Stephen McGibbon's Web Journal
Voices for Innovation is a global community working together to shape the technology policy debate. We want every voice heard. Join now and make a difference.

Recent comments


Durusau: "OpenDocument, among others, will lose if OpenXML loses."

Patrick Durusau has posted another thoughtful contribution to the document formats standardisation debate:-

Who Loses If OpenXML Loses?

As national bodies meet to cast their final votes on OpenXML, a checklist of who loses if OpenXML does might be helpful. This is just my list and isn't meant to be exhaustive:

  1. National bodies lose an open and international forum for further work on DIS 29500.
  2. Microsoft based third-party vendors may be excluded from contracts because Microsoft has no ISO approved format.
  3. ODF has no ISO-based formula definitions to insure compatibility between OpenDocument and OpenXML.
  4. ODF has no ISO-based definition of MS legacy features for an ODF extension.
  5. ODF has no ISO-based definition of the current MS format for mapping purposes.

As the editor of OpenDocument, I want to promote OpenDocument, extol its features, urge the widest use of it as possible, none of which is accomplished by the anti-OpenXML position in ISO.

Passage of OpenXML in ISO is going to benefit OpenDocument as much as anyone else. Here are some specifics:

  • OpenDocument currently lacks formula definitions for spreadsheets. (To appear in OpenDocument 1.2.) Many core financial functions in spreadsheets are undefined except for actual Excel output. That output varies by version and service pack of MS Office. What happens if OpenDocument and OpenXML reach different definitions of those functions?
  • OpenDocument does not presently support legacy features of Microsoft formats. That will be easier with a formal definition of those features. Without OpenXML, OpenDocument has no authoritative definition of those legacy features. That delays OpenDocument supporting them in some future release.
  • OpenDocument does not have a robust mapping to the current Microsoft format. That requires an OpenXML that has completed the standards process. If OpenXML is unclear, it must be fixed in order to create a robust mapping between the two.

The bottom line is that OpenDocument, among others, will lose if OpenXML loses.

Covington, 24 March 2008

Patrick Durusau

I said as much recently in an interview in Norway.


Anon said:


"So, sorry Mr Durusau, you have not convinced me that "OpenDocument will lose if OpenXML loses"

Wow Luc, I'm sure Patrick Durusau will be losing his sleep over this...

# March 24, 2008 7:11 PM

.NetBlogger said:

La Dirección  General  de Normas  emite mañana su voto, lo interesante es cual

# March 24, 2008 7:31 PM

Rob Brown said:


"Wow Luc, I'm sure Patrick Durusau will be losing his sleep over this..."

Thanks for your insightful contribution to the discussion.

Let's look at Mr. Durusau's points:

1) re losing a forum for further work:

If OOXML loses, then Microsoft either re-submits it via the "normal" track, and the forum is still available; or else Microsoft doesn't bother any more with ISO for this spec, and it loses relevance over time. I don't see how national bodies really lose in either case.

2) re Microsoft-based third parties:

Such parties will either re-focus on the changed marketplace, or they will die. It's business as usual. Don't forget that Doug Mahugh said that betting on OOXML before its adoption as a standard was a bad idea.

3) re Formula definitions:

OpenFormula is ready for standardisation. As we all know, standardisation takes a very long time but in the medium-to-long term (which is what standardisation is all about), the lack of formula definitions is a complete red herring. This is a very odd comment from the editor of ODF.

4) re MS legacy features:

Legacy features are "described" by OOXML, but no mapping is defined- neither from the legacy documents to OOXML, nor in the reverse direction. OOXML gives no guidance on how to process an Office XP or Office 2000 document. These are covered in the Microsoft binary format documentation. In general, mapping from legacy documents to (say) ODF has been achieved through reverse engineering and OOXML will not improve that situation greatly. Things that could be useful (like proper explanation of tags like "autoSpaceLikeWord95") have been promised by Ecma as additions to ECMA376, but they have not yet been delivered so their usefulness is currently unknown.

Also, it has to be noted that Microsoft itself passed up the opportunity to be represented on the OASIS ODF committee and contribute to Office-specific features for ODF. A lot has been written about the reasons for this, and it's shrouded in politics and conflicting stories, so I have no idea why it's happened like this.

My personal hope is: if OOXML is rejected at fast track, then the ODF proponents will extend an olive branch to Microsoft, which will be accepted. An accelerated program would begin, to add necessary features to ODF for Office support.

