One fast track. Two investigations. 3522 answers. Still no convergence….

The title really says it all: The Ecma applies for a fast-track procedure at the beginning of 2007; despite objections, it’s going through the process with a razor-thin margin. Remember the fast-track proceedings are about checking major flaws in the specification beforehand. Nothing comes out of it, but when it goes through examination by the countries, thousands of comments, most of them pointing out to alleged issues, hundreds of them actually highlighting major flaws in OOXML, are being produced and published. 

 Back to our time now: On the 14th of January 2008, the final set of answers to the comments  is being published on the Ecma web site (note that you still need a password to access those so the public is left out in the wild). The closed OOXML specification (just because it does not match the publicity requirement makes it lose its “open” attribute) will thus be amended.  

But will it really be that much modified ? 

I don’t think so. I’m browsing through the answers made to the comments by the Afnor. What strikes me as odd is how many comments are actually  approved by the Ecma. This sure does change with the “engineers” we met back then in Paris who would literally refuse everything, unless the changes were from an editorial nature.

So these are good news until you actually read the proposed dispositions.They are located at the bottom of each addressed comments. They usually start out like that: “Agreed.” – Some blurb follows – and then the least technical of us realizes easily that he’s been fooled. The Ecma may agree with the comment at hand,  but its detailed answer ends up by proposing something that does or doesn’t modify what the proposal made in the original comment with arguments that usually go along the lines of self-justification, polite rephrasing of the comment that bends it to the Ecma’s view and ended up confusing the reader. 

Now, I have to praise the work of Ecma, no matter how opaque, obscure, strange and closed it went. They actually read 3522 comments and replied to them. I hope the international delegates will be able to review this amount of data in 4 days (!), because that just looks impossible to me.  Seriously, 3522 comments in 4 days published over more than 2000 pages? That makes 880 and half comment per day (ouch!) and around 36 comments per hour assuming that the delegates will work 24 hours a day… So the BRM will either have to examine a part of the proposed dispositions, or it will have to be turned into a parody of an international meeting of experts. A good way to turn the BRM into such a joke has been made possible by cautiously separating each set of comments by country. Be sure that the Ecma will ask that each country delegate only discusses dispositions that are contained within the set of comments received by its country. That will be a way to “gain some time”, or rather, to buy time and vigilance. Of course, the BRM should be an opportunity to discuss all the comments regardless of the country they’re originating from, but there will be no time for this. After all, 3522 addressed comments on a 6000 pages-long specification only make up for more than one objection every two pages.

As for me, I was able to review a part of the answers to the comments from Afnor, but what strikes me as odd is how these were taken into account by the Ecma. 

What did the Afnor ask? You’ll have more details, half of them in French, here.  The Afnor laid out an ambitious proposal that aims to merge OOXML into ODF. In order to do this, several steps have to be completed. The Ecma and the OASIS have to talk, and OOXML has to be “cleaned up”, split into two distinct parts, one fit for becoming one day part of the future ISO standard, and another one (mostly VML and the problematic features mimicking past behaviors) that would be excluded from ISO examination. The Afnor also made clear that OOXML was not satisfactory at this stage and thus could not be thought of as a potential standard; as an example, the long list of comments, around 600 of them, were sent. What this means is that what the Afnor really wishes is convergence, not necessarily having every comment addressed, as the convergence should run at its own pace and proceed in its own way. Yet this convergence proposal was really the only official comment made by the French committee. And it’s precisely what did not happen yesterday. Every comment was addressed, except the one about convergence (unless I’m being misstaken, but I don’t find anything related to it in the available documents). 

I hereby consider that the questions raised by France have not been addressed. And I find this extremely disappointing, because as difficult as this proposal was sometimes judged, the Ecma did not even try to address it. All this does not build confidence neither in Ecma nor in Microsoft.

Other questions were raised elsewhere, such as the actual will of Microsoft to implement the amended version of OOXML. At this stage, too many points are left unanswered; it is not even obvious that Microsoft itself will be able to proceed with all the expected changes to its existing documents’ base. I am thus irritated and I do feel frustrated over all this. I should perhaps not be surprized to learn about the two simultaneous investigations ongoing in Europe targeting Windows, MS Office, and… OOXML. Perhaps should I not feel surprized by this study either?

So let’s rehearse all this for one moment:  OOXML gets the fast-track approval by the ISO, moves on to formal examination by the international standards committees, gets blocked unless some conditions such as fixing thousands of issues are being properly addressed (wasn’t the fast-track meant to avoid this kind of problems?) obtains a period of  4 more days (the BRM) during which international experts will browse through 3522 comments hoping to find a way out. Meanwhile, two investigations target this specification and its reference implementation, MS Office while a study links the affirmative answer of standards committees to the level of corruption measured in their country.

What else do you need? Ah, you need that below…