China's national standards body CNIS has a draft document out Guide for the Implementation of the Inclusion of Patents in National Standards. (For an English translation see the first column of this.)
Wang Yiyi's China's Approach to Standards-related Intellectual Property Rights Development - Legal Framework and Implementation is a recent (October, 2009) summary of China's (PRC) position. The slide I thought was most interesting, and which I strongly support as something all other governments (we are talking about regulations here) should support, is slide 11:
This is a big deal, and very heartening to me. Royalty-bearing standards, even RAND-licensed standards, have proved themselves (notably on the Linux desktop) to be very difficult for Free and Open Source software developers to make use of. The modern computing milieu often finds one or two large commercial companies selling product, with their competition being from FOSS. A national regulation that mandates an encumbered and license-bearing technology in effect stifles the FOSS industry: for nations outside the US intending to promote domestic competitiveness and technological leadership, this should be a source of serious concern. The appropriate words in the CNIS draft seem be from s3.5.3 which (from the ANSI document mentioned below) says
In case the standardization technical committee or organization in charge fails to receive the ... (RAND or RAND-RFlicensing) ... the proposed standard will be returned to the relevant technical committee to determine whether the standard can be modified, other technical options are available, or the standard should be abandoned.
The intent seems to be that standards committees should not get involved in bargaining of the best RAND terms: if a company is not willing to RAND-license, that technology is simply out for use in a Chinese national standard. (The issue can then be shuffled upstairs to the regulators to discuss other options such as compulsory licensing, though this is out-of-scope for the standards policy.) And, importantly, it makes it the responsibility of the patent-holder to prevent submarine patents: the more clarity on these issues the better.
IANAL, but it would be surprising to me if the Microsoft-style non-assertion licensing agreements would make the bar for the Chinese requirements: they add an extra constraint over RAND-RF.
The Chinese are proposing a three-year review cycle for re-affirming standards with patents. This is very interesting, because it suggests that a standard with even RAND or RAND-RF licensed patents should be regarded as more volatile and prone to replacement than an unencumbered standard. Perhaps it suggests an eagerness to get rid of them, where an alternative comes up.
Two weeks ago the US Body ANSI issued some comments giving the US point of view. Some are innocuous, some are more substantive.
The ANSI comment (parroting their IP-loving masters) about this section is that the committee should then investigate non-RAND-RF options, as I understand it. I hope the Chinese do not take this advise. A mandatory standard should only have RAND-RF technologies: if the mandatory standard is of national strategic interest or in a dominated market, this should be compulsory RAND-RF.
One sensible comment is on s4.6, that participation in an ISO standards process (e.g., under some expectation of possibilities of some license revenue) does not create any obligation under Chinese law. However, I do think that where there is no alternative to a patented technology in a mandatory standard in order to achieve a national goal (an essential patent), a government is perfectly entitled to either fix some reasonable or even peppercorn rent for using that technology, or simply to impose RAND-RF licensing: the US itself has a weak version of this 28 USC 1498. (Furthermore, even apart from strategic mandated technologies, I think all patented technologies which dominate their markets become anti-competitive market distorters rather than the market-creators they were intended to be, and therefore should be compulsorily RAND-RF licensed.)
The comments that struck my eye include that the ANSI is against mandatory patent searches, due to expenses. I think they are wrong: a trivial patent search requires no more than a simple Google, hardly burdensome or expensive. (In any case, when a company is granted a patent, I don't see that it is too much to expect that company to maintain records: if the company has not maintained a record, that should be prima facie evidence that the company is not intending to exploit the patent and the patent should be revoked.)
Certainly I would consider that a group developing a national or international standard would be derelict in their duty if they did not perform a minimal search: in the case of the ISO Schematron standard, I certainly did, and that was 10 years ago. And I corresponded with one patent holder to clarify that we were not treading on any toes. If there are committee members whose employers don't allow patent searches, the committee can still delegate the task, for example to the editor of the standard.
What the Monopolists say
For a view of what US big business says, see the US-China Business Council comments here (PDF). The Intellectual Property Owners Association (I thought that was the US Department of Commerce?) also has comments again encouraging RAND over RAND-RF.
These seem to have much the same whinging about the enormous difficulties involved: large companies seem to have no problems accepting income from the artificial temporary monopoly of a patent, I don't see why the difficulty of patent searches should not just be a cost of doing business. Remember, no-one has a right to a patent, it is something granted by the state to encourage markets and innovation: a company that is too disorganized or stingy to play along with the state has no grounds to complain if their so-called property is RAND-RF licensed against their will.
The world gets more value from standards than it does from patents: standards law and competition law need to reflect this reality. Not that fantasy that patents necessarily bring innovation or thriving markets. Nor, indeed, that innovation would somehow halt without patents or with RAND-RF licensing.
I would go much further than the Chinese. They are taking the approach that puts the patents first: if a technology is not available RAND or RAND-RF, it must not be used in a mandatory national standard.
I would allow the national competition authority the right to declare a technology to be of national interest or market dominating, and to start off a standards effort with compulsory RAND or RAND-RF licensing applicable. And I would make it an offense for a company with patents not to be able to instantly produce records of these suitable for the use by standards bodies and the market in general: the penalty being the RAND-RF licensing of the patents. This would prevent submarine patents and increase market awareness. (Of course, there is more: a patent that has not been exploited for a certain period of time, such as seven years should lapse, even if software patents are allowed.)
I think the Chinese material should also clarify that the undertaking to RAND or RAND-RF license material is not conditional on the ownership of the patent: if the original company goes broke or is acquired, the new company cannot renege on that undertaking. (I.e., the grant needs to attach to the patent not the assignee?)
The Chinese proposals provide them with a way to prevent fraud and marketing-by-standardization: I think it would be good if it also provided a stronger carrot and stronger stick for government negotiations on strategic and market dominating technologies.