5) re ISO-based definition of the current MS format:

This is a pretty egregiously misleading statement. The fact is that DIS29500 MkII *does not* describe the Office 2007 format. ECMA376 is fairly close, but not exact; DIS29500 has had changes which make it less accurate. Even though NBs working at the recent BRM were prevented from making changes which would obviously break compatibility with Office 2007, sections describing Office2007 functionality were moved to legacy annexes, and new tags were introduced which no Office 2007 application will understand.

DIS29500, as it is, may be useful to people implementing applications *consuming* OOXML files, as long as they were generated by Office 2007. Because it lacks semantic definitions, it is next to useless for people implementing applications *producing* OOXML to be consumed by Office 2007. That extra information will have to come from (sigh) reverse engineering. MS has produced a candidate "standard" which really works to entrench the requirement for Office 2007 in the document chain.

# March 24, 2008 9:16 PM

nksingh said:

@Rob, @Luc:

Let's ignore the fact that you guys are advocating the use of ISO to accomplish through law what cannot be accomplished through fair competition.

A much more plausible scenario is that OpenXML becomes the de-facto standard regardless of ISO standardization because private companies do not have any requirement to go with some national standard.  Office will include ODF capabilities in order to be compliant, but it will not be a particularly featureful implementation in comparison to OpenXML so it will more or less be a publishing format rather than a working format.  ODF sticks around for much the same reasons that OGG sticks around, but it will hardly be mainstream.  The only thing that will change is what particular lobbying battle is fought if governments start demanding an ISO format.  

# March 25, 2008 1:20 AM

RobBrown said:

Hi nksingh,

I'm *not* "advocating the use of ISO to accomplish through law what cannot be accomplished through fair competition". I do believe that "competing standards" in this area will be unhelpful. If you feel like discussing anything I've said then please go ahead, but dismissing my comment (and Luc Bollen's) with a misleading one-liner just looks a bit... lazy.

As for your prediction of the future, you may very well be right.

# March 25, 2008 3:03 AM

nksingh said:


Sorry, I don't mean to insult you.  I've found your comments to be generally quite respectful although you disagree with my perspective.  

If you plan to make a standard that purports to be 'the' document format for word processing, spreadsheets, and presentations, it stands to reason that such a format should take into account the needs of an extremely prevalent producer and consumer of such documents, namely the Microsoft Office family of products.  Trying to make an industry standard format without soliciting the input and requirements of one of the large players in the industry seems like more of a way to gang up on the big company than a way to produce truly interoperable documents that will actually help a real customer.  

My opinion is that OpenXML is a better starting place for the universal format (since it can already support much of the most featureful Office product on the market without any extensions), whereas ODF will have to be built up from a less complex and complete base.  It's a lot easier to strip down and generalize something that exists and works than to invent something new without proven implementations.  

I don't think we need competing formats either.  I just think that ODF is insufficient to be the one and only ISO standard document format.  I would not agree that it should lose its ISO status, because having a second format and committee would be good to keep everyone honest.  But in the long run, if ODF 1.0 were replaced by OpenXML as an ISO standard, the world in aggregate would not be a worse place.

# March 25, 2008 6:31 AM

Rob Brown said:


No offence taken. One thing that I should say is that if OOXML fails at fast track, I think it's a real possibility that IBM/Sun/whoever will take it as a victory over Microsoft, and start lobbying Governments all over the place to ditch MS Office in preference for other office suites just to rub salt into the wound. And I think that would be a great tragedy.

One thing this whole document formats saga has done is that it has raised the awareness of document formats and the possibilities that openness and standardisation can create. It's also raised awareness of dodgy dealings from both sides of the debate, and dragged ISO standardisation processes through the mud at the same time.

Any ideas on how to repair the damage and get back to the positive possibilities?

Rick Jelliffe said "the (immoderates on the) anti-OOXML side have so effectively poisoned the atmosphere by now, that any idea that Microsoft will be given a fair hearing in any ODF forum is ludicrous". I think that's overly negative, and I believe that should OOXML be standardised then Microsoft won't give *anyone* else much of a fair hearing. But it's obvious that not only Microsoft need to change their approach. If the OOXML fast-track fails, then IBM and Sun have an opportunity and a responsibility to try to turn the result into something positive for *everyone*.

