Where is IBM on Open XML?

by Matt Asay

Perhaps not perfect, but Microsoft has done a decent job of trying to standardize its Open XML format. As Jason Matusow writes:
Customers have been very clear with us over the past few years (as have many other vendors including our friends in Armonk) that they wanted to see our Office document formats become more open and standardized. So we did that. (OSP-licensed, Ecma standardized)

Governments have been clear that they need the ability to have interoperability between ODF and Open XML. The Open XML Translator is now in production, and delivers interoperability. In fact, we built that to enable ANY ISV to use the technology - not just Microsoft. Novell has already announced (back in December) that they are going to build it into Novell's OpenOffice. Sounds to me like customers are going to have greater choice.

I have attended open source conferences for the past 6 years, and sat on innumerable panels with various executives from IBM. I am really unclear as to the relationship between the rhetoric of openness and increased choice that they have been saying in that arena and how it lines up with the reduction of choice and closing of a participatory process in this arena. The message from IBM standards participants around the world has been consistant: don't even consider Open XML for ISO/IEC standardization. That is less choice for customers.
Why, indeed....

Well, Jonathan Murray thinks it may have something (actually, everything) to do with IBM's business model and its fiduciary duty to its shareholders:
IBM's position on the Open XML vs. ODF standardization debate is in no way altruistic. IBM takes the position it does, not to make life better for the Open Source community or to advance the position of free software. IBM takes the position it does because this position ultimately creates more value for its shareholders. Period. It is a pity that they seem to be doing this against the best interests of their customers.

The question is why IBM would be taking this position when the risk to its reputation with its customers seems to be so high. My personal belief is that the 180,000+ hungry mouths in IBM's global services division are a big part of the reason.

The importance of global services to IBM cannot be over stated. IBM is a services company far more than it is software or even hardware company today....

So what does this have to do with the IBM's campaign against the Open XML standardization process? In a word: complexity.

[I]f you are in IT services business...complexity is your friend. Any reduction in complexity dilutes the value you can offer to your customers. This, in my view, is why IBM seems to be so focused on preventing customers from having to right to choose between two open standards for their document formats.
Jason has apologized for making snarky comments about IBM's lack of enthusiasm for OpenXML, but he needn't have. IBM should be called into account for its continued intransigence vis-a-vis Open XML. It's not a perfect standard/process, but nothing ever is. I personally don't believe that this opening up of OpenXML is a big deal, as I've written before. An open file format is not going to move enterprises off Microsoft technology any time soon, especially given that SharePoint, not file formats, is the new battleground.

Regardless, Microsoft should be given kudos for OpenXML, and IBM should be a bit ashamed. IBM has been given a lot of love for its open source support, but look a little closer and all you see is support for Apache-licensed projects (and, of course, Linux). Like any good corporate citizen, it feeds itself before it worries about feeding others. But word on the street is that IBM won't consider buying an open source company unless its licensing is Apache-style. I guess it likes to consume open source but doesn't like free source-requirements to give back....


2007-02-12 10:18:28
Are. You. Insane? Have you even bothered to review the 6000 page OpenXML document? If you had, you may have actually noticed that 1900 is a leap year according to Microsoft! And, shockingly!, Microsoft doesn't seem to be terribly concerned about consistency with present standards -- they are more intent on inventing their own. Before you open your mouth about standards, you might want to do some actual research on them.... For example: http://www.grokdoc.net/index.php/EOOXML_objections
2007-02-12 10:38:56
IBM doesn't want OpenXML to become an ISO standard because it creates complexity that they cannot bill for. OpenXML, if implementable by a non-Microsoft party (debatable) is hugely complex and doesn't provide IBM with something they can point to as a value add to their customers. Therefore they would have to underwrite support of this complexity out of pocket. How does that help the shareholders?

As for your armchair economics, you should think harder about conflicting "standards" does to the customer. This is not two car companies providing essentially equivalent cars, and letting competition drive down the cost of those cars. This is two car companies trying to sell you a car that only works on one kind of road, and then proliferating those roads and passing the cost on to you. Open data is like the vote - useless if not completely interchangeable.

2007-02-12 12:57:19
William: Explain to me how it is OK for HTTP and FTP to co-exist, yet OpenXML and ODF cannot also co-exist? Why is it OK for PDF and ODF to co-exist, yet OpenXML is not OK. How can PNG and JPEG 2000 co-exist w/o causing confusion? Heck, why even let ODF become an ISO standard when ODA (Open Document Architecture) has been a standard since 1999?

