Enclosed is the letter I wrote to the relevant people in the government about this. (It has been slightly edited to correct some usage).
To Whomsoever it may Concern
As a member of the Mathematics Advisory Board of the international arXiv.org e-print service, I am one of the persons quite actively involved in the wider dissemination of research in the areas of Mathematics, Physics and Theoretical Computer Science. Thus standards are important to me and I feel that I should comment on this draft policy.
Overall, I am very favourably impressed with the draft policy and I commend the relevant departments for its formulation.
However, some clauses in section 6 seem to negate the entire meaning of the exercise by providing escape routes for those who do not want to comply with the policy for one reason or the other. This will defeat the purpose of the document.
There is good reason to suppose that a standard which is maintained by a not-for-profit organisation will work well and that seems to be the reasoning behind the original draft. It is well known that the standards that drive the internet are maintained by such organisations. The purpose and meaning of "relaxation" in 6.2.1 is then just to subvert the well-known process of standardisation. Standards are never maintained by one or the other for-profit entity or one with some other motivated interest. Each such entity participates in the discussions that take place under the non-profit organisation's umbrella and a consensus is evolved to approve a standard or modify an existing one.
The term Reasonable and Non-Discriminatory (RAND) is an exceedingly ambiguous term since what is reasonable for one entity may not seem reasonable for another entity. Thus 6.2.2 opens the door for standards to become a matter of litigation --- surely something that should be avoided as far as possible. Hence RAND terms and their variants (like FRAND) should be kept out of standards.
If mandatory clauses are to be "relaxed" as suggested by 6.2.3, then there is no meaning to the term "mandatory". The standard then lapses into a guideline which is not at all what is aimed for.
It seems to me that 6.2 could be used as an excuse to make a minor enhancement of some feature already present in an existing standard and advertise it as a major cannot-do-without feature; this section would then be used as a means to go outside the agreed standard.
There is no meaning to a standard when there is more than one of them. Each office would then need to maintain conversion tools that would convert documents from one to another. Almost no such conversion will be entirely loss-less.
Let me provide an example which is within my domain. Most scientific research in the Mathematical sciences is e-published in the "(La)TeX system" --- this has been the de facto standard since the late 1980's. However, there was no uniform policy and there were many competing "ehnancements" to TeX system in the 1990's. As a result a significant portion of the arXiv software is devoted to correctly detecting and processing these legacy formats and documents. The situation has significantly improved now that LaTeX2e has become the new standard that all new documents must (must as in "mandatory") comply with.
In summary, it seems to me that some amendments have been suggested to the policy by people who:
have not done enough homework and have thus used catch-all phrases like "relaxation of mandatory requirements".
do not understand the meaning of the standardisation process which is to find a common ground that is acceptable and accessible to all. Such a process takes time and involves some give-and-take and clear disclosure of one's requirements and goals.
It is thus clear that these parties have not participated in the process so far and are trying to subvert the process at the eleventh hour. I hope that the relevant government departments will stand firm and not let such people wash all their effort down the drain.