Slashdot Mirror
The Good Old Patent Law - Revisited
trifakir writes "Scientific American talks about the imperfections of the current patent law, subject to the book of two authors from the Harvard Business School. It seems that even business people start seeing the insanity of the current patent system. This time it seems that they are not only criticizing, but suggesting some procedural amendments (e.g. patent conflicts resolved by a judge and not by a jury). Do you think that any of these has chances being heard by the big wigs?"
...the patent office makes more money allowing retarded patents to get through than by being sensible.
It seems that even business people start seeing the insanity of the current patent system.
As long as 'business people' in the form of very large companies are trying to get something similar to US patent law into European Union law, I won't believe in a change of opinion at the top. Everyone knows that the US system is broken, but the odds are still on it being adopted in the EU.
Mielipiteet omiani - Opinions personal, facts suspect.
The imperfections may be subject of those two books, but I really doubt that they are subject to those two books.
"The significant problems we face cannot be solved by the same level of thinking that created them."
[ I can not bring myself to believe that if knowledge presents danger, the solution is ignorance ] -- Isaac Asimov
Instead of fixing the patent system, the authors seem to suggest that a faster-track court system will fix the problem. However, those who an invalid patent will continue to bare the entire financial and legal burden to do so.
Thought experiment: I'd like one of the authors to discredit any flawed patent using their own methods. Will it be very expensive? My bet is "yes".
Will this stop the increasing flow of bogus patents? No.
Nothing changes unless those with the power to change it are effected. It works this way at work, in business, government...
So due to the patenting policy of all the big companies, no new idea is rewarding for his inventor in the field of software patents. Because a software implies so many ideas, it is subject ot a lot of patents, that is the main difference with the other industry fields.
You should have a look at Richard Stallman's talk about patents, it is far more informative than this article, and also the presented books.
Patents are killing inovation, and let big companies use every new idea thanks to cross-patent licences...
Have I read a patent? I have worked in R&D organizations whose goal was often to develop patentable technologies. I have read thousands of patents. I have also written about 18, and am listed as an inventor on 12 US patents, as well as the foreign equivalents in dozens of other countries.
As far as reverse engineering goes, there are plenty of cases where that doesn't work because of the complexity of modern biological systems, or the technology in question is a processing step needed to produce something very difficult to measure. And don't forget that reverse engineering can be barred through legal means, i.e. a license. And of course because of patents we don't have many companies trying to conceal their technologies to any great extent. Make patents difficult or impossible to obtain and you will get all sorts of creative (and ultimately damaging to the spread of new concepts - and the overall economy) efforts to conceal, obfuscate, and legally hinder any efforts to discover trade secrets. The possibilities are endless, and include stuff we haven't seen for centuries.
Before making a large scale attempt to destroy a major institution that has been in place for centuries, you had better be damn sure that you FULLY understand what forces caused it be put into place, and what the consequences of removing it will be.
Much (not all) of the present problem with US Patents derives from changes made in the 1980's.
At that time, a typical patent application might sit in multiple bureaucratic queues for years, sometimes decades, as each patent had to be rigorously analyzed and proved to be substantive and unique, new tech, etc. Basically all the work was at the front end of the process. This collided with the rapidly accelerating pace of new tech and new product lifetimes - many products were introduced, sold and finally end-of-life'd before the patent was awarded. Getting a patent was expensive, slow, and difficult.
The Patent Office was basically told by business, Congress and the Reagan administration to Do Something, without spending more money. USPTO quit being so rigorous, and just began to award anything that seemed reasonable, i.e., everything except perpetual motion. The analysis was pushed to the backend, putting the onus of proving the patent upon the patent holder and the courts. This removed a multi-year logjam at the USPTO, without requiring the hiring of hundreds or thousands of expensive new patent examiners with expertise in all the new tech fields. The business community was much happier with this approach, perhaps shortsightedly.
So, since 1980's a patent is more like a timestamp than a proof of innovation. It's a piece of paper that gives you the right to defend your invention. This is arguably a reasonable way to go, in principle. It can mean that the system only invests the cost of rigorous defense on patents that have commensurate value. Individuals and small players can get their timestamp at minimal cost, and then begin trying to make money with it. As we know, there are problems with the new system, and perhaps the proposed changes will fix them.
