I am publishing this post almost three months later than my initial plan, mainly because the past three months were very complicated ones.

There are many “open standards” definitions, and this great amount of interpretations or views about the subject, allowed distortions to make possible that almost anything on IT market could be classified as an “Open Standard” by marketing departments.

For me, and certainly for many of you, an open standard must meet two basic requirements:

1. Have been developed and is maintained through an open and transparent process.

2. Have open licensing.

The first point is self-explanatory (just a visit the committee that develops the ODF, a look at our mail archives and you’ll understand better what I am talking about). The second point is much more complicated, and therefore decided write this post.

Basically there are 4 types of “licensing models” (I do not know if “licensing models” is the best term to mane it, but for lack of another term, I’ll use this one) that are used in the development of standards (in fact, there are 3 models and a huge distortion that we may unfortunately consider as a fourth model).

The first and simplest of them is the so-called royalty free. Through this model, all the intellectual property rights related to a standard is available for licensing without paying a fee (free as free beer). This model allows that a person with a good amount of talent, an open standard on hands and some code development work will result in natively interoperable products (and services) without no additional entry barriers to entrepreneurs. ODF is licensed through this model.

The second one, is the so-called RAND (Ressonable And non-discriminatory). One don’t need to “force the brain” to realize that this model is quite complicated to use in practice, because what is reasonable to me may not be for you and to believe that something can be non-discriminatory on our current world is a little bit difficult task, OK ? (if someone disagree with that, please explain me how a Cuban company may RAND license technology from an USA company ?). ISO accepts RAND licensing and OpenXML can be cited as an example of this type of licensing (indeed, if there is anyone encouraged to answer the previous question, please improve the response explaining me how Cuba - an ISO member - can get a RAND license to develop an OpenXML based software… ISO is really a complicated institution.). In other words, for the USA economic reality, a technology license that costs U.S. $ 500 K may be reasonable, but I can’t assure that this resonable in Brazil (same on other countries)… a value like that would be considered “reasonable”? Maybe the participation on international standardization process to latter offer RAND licensing of its IPR may be a good deal.

The third model is known as ex-ante (from the Latin “before the event”) and basically says that all patents and intellectual property rights involved in the development of a specific standard are declared before the beginning of the standard development process (or before a particular company that owns the IPR begins to submitt its contributing to the standard development). For those who played soccer (or football) in the street as a kid, this is the famous “ball owner”, the one that may define the game’s rule and the other participants has only two options: either accept or forget soccer (or football) on that day. It is also important to emphasize that the IPR licensing mentioned can be done in a reasonable and non-discriminatory terms, but the important point here is that you now know exactly where are you putting your feet (I recommend everyone to read the licensing document of OpenXML to understand this difference. On that document, you will not find a list of the patents or rights involved, but only a promise not to sue … and promises is probably the easiest thing to find in Hell ).

The fourth model, as already stated isn’t exaclty a model, but a “good faith’s rape” (excuse the harshness of the words, but I cannot express what I feel in more light fashion). It is the use of “submarine patents”, which are nothing more than patents registered by a company (usually a patent poll, an association of companies that normally would not like to see their names directly involved in such delicate issues) that end up being used , or in some cases just “touched” by any standard. These companies are not always aware of the use of its IPR in a standard during its development (OK, a simple accident), but what we have seen lately is that patent polls silently tracks the development of standards. When the standard is ready and already in use by a significant number of businesses (and people), the patent poll gets out from the shadows and presents their account, retroactively charging by the use of their IPR (fair and ethical, OK?). These patents are called submarine patents, because when least expected, they emerge.

As you can see, an specification called as an “open standard” can be licensed by one of the three models mentioned above or be the victim of the fourth type and I write this text to warn everyone about this problem.

Unfortunately in Brazil there isn’t the culture of studying IPR issues in our IT colleges and universities and we have many people out there in good faith that is under a huge risk to be sued a millionaire at any time, simply because they did’nt realized that this simple detail exists (OK, in Brazil it isn’t possible do register a software patent, but today I really thionk that Brazilian companies don’t really want to develop software exclusively for Brazil.). If someone want to have a better idea of what I am talking about, do some searches on Google Patent Search, and will be amazed with the things they can find it (hint: look for some data sorting algorithm, for example).

I await your comments, questions and views about this post and please never forget: The devil lives in the details