Last Days for Office Open XML Ballot at National Bodies: plus Office XML Lite

by Rick Jelliffe

The deadline for the National Bodies to vote on DIS 29500 Office Open XML (fast-tracked from Ecma 376) is coming up on September 2: this is the vote that comes at the end of the 5 month review, which is 7 months since the draft was submitted, and probably not the end of the line. Here is some of the news, as a companion to my next blog item, which is about what I recommended to Standards Australia for our vote.

There are not many informed ideas of where the votes will go. The US body looks like it will vote Yes (I predicted an abstention there) as seemingly will Germany (I would have predicted a "No"). India looks to be voting "No with Comments", which I am pretty happy about since that I commended that vote to them when I was there.(Note that that news story gets it wrong about the impact of a "No" vote—a simple"No" and "No with Comments" are utterly different beasts, as Jon Bosak has prudently pointed out. (But even a "No with Comments" may not, in effect, be a conditional yes if the comments are impossible to fulfill: obviously no National Body will put in merely vexatious comments, they don't want to wast time, however, they will state their requirements in a clear way that allows rapid resolution of issues.)

It seems likely that there will be a Ballot Resolution Meeting: if there are not enough "Yes" votes and enough "No with Comments" (i.e. 'conditional yes') votes, then a meeting is scheduled to see what technical changes need to be made to satisfy enough national bodies' requirements. A meeting has been scheduled for Feb 25-19 in Geneva, with UK's Alex Brown appointed as convenor.

There is a last minute frenzy on the contra side: IBM's spokeman is claiming lots of alarming shenanigans without actually giving us the benefit of any details: names of countries, parties, dates, anything tangible. Stephane Rodriguez is complaining that he has to look at the schema and documentation when editing Open XML files; his new blog is notable for the number of times it says that there is a problem with OOXML but actually refers to the some implementation issue in Office 2007

Behind the scenes, Patrick Durusau (the ISO ODF editor) has been working on a really interesting and useful project. While he is not keen that people use ISO Open XML, he is keen that the quality of ISO standards should be maintained and he sees OOXML as a way to get MS' technical requirements on the table to help future ODF improvement (whether by cherry-picking, mix-n-match or knowing what to avoid.) I suggested to him a time ago that one approach to fixing DIS 29500 would be radical surgery: removing all the explanatory and non-normative material. At the moment it is far too tutorial. That is fine for the Ecma version, but gets in the way of an ISO-quality standard. I had also suggested that the schema fragments were otiose too, and that the 11pt body text should be 10 pt. . So Patrick has gone ahead and stripped out the fluff from the WordprocessingML chapter and with tighter formatting he was able to go from 1874 pages to 607 pages without altering the technical content!

Knowing Patrick, I expect he would not release his Open-XML-Lite because having extra drafts floating about in public just provides more fodder for the lunatic fringe, however he sent me a copy and I think it is great. It is a real proof of concept that there is indeed a workable spec lurking inside DIS 29500. I would really urge National Bodies to include comments that request or require that the non-normative material in DIS 29500 be removed. That will make maintenance, editing and use much clearer. The Ecma TC45 got it wrong here; or, at least, they went with the "friendly" view of a standard, where it is best that a large ISO standard avoids being too tutorial.

I am hoping that once the vote is over, the PR considerations of the big boys will take a back seat: with a couple of "Yes" votes from some large countries, MS has its marketing material to say that Open XML is credible; with a couple of "No" votes IBM has its marketing material to say that ODF is the way forward; and with enough good comments and a sharply-run at a Ballot Resolution Meeting, the baying mobs will lose interest; and the nerds can get down to improving the shortcomings exposed in both OOXML and ODF., If things go as the process is geared to make them go, IS 29500 OOXML and IS 26300 ODF should ultimately provide a really useful pair of technologies.


