Ecma’s value?”, they state:

==========================

Offers industry a “fast track“, to global standards bodies, through
which standards are made available on time;

… balance between timeliness and
quality …

Balances Technical Quality and Business Value

Offers a path which will minimise risk of changes to input specs

==========================

The emphasis shown above is as in the original text. In other words, Ecma themselves claim that you should use them to ram your proposals through the ISO as quickly as possible and without changes. This isn’t some wild claim being made by a Microsoft opponent; these are Ecma’s own words on why you should go through them. I can’t imagine a more damming indictment.

Given this and given the current fiasco, it is difficult to see why the ISO should maintain the current fast track special relationship with Ecma. Ecma can’t provide a reasonable review process which substitutes for the ISO’s own; indeed as they admit themselves, they don’t even try to.

TC45 was the Ecma committee that produced OOXML. TC46 is the Ecma committee that is working on XPS, which is Microsoft’s proposed replacement for PDF. In the XPS presentation (from the Ecma web site), Microsoft draws out a comparison between OOXML and XPS, where they see XPS as following directly from OOXML. In other words, we are looking at repeating this whole process again with another standard.

The ISO should suspend Ecma’s fast track privileges until and unless Ecma can show that they can provide adequate review of standards with openness and proper public participation. If the ISO doesn’t take this measure, then Ecma will be a millstone around the ISO’s neck dragging it down into disrepute and irrelevance.

I think it’s pretty likely that the strategy will be to move more of the yes-voting O members into the P group. That is certainly the path of least resistance here.

Then it’s only necessary to change a a single No to a Yes.