So much quarreling about open standards. Jason Matusow advocates for a document format with RAND licensing conditions for the patents. What does he mean when he talks about RAND? RAND stands for "reasonable and non-discriminatory". But Jason Matusow's company Microsoft lacks honesty when it talks about "reasonable and non-discriminatory" conditions.
We need to be precise about what reasonable and non-discriminatory actually means. A restaurant in apartheid South Africa said it allowed both Boers and English, so was "not discriminatory". It even let some Jews in. However it banned non-whites.
Reasonable and non-discriminatory in patent licensing means "we apply a uniform fee". However with respect to Microsoft's legacy OOXML format, one party controls the standard and the associated patents. All market players need to license except the patent owner. For dominant standards it is a tax on the market. It seems highly unreasonable that such standards should become international standards, mandatory for government users.
You may find it unreasonable for an ubiquitous standard. But there is a more insidious aspect. RAND patent licensing conditions are a tool to ban Free Software, which is entirely incompatible with RAND licensing conditions. Now one side of the debate blames it on the patent licensing conditions, the other side on the software licensing conditions.
The reason I agree with the statement about patents and Free Software not mixing is that there have been terms written into GPL licenses that explicitly conflict with software patents. Okay, that is the choice of the authors and users of those licenses.
It sounds a bit like: well, you chose to marry an African woman, so we cannot let you into the restaurant. Free choice, right?
Yes, Matusow calls his standards with RAND conditions "open standards" and contradicts the commonly accepted definition of "open standards". We should speak about shared standards. These shared standards appear to discriminate less, but they still discriminate against the only real competitor to Microsoft's hegemony.
It is true that ISO, driven by simple pragmatism, allows shared standards. From the ISO/IEC directives:
14.1 If, in exceptional situations, technical reasons justify such a step, there is no objection in principle to preparing an International Standard in terms which include the use of items covered by patent rights – defined as patents, utility models and other statutory rights based on inventions, including any published applications for any of the foregoing – even if the terms of the standard are such that there are no alternative means of compliance.
Generally international standards and patents are like water and oil, and RAND conditions are the soap that allow them to mix. But as the move towards Open Standards evolves, shared standards get more and more unacceptable. Shared standards do discriminate and do appear to be unreasonable.
It is time to adapt the legal definition of reasonable and non-discriminatory to common sense.