M. David Peterson
2007-08-28 01:14:20
Fantastic summary, Rick! I'm looking foward to reading your mentioned follow-up analysis and recommendation for Standards Australia.
2007-08-28 23:40:58
Amusing is that when Microsoft submitted it's standard to Ecma is was about twothouands pages.
Ik could be that that ias actually it's logical size when removing the fluff (as you call it). However I must say that the use of examples makes such an interpretative spec as a big Office spec a lot more easy to implement.

So I actually think you might be wrong there. Maintenance may be more complex because of the added examples and explanations but as a pure syntax I think there might be a lot of trouble getting interoperabel documents.

Especially seeing a Micrsoft has such a big headstart in it's implementation it is quite usefull to have more guidance.

Robin La Fontaine
2007-08-29 03:04:44
Agree with you re normative material only in the spec, it is far too large to read as it is and many repeated examples which clutter it up. A clear distinction between normative/tutorial is always good.
Stephane Rodriguez
2007-08-29 10:55:11

Implementation issues? Not really, there is a ECMA 376 excerpt in almost every section of the article. And it's a different one each time.

Granted, the article is not complete yet. If you have implemented even 1% of this mega crap, you know there is no way there are only 10 problems with it.

It's important to read the article with the right angle. Miguel de Icaza failed that, most notably in his pseudo-rebuttal of my article (which got slashdotted).

I have just added a section today related to document security (or lack thereof). Have a good read.

Rick Jelliffe
2007-08-29 11:09:36
Stephane: You should get hold of the report commissioned by the French Government on security in ODF and Open XML late last year. (It has not been publicly released, but some of your contacts are bound to have a copy they can share with you.) It found very large potential problems in both, though to me many of the problems looked easily solvable by simply not running on Windows.

You are a passionate guy, and one with the courage of his convictions. I am not a spreadsheet user, my expertise and current interests are in schemas+validation, publishing, governance, internationalization and standards in roughly that order, so I will let users who are interested in spreadsheets pursue it through your and Miguel's wrestling.

2007-08-29 11:23:09

"You should get hold of the report commissioned by the French Government on security in ODF and Open XML late last year."

And what do you say when you learn that document security in Excel's BIFF was better than Excel's SpreadsheetML because the audience were unable to find and delete the password, and now they can? Does not this qualify as a regression?

I don't care Miguel, he has been working for Microsoft for the last 5 years (taking a pride not to be on their payroll officially despite the fact it makes exactly no difference in practice). Imagine if the guy said anything negative about this stuff : bye bye Redmond, bye bye the contacts, bye bye the protection from lawsuits. I work as an independent.

Stephane Rodriguez
2007-08-29 11:23:54

The anonymous above is me.
Rick Jelliffe
2007-08-29 12:19:41
Stephane: Oh, nothing surprises me. The "security" in SpreadsheetML is more like a field-locking feature for the user interface IIRC, but I haven't looked at it for months so I doubt my recollection. The OPC security features provide a better level of protection, but they are just documents.

This is where one would expect the ISO security committee (whoever they are) to come along and make a meta-standard for document security. Armed with this, standards makers and reviewers can develop the standard more reliably. (I happen to think this kind of thing is more likely with MS inside the house rather than lurking outside.)

Are Gulbrandsen
2007-08-30 07:32:21
Big headlines in Sweden and Norway now. Sweden woted yes after a coup by Microsoft partners.

English summary:

Rick Jelliffe
2007-08-30 12:46:09
Are: Yes, I think there are a lot of businesses in the MS ecosystems around the world who feel that a standard would be beneficial, and who (belatedly) won't let rivals (who have their preferred standard already) get away with blocking the MS technology. It is hardly surprising, except perhaps to people who think that MS somehow has no users or co-stakeholders. (The real Stockholm Syndrome :-) ? )
2007-08-30 15:20:43
Why is it that even Microsoft Certified Gold Partners, some of whom have businesses dependent on Microsoft Office, won't naturally vote in favour of MOOXML without 'market benefits'?
Rick Jelliffe
2007-08-31 03:09:37
Daron: Same as why Open Source people have avoid standardization at ISO. Not their habit. Also, probably they have a client mentality rather than a community mentality.