X
Business

War rages on over Microsoft's OOXML plans

What is it about Microsoft's proposed OOXML standard that has boffins hurling death threats at each other?
Written by Brett Winterford, Contributor

What is it about Microsoft's proposed OOXML standard that has boffins hurling death threats at each other? Brett Winterford investigates.

One would assume that taking a job at the ISO (International Standards Organisation) would be uncontroversial enough. Lots of paperwork, some good frequent flier miles, but nothing too explosive.

Try telling that to Australian programmer and standards activist Rick Jelliffe: a man who has had an unnatural amount of slandering and mudslinging directed at him.

It may have something to do with his controversial choice of roles. He recently hit the headlines after being offered to edit some Wikipedia entries on Microsoft's behalf.

These days, however, he finds himself an invited expert to the JTC1 SC34 steering committee for document description and processing languages -- responsible for arranging maintenance of Microsoft's proposed Office Open XML (OOXML) standard, possibly the most controversial in the technology industry's history.

OOXML is Microsoft's alternative take on ODF (Open Document Format), a file format approved as an ISO standard in May 2006 that describes the data held in office productivity documents such as word processing files and spreadsheets.

Microsoft proposed its own standard to take advantage of its stranglehold on productivity software via its Office application. But the proposal has been met with strong opposition -- both by competing vendors that had already put their energy behind the ODF effort, and by government and industry stakeholders from around the globe.

Microsoft's specification, officially termed DIS29500 by ISO, failed to achieve the minimum number of votes required to immediately be approved as an ISO standard during a ballot in September 2007.

Consequently, Microsoft has been given six months to address the issues raised by those national standards bodies that voted no.

Some 87 national standards bodies will meet in Geneva, Switzerland in February for a Ballot Resolution Meeting to decide whether the software giant has done enough.

The controversy
The controversy over OOXML has to date been characterised, perhaps too simplistically, as a stoush between the standards of Microsoft and IBM.

Microsoft has certainly made noises to that affect -- claiming that opposition to its proposed standard was sponsored by an IBM smear campaign.

But at a recent symposium at the Cyberspace Law and Policy centre of the University of NSW, it was clear that there are far more stakeholders involved.

On one side was a wall of Microsoft technologists and lawyers. Dotted around the rest of the room, representatives from IBM, Google and the Open Source Community, government users and developers from Australia and New Zealand, cushioned by Jelliffe and Standards Australia program manager Alastair Tegart.

"We feel that the best standards are open standards," technology industry commentator Colin Jackson, a member of the Technical Advisory committee convened by StandardsNZ to consider OOXML, said at the event. "In that respect Microsoft is to be applauded, as previously this was a secret binary format."

Microsoft's opponents suggest, among a host of other concerns, that making Open XML an ISO standard would lock the world's document future to Microsoft.

They argue that a standard should only be necessary when there is a "market requirement" for it.

IBM spokesperson Paul Robinson thus describes OOXML as a "redundant replacement for other standards".

Quoting from the ISO guide, Robinson said that a standard "is a document by a recognised body established by consensus which is aimed at achieving an optimum degree of order and aimed at the promotion of optimum community benefits".

It can be argued that rather than provide community benefit, supporting multiple standards actually comes at an economic cost to the user community.

"We do not believe OOXML meets these objectives of an international standard," Robinson said.

"OOXML is philosophically a problem," says StandardsNZ's advisor Jackson. "The world already had a standard [in ODF]."

"There has been insufficient technical rationale as to why we need another standard," added open source advocate Jeff Waugh.

But Jelliffe insists that it is unexceptional for there to be multiple standards in document processing languages, just as there are in graphics, programming languages, schema languages, even screwdrivers.

Microsoft regional technology officer Oliver Bell says a single standard is not something that he would expect to see anytime soon, especially when one considers that a third standard is already in the works. Dissatisfied with standards and standards processes that it feels are "discriminatory" and biased to the needs of the West, China has proposed its own national document standard -- the Uniform Office Format.

Microsoft also argues that without submitting Open XML to the wider industry, there would be large number of legacy documents -- the lion's share of the world's office documents have been produced using Microsoft Office -- that might not be accessible in the future. Considering the wide adoption of Microsoft Office, the vendor argues that "it is better to have a published, adopted standard than a de facto one."

Microsoft's Bell argues that as Redmond would most likely not be welcome in the technical committee that put the ODF standard together, co-chaired by IBM and Sun Microsystems, and given that there are difficulties in reproducing older Microsoft Office programs within the ODF format, there is a clear market need for this additional standard.

Open XML is 'technically flawed'?
The second main criticism of OOXML concerns the quality of the specification itself.

Microsoft's opponents are wary of the fact that DIS29500 was submitted for consideration as part of ISO's 'fast track' process -- which enables it to be approved within one year. By contrast, most standards take two to three years to go through the ISO process.

