A few weeks ago Morten Andreas Meyer the Minister of Modernization in Norway gave a speech about the Norwegian government’s desire to only use products which support open standards. Specifically he talked about the desire to implement open file formats. Norway , of course, is also the proud home of Opera Software. Paul Festa at news.com covered the minor war of words that resulted when Asa Dotzler, who is community lead at the Mozilla Foundation took the minister’s speech as an opportunity to be a little provocative and to ask what this move by the Norwegian government would mean for the company.

I’m not interested in commenting on the exchange between Asa and Opera. However, this debate did raise an interesting issue about the confusion between open source and open standards. A lot of energy seems to go into drawing an equivalency between the two. This is strange given the reality that the supposed equivalency does not in fact exist.

The Norwegian government’s has an admirable desire to ensure that they are using document formats based on open standards technologies. This was the main thrust of Minister Meyer’s speech. In addition he said the government should also investigate the role that open source should play in the government’s IT ecosystem. Again, there is nothing wrong with that. I know our team in Norway has been and will continue to be engaged with the Norwegian government to explain how Microsoft already provides a choice of open xml based document formats for it Office products. In fact they are now able to point out that this support gets even better in the next version of Office when we make XML the default document format for our Office products. In addition the open standard VC-1 codec for media streaming should address the Minister’s concerns about using open standards to stream his speeches in the future.

The problems begin when an equivalency is drawn between open standards and open source development. The fact is that open source development has no inherent advantage over commercial software companies in the implementation of open standards. In certain situations the reverse may actually be true. Why is this the case.

An open standard has no relevance while it remains only a specification on paper. Real open standards need to be implemented in software. Further more, just because you write some code which implements the standard means very little in reality. Most open standards are highly complex specifications with lots of optional components and implementation dependant interface choices. Implementation of an Open Standard means nothing till the system has been integration tested with other systems and proven to be able to connect to the wide diversity of other implementations of the open standard on other platforms. All of this takes serious time and serious resources. Unfortunately it’s highly unlikely that most open source development projects will have the wherewithal to prove the correctness of their implementation.

Even very well funded open source projects, for the sake of argument let pick Mozilla, don’t get it right. Just send your copy of Firefox over to the Acid2 test…oops! Internet Explorer's poor CSS support is well known. I am really hoping our IE team takes up Hakon’s challenge to fully support the CSS standard. This recent update from the IE7 team does seem to be pointing in the right direction.  In the first draft of this piece which I wrote on a flight from Zurich to Seattle today, I said that I expected Opera would ace the Acid2 test. I was wrong! I tested the latest version of Opera and it seems not to render the test correctly either. I guess that just makes my point about how hard it is to do open standards implementations.

The reality is that to do open standards well requires significant resources and great engineering design and implementation discipline. It would be tough to find a company that has done more to implement XML in the real world than Microsoft. The breadth and depth of XML implementation across the Windows platform is pretty impressive. We already implement the next version of the IP v6 open standard in Windows. The next version of windows takes this to a whole new level.

The inherent problem for government policy makers is that they are being told that a policy mandating open source is a prerequisite for advancing their desire to base their systems on open standards and open file formats. Even Bob Sutor, IBM head of standards, who has been occasionally known to connect the two concepts, admits that this is a problem.

The two issues of open source and open standards are wholly independent. The process of evaluating open standards and open formats should be separate from the process of evaluating whether open source software may have a role to play in government systems.