Open-source advocates are always about "choice" (Linux distributions, text editors, programming languages) but not when it suits their interest.

Fed up with this Shillery
2007-02-12 18:22:35
How much did your soul cost? Must have been cheap.
M. David Peterson
2007-02-12 20:14:55

You are aware that the 1900 leap year bug is actually a Lotus 1-2-3 bug that MSFT has purposely perpetuated for *compatibility* reasons, correct? Of course, given that you seem to know so much about this stuff, you would then know that Lotus is an *IBM* brand, correct? See: http://www.oreillynet.com/xml/blog/2007/01/odf_vs_ooxml_from_an_insider_o.html#comment-477405 for more information.


It's always nice to see people spending the time to dig a little deeper to find the real story behind the madness. Thanks!

M. David Peterson
2007-02-12 20:16:37


Aaron 'Teejay' Trevena
2007-02-13 02:29:51
Right.. so it's a GOOD thing that OOXML preserves a bug required in ancient versions of microsoft software in order to be compatible with a bug in even more ancient and no longer used spreadsheet?

Sorry, the problem most OSS people have with OOXML isn't that it's Microsoft, it's that it is so deeply tied to the arcane intracacies of Office, and that includes a huge ammount of cruft that's accumulated to ensure decades of backwards compatibility to ancient workarounds, bodges, etc.

This means that there will NEVER be any application that implements OOXML fully except Microsoft Office, even with the efforts of Novell and Microsoft, the OOXML standard and the documents that will be almost as arcane, complex and undocumented as the current Word Document file format, and we'll see the same niggling incompatbilities as we see now.

I've seen examples of OOXML and ODF XML and I'd much sooner be dealing with the latter than the former, even though the spec is handwavey - there are multiple compatible implementions.

Also HTTP and FTP aren't really comparable - they do different things for different jobs - OOXML and ODF do exactly the same job, PDF and ODF do different jobs - PDF has always been primarily for Print and readonly distribution, ODF is general purpose.

As a user of software I want choice of software, as a developer of software I don't want to have to write the same thing twice but to deal with different formats.

From what I have seen OOXML, seems to have more to do with Microsoft wanting to retain control of the Word Processing and Spreadsheet market through it's formats, rather than use an already established and implemented standard. OOXML is clearly trying to solve two problems, neither of which are anybody's but microsoft's :
1) the Office document formats are a trainwreck (much of which is preserved in OOXML to ensure backward compatibility)
2) Full compatibility with ODF would allow any organisation to exchange documents internally and externally regardless of what software was being used.

In my experience, the only thing holding back most people from switching away from MS Office is the fear of incompatibility, not retraining or anything else - if that obstacle was removed then the productivity software market would be vibrant again. I was surprised by how many local organisations have moved to Open Office, and that they had no retraining issues and were very happy - the only problem was dealing with some arcane macros, templates and other niggling incompatibilities.

2007-02-13 06:58:28
AndyB: I will explain how FTP and HTTP coexist. FTP was created long before HTTP, and has been made largely irrelevant by HTTP, even though they are structured to perform different tasks. If we let FTP support go and retain HTTP, we move forward with less complexity and equal function.

The JPEG 2000/PNG comparison is *patently* ridiculous. JPEG 2000 is lossy, and encumbered with dangerous intellectual property and a license. PNG is lossless, and free. Generally, JPEG 2000, if you are willing to support it (apparently many are not) is a better choice for photos, whereas PNG is a better choice for archives, lines drawings and images with sharp delineations.

I am not about "choice", I am about "freedom". This particular issue is about neither - it is about complexity and subverting a standards process to perpetuate vendor lock in.

2007-02-13 11:39:39
"Also HTTP and FTP aren't really comparable - they do different things for different jobs - OOXML and ODF do exactly the same job, PDF and ODF do different jobs - PDF has always been primarily for Print and readonly distribution, ODF is general purpose."

Well, then OpenXML is for people who want maximum backward compatability with old & new Office documents. See, there is no conflict.

"as a developer of software I don't want to have to write the same thing twice but to deal with different formats."

