Are Standards Organizations Relevant?
Recently I was in a brief conversation about the use of the IEEE. They are an institution that, like many (all?) other standards organizations, seems to exist largely or entirely to standardize the reason for their existence. While it is generally agreed that we would greatly prefer a world with standards organizations to one without, it's evident that standards organizations need to do more than they currently are to remain relevant. Given the complaints about standards organizations, the solution to this may actually be more standardization.
Rather than simply dispose of standards organizations, or trust in existing standards organizations to take the necessary steps, some action needs to be taken to reform the entire standardization process from the inside out. Organizations creating international standards ought to be internationally regulated, with membership defined such that no single nation has the ability to take a controlling interest. To prevent issues with multiple organizations claiming jurisdiction over standardizing anything, perhaps have a single umbrella organization that creates groups as needed. The transparency that must be mandated from this organization must be second to none. Even national governments should be permitted more secrecy than this proposed standards organization; after all, national governments have national security to consider.
Another important point is careful control over what exactly becomes a standard. Standards must be fully defined with nothing critical being dependent on any particular implementation. If I pick up the document defining the standard for a document format, that document should contain all the information I require to create an application which handles that format in a manner identical to the reference implementation. For the Open Document Format (ODF) standard, this is possible. For the Microsoft Office OpenXML (MSOOXML) document format, this is impossible. The MSOOXML definition refers to the behaviour of previous versions of Microsoft Office applications without providing any documentation on how to properly produce that behaviour or even what the correct behaviour should be. Even worse, the MSOOXML definition defines an "arbitrary binary data" field — neither open nor XML, and by definition impossible to define. It is therefore possible for an application to create a MSOOXML document which would appear to be completely adherent to the document specifications but which could not be properly read by any other application.
Yet another issue that needs to be addressed is how standards can differ from region to region. A prime example is the IEEE 802.11 standard, which is an umbrella standard for wireless networking technology. 802.11a, 80211b and 802.11g are the most well-known standards, and 802.11n is already starting to be integrated into all new wireless equipment. However, the 802.11 standards do not function the same way in North America as in Europe, which is a problem. A standard must work seamlessly in all regions, so that users from one area of the world can go to other areas and have confidence that their standards-based hardware and software will function the same way everywhere. With 802.11, not all channels are available to you in all parts of the world, so your wireless networking hardware from North America may not be legal in Europe.
Finally, patents are a huge issue, not only in standards but in hardware and software development in general. There is no easy way around this; either companies must freely and automatically provide free, irrevocable, worldwide royalty-free licenses for any patent any developer reasonably needs to implement any standard, or companies must be forbidden from including any patent-encumbered technology in any standard. The former seems to me to be the better idea. A prime example of why these assurances are necessary is to help prevent objections similar to the objections around the use of Mono, which provides the .NET runtime for non-Windows users, as a development platform for Linux. Microsoft owns all the patents related to .NET and so, in theory, could shut down Mono at any time. Although they have not done so yet and have stated they do not intend to do so, in the absence of any formal legal protection from patents this possibility is something that must be seriously considered when deciding on a development platform.
Many people disagree however, and think that things are perfectly fine just as they are now. These people have nothing to worry about; my ideas would require many different people from many different nations to hold open and transparent negotiations, come to a mutually-agreeable conclusion that benefits everyone and radically change the status quo. None of these things seem to be high on any politician's priority list, so things are most likely going to remain exactly the way they are now.
