My assertion is that the 662 documents are a private Ecma consultation with NB’s, not part of the ISO process. You’ll have to take my word for it — and no one has contradicted this — that these 662 PDF’s are hosted on the Ecma website, with an Ecma controlled password, with Ecma assigned ID’s. They have not been submitted to ISO, ISO is not distributing the documents and they have not been assigned ISO “N” numbers. Therefor, ISO has no say in their access control.

There is an important principle at stake here — NB sovereignty. If Microsoft can successfully argue that a private, unofficial communication between two JTC1 member is covered by ISO access controls, even though the document is not submitted to ISO, then every internal document shared from one NB to another is similarly protected.

Consider also JTC1 Directives, 9.9:

“When a document is out for ballot at Stage 3 or higher, NB/Liaison organisations are free to circulate their comments to other NBs provided they do not use the formal SC or JTC 1 documentation distribution system. Formal distribution is prohibited because it could create confusion as to the status of the ballot. Documents out for ballot at Stage 3 or higher are not to be subject to formal discussion at any working level of JTC 1 during the balloting period. Therefore, NB positions on the document under ballot are not to be formally discussed at any working level. Circulation of such comments shall have no formal status within JTC 1 or its SCs, i.e. they shall not bear any document number nor shall they be considered in any ballot resolution meeting unless they were formally submitted to ITTF as comments accompanying the ballot.”

In other words, these comments created by Ecma are not, and cannot be, official JTC1 documents. So one cannot argue that they are governed by ISO access controls, which (according to JTC1 Directives, Annex HD) apply to official JTC1 documents, hosted and controlled by JTC1.

It does feel a bit like you are playing fast and loose with the facts here. While it is true that “ISO has no control over the process under which a standard is developed in OASIS or Ecma.”, it seems you are almost purposefully confusing the process that came before the vote in September with what comes after. As I understand it, the rules about openess NOW are set by the JTC1, and Microsoft is not keeping things quiet, but just following rules. It may well be that they have made the comments inacessible, but the vast majority of your readers have no way of knowing that except taking your word on it. This is not one of your better posts, as it seems to willingly to fall into the “anything Microsoft does must be wrong” category, which is not only annoyingly partiosan, but actually hurts your case about OOXML by playing into the hands of those who find you overly partisan. Just my opinion, of course.

I’d like to say that IBM made some great effort and fought to make the OASIS ODF TC documents available to the public, but I can’t claim that distinction. The openness is standard procedure for OASIS and part of their rules. All OASIS TC’s have their mailing list archives open to the public, etc.

If ISO (or ECMA, or anybody) accept the odd hundred changes as desirable, and then DIS29500 is changed and a standard is issued, will anybody produce a conformant product ?

Microsoft Office 2007 is what Microsoft Office 2007 is. Maybe it conforms to DIS29500; maybe it doesn’t.

I presume that tying down OOXML will be like tying down Microsoft’s ‘Win32′ interface. If you wrote an application to ‘Win32′ as expressed in Windows95 (and I have a fair number of ‘Edutainment’ CDs which were), most of them don’t run now.

‘Win32′ changes. The old applications aren’t supported any more; I can’t get them revised, no more than I can get a 5-year-old newspaper revised. It’s water under the bridge, history, gone.

Important stuff needs stable standards. POSIX and ISO26300.

No, had that been the case, there would be 662 OOXML files, not 662 PDFs, on their website :-)

What we have is an artifact of the process. The review and comment-generation phase was done in the open marketplace of ideas. We had the wiki on GrokLaw, numerous observations on my and other blog, the BSI wiki, open forums, public input, debates, etc. So there is no surprise that a number of issues show up on more than one NB’s list. I think this redundancy makes the process more open, robust and less suceptible to manipulation.

Imagine the contrary. For example, imagine that only one member of one NB notices a particular serious security flaw in OOXML. The flaw is reported, in secret, to the NB, discussed in secret in the NB, and then squashed by Microsoft who shows up at the last minute with 20 business partners and votes to approve OOXML without comments. The security flaw would then never see the light of day.