At this point I start to feel like a starry-eyed dreamer waving a peace flag and singing "Why can't we all just get along?", so I'll stop right there ;-)

# March 25, 2008 10:52 AM

Luc Bollen said:


Like Rob Brown, I agree with part of your prediction of the future: it makes no doubt that OOXML will become *a* de-facto standard, regardless of ISO standardization.

Will ODF never becomes mainstream ?  I'm not so sure.  Linux is becoming more and more popular, for governments, big and small companies and general public.  A major weakness of OOXML is that there is no good, featureful implementation of an office suite making use of OOXML on Linux.  So, we can predict that ODF and OOXML are here to stay for quite some time.

In the long run, at least one will disappear (see Betamax and VHS), maybe both if a better format appears (like VHS has been replaced by DVD).

In the meantime, Microsoft will have to compete fairly, and abusing the ISO fast track process to try forcing approval of an immature document was not competing fairly.

Note that the claim by Microsoft that the spec is much improved from what it was a year a go is a recognition that Fast Track was not the right approach.  The purpose of Fast Track is to approve "an existing standard [...] submitted without modification" (see JTC1 directives).  The need to completely restructure the document between the BRM and the final vote says enough.

I'm against two standards to do the same thing, but in any case, if Microsoft want to get an ISO stamp, it has to gain it via the normal track, which will allow the spec to mature before being approved.

# March 25, 2008 11:14 AM

Stephen McGibbon said:

Luc I think it's great that you express your point of view here - I don't want this to be a Microsoft echo chamber, so let me start this post by saying a sincere "thanks" for contributing.

That said, I don't know what "angle" you approach this from as I know nothing about you save your name, but if you will forgive me your comments seem to be written from a sense of despair. I am honestly very optimisitc about what's happening right now, even in the furore of the document formats debate.

>A major weakness of OOXML is that there is no good, featureful implementation of an office suite making use of OOXML on Linux.

Well the next version of will address that for you I trust. 

but why should that be seen as a weakness of OpenXML in any case? Is it now a requirement that any IT standard has to have Linux implementation? Is it also a requirement that any IT Standard should also have a Windows implementation? or an OSX implementation?

>So, we can predict that ODF and OOXML are here to stay for quite some time.

Well we agree here at leastSmile. I hope ODF will be successful - the competition is good for Microsoft too - but you need to be realistic about how long that will take too. That's why I said in the Norwegian interview I linked to that ODF would be the loser if DIS29500 isn't approved. I don't think ISO26300:2006 is ready for many customer scenarios, ODF 1.2 will help but that's not what we have today is it?

>abusing the ISO fast track process to try forcing approval of an immature document

Microsoft haven't abused the ISO fast track process at all. Ecma 376 is certainly more mature than IS26300 was when it was submitted under the similar PAS process.

>was not competing fairly

in what sense is this not competing fairly? I think this is a case of damned if you do and damned if you don't.

>I'm against two standards to do the same thing

which two standards do that? It's generally accepted that DIS29500 and IS26300 try to do different things. This is good not bad. And in any case, as long as there are people who don't mind, or even welcome, competing standards the only way for your view to prevail is to dictate standards. If you do that you'd better have a much better governance model for the process of choosing which one to dictate - in my view any sensible assessment would lead to that being OpenXML and not ODF.

>if Microsoft want to get an ISO stamp, it has to gain it via the normal track, which will allow the spec to mature before being approved.

Here you go on the "need to mature" route again. Can you be more specific?

You clearly don't like OpenXML. Fine, you're free to use ODF. But then you object to other supporters of ODF who think OpenXML as an ISO standard will help ODF. Maybe you should consider running for office somewhere! I hear Digistan is looking for a benevolent dictator Wink.

# March 25, 2008 11:55 AM

hAl said:


[quote]Why would MS spend so much effort and money to have OOXML fast tracked,[/quote]

Why would IBM spend so much money to block ISO/SEC standardization of Office Open XML whilst others are already widely implementing Office Open XML.