Supporting one does not prevent you from supporting the other one. It is more work, but have you seen the specs for MPEG-4, another ISO standard? In addition to having to worry about MPEG1 and 2, you also have to worry about all of the different video & audio codecs the standard supports, for example.

As a programmer complaining about an additional format when there are 3 at most (PDF, ODF, OpenXML), talk to someone who writes a media player that supports WMV RealVideo AVS VC1 H264 AAC WMA Vorbix FLAC MP3, etc.

"This means that there will NEVER be any application that implements OOXML fully "

You mean to tell me that the open source world, with a fully specified and documented ISO standard (when complete), with all it's resources, the same group that has managed to create applications that support almost any standard out there, that manage to reverse-engineer most formats and technologies, break most DRM schemes, cannot handle one 6000 page spec?

As another example: XMLHttpRequest is not part of the HTML 4 spec, but can you imagine a new browser that doesn't support it? Sometimes you just have to support de facto standards.

Is it possible for one person to write a modern web browser? No. Will making OpenXML an ISO standard make things easier for everybody - yes. Are some people delusional in thinking that everything else but ODF will go away? ....

2007-02-13 13:55:37
AndyB: You write: You mean to tell me that the open source world, with a fully specified and documented ISO standard (when complete), with all it's resources, the same group that has managed to create applications that support almost any standard out there, that manage to reverse-engineer most formats and technologies, break most DRM schemes, cannot handle one 6000 page spec?

Of course the OS community can handle one 6000 page spec. The issue is that they clearly don't want to. People with a passion for software are not volunteering their time to write bloatware - there's no pride in it. Much of that 6000 page spec is carefully defined bloat. Some is less carefully defined bloat. It is distasteful, hence the resistance.

Rick Jelliffe
2007-02-14 01:04:37
I wish the anti-OOXML people would explain the mechanism by which an ISO OOXML would stop the progress of ISO ODF (or cause a 'barrier to trade'.) Governments who mandate ISO ODF won't, just by that, be ruling MS out of contracts. (Indeed, they run the risk of doing the opposite in the short term: reducing the running for government contracts to the comercial implementations that conform ISO ODF may rule out slower open source applications and leave the running to the commercial big boys.) Similarly, an ISO OOXML will not prevent governments from mandating ISO ODF for data interchange. Just because a standard exists doesn't mean you have to adopt it where it is not appropriate. Being pro-ODF does not require you to be anti-OOXML; what is required is for regulators to understand the technical and economic tradeoffs when regulating which standards are appropriate in different areas. ISO is the wrong forum for anti-trust disputes.
2007-02-14 12:32:39
Matt -

C'mon, the idea that IBM is trying to increase complexity to sell more services is quite a conspiracy theory. So all the people at IBM making this argument don't really believe the facts of their arguments, but are making these arguments so IBM can sell more consulting? And all the people involved in this, to a person, have kept this big secret?

I've been in consulting for over 20 years, and lots of stuff happens that isn't necessarily in the customer's interest, but trying to kill an international standard to sell consulting would take more planning, foresight and secrecy than any consulting company could ever muster.

Excuse me while I go put on my tin-foil hat!

(Forgive the snark -- I couldn't help myself.)

2007-02-14 13:11:46
"Much of that 6000 page spec is carefully defined bloat"

It is very hard to take you seriously when you make a statement like that. ODF is a smaller spec since it is missing a good amount of material - one big example being that formulas in spreadsheets is missing - which is why they are working on an entire new spec for that.

Please provide some concrete examples of bloat. Maybe the fact that OpenXML describes how to include a table in a presentation whereas ODF does not is an example of bloat to you, To someone using that feature, it is absolutely required.

So far, the only argument I keep hearing about the spec is which company it originates from, not anything else.

From http://tirania.org/blog/archive/2007/Jan-30.html

"Christian Stefan wrote me to point out that the OOXML specification published by ECMA uses 1.5 line spacing, while OASIS uses single spacing. I quote from his message:

ODF 722 pages
SVG 719
MathML 665
XForms 152 (converted from html using winword, ymmv)
XLink 36 (converted from html using winword, ymmv)
SMIL 537 (converted from html using winword, ymmv)
OpenFormula 371

Now I'm still missing some standards that would add several hundred pages and changing line spacing to 1.5 will bring me near the 6000 pages mark I guess. This is not very surprising (at least for me) since both standards try to solve very similar problems with nearly equal complexity."