Patents

As I note in a discussion of ideas, patents are often given for trivial “inventions”; and it can be well argued that patents do more harm than good. (See e.g. http://www.gnu.org/philosophy/philosophy.html^e for one of the more far-going views on the topic.) Nevertheless, the pro-patent groups have some very good arguments for their cause, including rewarding inventors, justifying R&D investments, ...

What could a good compromise be?

From my POV, the following outline seems a good starting point:

1. Patents should be awarded, but only for inventions that are truly creative. I would suggest at least two criteria that both would lead to exclusion from patentability:

   Firstly, if a highly intelligent expert in the field was given the problem which the invention proposes to solve, would he be likely to find the suggested solution, a variation or prototype thereof, or a better solution within, say, a day of thinking. Notably, many (most?) of the current patents are things that such an expert would “invent” in five minutes, often even suggest of the top of his head.

   (Note that this opens a series of questions, like “What makes a highly intelligent expert?”, that would need discussion and experimentation to be resolved in a relevant manner. Similarly, “day” may be the wrong interval, possible a week would be better.)

   Secondly, is the invention a comparatively trivial (if somewhat original) assembly or modification of noticeably more significant inventions? Alternative: Would the same or a similar invention have been likely to be made within one year by someone else, had the current inventor failed? The telephone would be a good example of such an invention.

   Note that the above will require highly competent experts to deal with patent issues. This should not be a major problem, however, in light of the many trivialities that will not be submitted at all, or can be ruled out out of hand: Most current patents would not stand a chance to be included—while the truly remarkable and important would still be protected. Further, a system of tentative patents that are easier to over-throw than grant could be used, where the burden of proof is placed largely on the inventor, should the patent be challenged. Further, there is a fair chance that the existing system of scientific publication and peer reviews can be utilized to cover a large part of the effort—unlike with the many trivial patents of today.

2. Patents would be handled centrally in one global agency, in order to cut out dead-weight and duplicated testing, to make rules more consistent, etc. Per default, a patent is valid globally; but the applicant may chose to limit the scope of the patent to a smaller area (with limited, pre-defined choices); if he does, no-one else may patent the invention in other areas.

3. Obviously, existing restrictions on patents, including that the invention is not already known to others, would still apply.

4. Patents should be expensive: A high fee, possibly linearly increasing with the number of already held patents of an organization or individual, prevents the construction of “patent portfolios”. The fee may further depend on the size of the organization, or whether an individual or organization applies for the patent. (Care must be taken to prevent artificial lowering of fees, e.g. by distributing patents on subsidiaries.)

5. A patent holder is legally required to license his invention to any and all askers for a reasonable fee, without undue discrimination. Exceptions could possibly be made where the licensing would be contrary to the personal believes of the inventor, e.g. for a patent by a pacifist to be used in the weapons industry.

   Fees are a tricky issue: On the one hand, competition should not be blocked; on the other a fair remuneration must be possible, and under no circumstance should the inventor see himself driven of the market by a larger competitor who uses his invention. Possibly the first three years should excepted; possibly very-high-but-not-entirely-unreasonable fees should be allowed, a choice between percentage of revenue and a fix sum (on the discretion of the inventor) would be possible.

   As for discrimination: Effectively everyone should have the opportunity to license at the same or, at least, very similar terms. If IBM is given good terms, then so must Sun. Exceptions would exist for organizations and individuals of close affinity (e.g. the inventors own company and his cousin). Further exceptions may be made for non-profit organizations; possibly, the equality of terms would apply only between direct competitors.

6. If it can be shown that an invention, while comparatively trivial, still took costly research to arrive at (and this was not through inefficiency), exceptions may be possible. For instance, it is often the case that biochemical companies go through a great number of dead-ends before they come up with a best-selling medicine—not to mention lengthy approval processes, research into side-effects, etc. Seeing that the eventual medicine can often be produce for almost nothing, once the research is done, additional protection is likely justifiable. (Whether as large as the companies want it to be, is another question.)

7. Patents should be strictly limited in time. The exact lengths would have to be determined later, but possibly ten years for a typical patent. (Twenty years, I believe, is a typical limit today.)

   Additional rules may be needed to handle inventions that cannot go into production until a non-trivial part of this time has passed. Medicines, as above, would be a likely candidate. Other cases could conceivably base on technological infeasibility of mass production at the time of patenting, e.g. something based on manipulations of individual atoms into a certain configuration that is currently only doable in a research laboratory; or a temporary lack of market, e.g. a new kind of memory chip that is only usable with a new generation of mother-boards. In all cases, however, this extra time must not be unduly large: Going to market in year nine, and expecting a full additional ten years of protection, should not be possible—let alone doing so fifty years after the original patent. (An exception would be if governmental regulations impose a de facto or de jure limit on the beginning of large scale commercial sales, as could be the case with medicines.)

   Other patents, notably many software patents, may warrant a shorter protection, possibly five years. (Noting the shorter “time-to-market” and the high rate of technological development.)

   Similarly, copyrights should be restricted: A fair criteria could be “until the author’s death; with an extension to fifty years after the works creation, should the author die before fifty years have passed”. (This to be contrasted with the excessive “seventy years after the author’s death” that applies in e.g. the US and the EU today.) For copyrights owned by organisations, or sold by the author, a uniform maximum of fifty years should apply, possibly with an amendment giving some minimum use, say 5-10 years, should the fifty years have expired or be close to expiration at the time when copyrights are first transferred from the author.

   By law, it should be impossible to increase the copyright/patent life over the terms at its beginning—even by new legislation. (This to counter-act problematic developments, like in the US, where copyrights stand a risk of being extended to perpetuity.) Exceptions should be reserved for extraordinary circumstance, e.g. to compensate for wars.