There is relativly little problem in implementing Office Open XML compared to implementing ODF (which essentially requires reverse engineering showing already that the formathas much less technical problem than might be expected. It is much easier to reuse software in use for the binary formats for the new formats as the Office Open XML essentially supports the same featureset whereas supporting ODF is challenging because excetp for a few most do not have software that support the same features as ODF does. Also the lack of spreadsheet formulas almost three years after ODF was standardized in OASIS is a shocking sign that ODF maintenance is not going well. The ODF formula's alone are taking significantly longer than the entire ODF original specifcation has. This does not boast well for the complexity of implementing those totally new formula's especially if it will again be scarsely documneted and thus mostly about reverse engineering how OpenOffice implements them.


Who is paying all those organisations, like the ODF alliance and the OFE to oppose Office Open XML?

# March 25, 2008 12:46 PM

nksingh said:


I don't know if that's fair.  Maybe the ODF committee is waiting for OpenXML to be finished so that they can make formulae that are harmonized with Excel's.  

# March 25, 2008 4:32 PM

Rob Brown said:

@Stephen McGibbon:

When Luc referred to the lack of a Linux implementation of OOXML as a major weakness, I think he meant a weakness in the marketplace rather than a reason for not making it a standard.

You said "it's generally accepted that DIS29500 and IS26300 try to do different things". I don't know about "generally accepted", at least I don't accept it ;-). It all comes down to whether you believe that to represent a specific feature set, you need to provide explicit tags for those features in the markup. Debating that point is probably not going to be productive.

I'm seeing a lot of non Microsoft, pro-OOXML people assuming that a standardised OOXML will definitely be open, consensual, and independent of the mothership. For me that will depend entirely on the outcome of the ISO plenary in Oslo which will decide on the maintenance plan for IS29500 if it succeeds. I'd be really interested in other people's views on that.

@hAL: your trolling is very old, you need to change the record.

# March 25, 2008 7:16 PM

Brian Jones: Open XML Formats said:

I try to keep the discussion on this blog primarily focused on the area I care most about, the technology

# March 25, 2008 11:58 PM

rinkevichjm said:

Microsoft is  a convicted US Monopolist.  They were convicted for among other things using Windows to unfairly gain market with MS Word and Excel.  Any claims that not having OOXML as an ISO standard, right now as is, is unfair are sheer lunacy.  If anything the market needs a single standard that Microsoft didn't dominate to cure the monopoly issue.  OOXML right now as is, is an attempt by the convicted felon Microsoft not to have to compete and keep their monopoly.  OOXML needs MAJOR error correction (standard should be less than 1 error per 2 pages, OOXML appears to be at >1 per page), additional sections to describe how to convert pre-OOXML Office into OOXML format, SELF-HARMONIZATION, and harmonization with ISO/IEC 26300 among other things in order to become a viable standard.  

What does this mean?  It means it belongs in a JTC1/SC-34 WG at stage 2, not at FAST-TRACK.  

V1 should be voting NO.  Microsoft brought in partners and stacked that body in its favor.  ANSI rules require that committees such as V1 be free from undue influence of one party.  Mr. Durusau as chair of V1 is violating that policy.  He appears to be unethical in fact.  Maybe NIST should take control of ANSI like the government does in India with BIS.

IMHO MS is spending loads of money to get DIS 29500 approved, while IBM is spending the same amounts it usually spends on standards issues.  IBM certainly isn't bribing other businesses to join these standards committees, OTH MS has.

There is no reason to make ECMA-376 an ISO standard.  If you need OOXML, right now as is, use ECMA-376.  Otherwise, let it go through a standard ISO process beginning at stage 2.

# March 27, 2008 1:09 AM

Stephen McGibbon said:

James I hope you feel better getting that rant off your chest.

I don't see any reason for you to insult Mr Durusua just because you disagree with him. It reflects more on you than him.

As for the rest, well, stick to Groklaw and /.

# March 27, 2008 1:49 AM

Robert Millan said:

nksingh has a very funny idea of what is "fair competition".

You see, in year 2000 I created a PowerPoint document as part of my school work.  Shortly after, I abandoned Microsoft software completely.  The result?  To the date I've been unable to open my own document with any other presentation program.  It would simply not render correctly, because the reverse engineering work wasn't complete.  Microsoft effectively owns my data.

Today, if you create a document in Office 2007 format, an OOXML implementation won't get it right either.  Users are effectively locked in, and this is what "fair competition" means to Microsoft.

Face it.  MS-Office couldn't handle fair competition for more than a few years, and this is the whole reason it doesn't support ODF yet.

# April 1, 2008 8:12 PM