Jackson says that many national standards bodies became suspicious of Microsoft's motives after only being made aware of the fast-track request when they returned to work in early January, 2007 -- giving them little time to assess the proposed standard.

Time would be required, he suggested, especially as Microsoft's draft was an exhaustive 6,000 pages long. Few in the symposium admitted to being able to read the whole document.

IBM's Robinson calls it "rushed", StandardsNZ advisor Jackson calls it "undercooked".

"No standard is perfect, but in its current form, OOXML doesn't even meet the basic requirements of a standard," IBM's Robinson says.

However, DIS29500 has the same error rate as most submissions to ISO: one for every six pages. Jelliffe expects that a second draft could be reduced to 1500 pages should Microsoft set about culling redundant sections.

All at the symposium agreed that Microsoft is on track to address these technical issues before the Ballot Resolution Meeting.

Microsoft has promised to address these issues by February 2008. Even open source advocate Jeff Waugh conceded that all of these problems look on track to be resolved.

Waugh, who opposes the standard on other grounds, said: "None of the technical difficulties were showstoppers. "

There are, reportedly, other devils in the detail, however. Microsoft has been accused of embedding proprietary features -- in this case, binary data -- within OOXML which serve to tie users to Microsoft products, a charge the software giant denies.

Jelliffe says these concerns highlight that many in the industry have misguided ideas about the function and purpose of an ISO standard. "There is a difference between adopting a standard and having a standard," he said.

"ISO standards are voluntary," he said. "They are not laws and ISO does not make regulations. ISO provides the standards, the adopters [national standard bodies] determine whether its law."

Can Open XML and ODF be harmonised?
Google developer Lars Rasmussen says that ultimately, users would like to have a single, open XML document format with which they can exchange documents.

"One would hope that in a few years we can laugh at this situation," he said. "We will be able to open, read and edit document, regardless of the application we are using."

Rasmussen says there are often multiple standards in technology, but only when there is a "good solid technical reason".

"What I would like to see is that we work together to a single standard for exchange of documents," he said. "What I want to ask Microsoft is: what would be the downside if OOXML wasn't accepted and we worked towards harmonising with the ODF spec?"

There is some debate as to whether Open XML and ODF can be 'harmonised'.

Harmonisation, says Matthew Cruickshank, lead developer of docvert.org and formerly the lead developer on the New Zealand Government's largest XML Portal, can be loosely defined as feature compatibility consolidation of multiple standards by recognising similarities, redundancies and consensus.

"Personally, I think there is a market requirement for harmonisation," he said. Microsoft has said it cannot harmonise the two standards are due to differences in the implicit page style model of ODF versus the explicit page style model of OOXML, differences in table models and differences in the style information associated to spreadsheet cells.

Cruickshank describes these excuses as "facile".

"They don't really stand up. These are not big problems. Any technical person would understand that you can address these issues." Jelliffe again disagrees. He says "no one really knows what all the differences are" between Open XML and OpenDocument.

"Until you know that, you can't be too dogmatic about whether there can or can't be harmonisation," he said. "At the moment we can't get most of the players in the same room together. There is all this mutual suspicion and venom and heat about it. With ISO, at least we get them in the same building."

Microsoft argues that a harmonised standard would be unlikely to reproduce older Office documents in full fidelity.

Further, Microsoft's Bell argues that a process of harmonisation would be unlikely to result in an enriched version of either OOXML or ODF given the broad market adoption of both the formats already.

"I would expect the outcome to be a third standard based upon a mix of requirements from both," he said.

Tegart, representing Standards Australia, is already predicting such an outcome.

"Essentially, a standard is what people agree on," he said. "There might be a path for a third standard, which is what [both parties] agree on. We would naturally prefer to come to a resolution."

Cruickshank says users would be better served by a truly harmonised, single open standard, one which would blur Open Document and Open XML together.

"Yes, you could say that it would be introducing a new standard, but only in so far as ODF 1.2 is an update on ODF 1.1. Open XML would merge into the core of ODF and be known as ODF in the future."

The need for resolution
Jackson says it is essential that the user community -- particular in government where archiving of data is a key requirement -- is provided with a single workable open standard.

"From a government perspective, digital sustainability is essential," says Jackson. "Agencies need to know that they can they get access to documents in 100 years time." Greg Stone, regional technology officer for Microsoft in Australia/NZ, said that these messages are being registered. Microsoft, he insists, had users in mind in submitting the spec for critical analysis.

"I don't think anyone is denying that harmonisation would be a noble goal over time," says.

"We at Microsoft felt, 'let's step up to the plate and make this easier by providing as much information as possible'. We can make the spec available so the market can choose to take it up or not."

Jelliffe said he wouldn't personally recommend the use of OOXML for public government documents. "But just because I don't think it should be used, that doesn't mean I think the standard shouldn't be available," he said. "It's not OK to say I don't need this, therefore you can't have it. That's not the way the system works."

Editorial standards