A more reasonable process, IMHO, would be to have JTC1 NB’s contribute to a running list of ballot comments, that can be edited and viewed by all JTC1 participants. But when it comes time for the ballot, each NB votes on its own and references the comment numbers of the comments that they are concerned with. That would eliminate, or at least reduce, the duplication issue while still allowing a fuller consideration of the issues.

Of course, ISO has barely moved off of an entirely paper-based process, so it will take some time before they progress to a level where they can manage a real-time collaborative endeavor.

I don’t expect to find an ISO standard for a videotape. I do expect to find an ISO standard for a DVD. If I want to play a videotape, I’ll get a videotape deck (JVC, single vendor); if I want to play a DVD I’ll get something that works according to the ISO standard.

Microsoft Office works just fine for what it’s intended for; productivity for office workers, who happen to be using Lenovo-type Personal Computers.

But the documents it produces are useless for doing anything else with. Indexing. Machine parsing. Filtering on mail gateways. Editing on XBoxes, Playstations, and Wii’s. Editing on Apples.

ISO issuing a standard won’t help any.

And yet, I get the impression that Microsoft has almost entirely succeeded in eliminating the question of competing “standards” from the ISO deliberations, reducing debate to relatively minor technical clarifications. (Even our supposed friends from GNOME have been claiming that the question of whether we should have two standards is “political” and “not relevant to ISO’s process”.) Am I mistaken?

If not mayby someone else can. That way everybody can just use a link to reference the document and also it could be upgraded when more comments become available?

It is a bit strange that you are complaining about the duplicates in the responses. It seems that a lot of duplicates are a result of IBM representatives submitting a large number of identical comments to the national bodies of more than one country. Mayby you could enlighten us in this area ?

OCXML:

Office Closed XML

Since its beginning all the process of its development has been kept closed.

At this stage, things haven’t changed.

Shame on ECMA, Microsoft and ISO.

We ( the final users ) should be the main part in this process ( we are the ones who “suffer” the result of this standardizations ) and we are kept outside of it.

Shameful

we have all the original comments available and being consolidated in a number of ways, such as duplicates, not relevant, issue of substance, complete show stopper etc.

This is all freely available at http://www.dis29500.org

In the end, the divergent behavior of this function reflects Excel’s divergent codebases on Windows and the Mac. You see this in other places as well, such as the way dates are treated. Date 0 is January 1st, 1900 on Windows, but January 1st, 1904 on the Mac. OOXML perpetuates this mess, throwing in a legacy bug in leap year calculations for good measure.

IMHO, an ISO standard in 2007 should not be bringing along legacy bugs and hardware constraints from 1990.

Back to CHAR() — the easy way to fix this function is to have an optional 2nd parameter, so the user can pass in a string indicating character encoding they want to use for this conversion.

Um, WTF? Why have a function work differently on different *platforms*? That’s just insane. Having it work on the character set of the document would make some kind of sense, but on the character set of the platform?

Further, what is the “ANSI” character set? If it’s a “Windows code page” (which are not defined by ANSI[0]) then it should be specified as such instead.

It also raises the question as to what should be done on Linux systems that for the most part use ISO-* character sets, and not the Windows-* deviants of these.

And I’d wonder whether a Linux system should use the system’s default character encoding, or the user’s current character encoding. (Can individual users specify which character encoding to use for their applications on Windows?) In which case you could get the same document giving different results on the same computer when viewed by different people.

One thing to note here is that MS explicitly do not support UTF-8 as an non-UCS2 encoding[1], while most Linux distributions are moving towards putting everything in UTF-8. So it would likely be the case in the near future that Linux and Windows users would not share a common platform character set, even if they spoke the same language. (e.g. Windows English British in Windows-1252, and Linux en_GB.UTF-8) That’s going to make for some fun debugging, I’m sure…

[0] http://en.wikipedia.org/wiki/ANSI

[1] I can’t bring myself to call UTF-8 a “non-Unicode encoding”, which is what Microsoft terminology demands that it is. They call UCS2 “Unicode” and all other character sets “Non-Unicode”. But that’s clearly wrong as there are a number of non-UCS2 character sets that are also Unicode encodings. But if you try to explain this to someone who’s only read MS literature, they just don’t get it. You can’t call non-UCS2 sets MBCSs either, as ISO-* are not multi-byte.