That period is also when they decided, erroneously IMHO, to accept patents for software. Previously software was considered to be algorithms, which are math, which is "discovered" natural law, not invention. The first awarded patent involving SW as such, IIRC, was a Honeywell patent for a HVAC (or security?) control system that included software in the controller's CPU. Acceptance of SW patents was partly driven by the successful patenting of hardware implementations of algorithms that could also be done in SW. Challenges to the USPTO policy were based on that disconnect. IIRC the original virtual memory patent, was awarded about 1962 for a hardware implementation by a British computer company whose name escapes me.
Unfortunately, opening SW patents at that late date was destructive, because two generations of software builders had worked under a paradigm that depended on either "trade secret" control, or public sharing. This allowed companies to protect their employees' work for a time, but maximized the velocity of innovation by rewarding innovation with respect and increased employability. Patents broke with a long history and tradition. The effect has been to reward minor tweaks while failing to reward those who arguably contributed the most seminal work. While my own work can by no means be considered major, a lot of my work in systems architecture and image processing would now be patentable.
Sharing of information greatly increases its utility, a la the "network effect", and the velocity of software innovation is so high that patents are a poor solution. A SW patent with a term of perhaps 2 years from the date of patent, plus the period between application and award of patent, might be useful. Certainly the mere transformation of a process from hardware to software, or from local to networked systems, should not be patentable, any more than building something in plastic rather than metal should be.
It's easier to be a result of the past, but more fun to be a cause of the future! http://www.spacefinancegroup.com/
We need to move more towards a system that emphasizes testing the obviousness more than searching for prior art. Prior art searches are expensive and can never be even close to exhaustive, and so will always have huge holes.
How can that be done? The current system makes obviousness impossible to test for, because once the patent is published it looks obvious after the fact.
So in order to test the obviousness, a short summary of what the invention does, without any details of how it does it, should be published while keeping the rest of the patent hidden. If within some predefined time period (say 90 days) somebody can come up with a detailed description of how to do the same thing, or actually implement something that does it, and that implementation or description is similar to the methods described in the patent application, it should be regarded as obvious and the patent not awarded. That somebody else could create it in such a short time, and is willing to do so knowing they won't get a patent for it, is a strong indicator that the alleged invention is simple enough that it does not need the incentive of a patent in order for it to be created.
Of course, some will object saying that for some inventions, the problem or goal is in itself nonobvious and innovative, even if the implementation is obvious. Well, I say even those still don't deserve a patent. Those kinds of inventions do not require the incentive of a patent, because the first person to think of it or someone else would almost certainly create it anyway if the implementation is obvious. If someone else encounters the same problem, they will solve it in an obvious way. If no one else encounters the problem, there isn't much harm done in not having the patent granted.
---------
There is inferior bacteria on the interior of your posterior.
A better measure of performance than number of patents is the percentage of patents which withstood challenges in court during their lifetime.
The number of patents granted, large or small, tells you little about the quality of those patents.
What?
" If we permit the patent office keep their fees, then they have more incentive to pass more bogus patents."
No, you have it backwards. They have more incentive (as an agency) to reject patents if they keep the fees. The fees are paid for the *application*, rejected or not. If they reject patent applications, the applicant might refile (so they get paid again). If they accept patent applications, then no more fees.
The problem is that 4 hours (assuming two applications to process per day; I believe that is what I read in one of these discussions) is not enough time to find the prior art and evaluate the obviousness properly. Thus, the patent office needs more time (money) to function properly.
Its also worth noting that government agencies tend not to give raises based on performance. Instead, they give cost of living adjustments or promotions. Thus, your scenario is a bit far fetched even if you had gotten the incentives correct. Increased funding would tend to result in hiring more people (rather than giving raises), as supervising more people is a promotion (which gives the manager a raise).
It would also help if the patent office reviewed every patent when a claim was made. If IBM (or whomever) had to submit its patent infringement claims to the patent office who would then forward them to the defending company, it would be easier for the company to contest the patents and harder for IBM to engage in extortion ("Ok, maybe these patents are no good. What about our other patents? Are you sure that you can beat all of them?"). The process would go something like this:
1. Someone gets a patent on whatever.
2. They find someone else who they believe use technology that infringes on their patent and inform the patent office.
3. The patent office sends an infringement notice and asks if the defendant wishes to contest the patent (or its relevance to the defending company).
4. If the defending company contests, there is a discovery period where each side sends the other (and the patent office) its claims. If one side concedes during this, the review of the patent is called off. If neither does...
5. The patent office schedules an examination to review the claims. At that time, they can choose to uphold or cancel the patent. The losing side pays the examination fees.
This process moves the burden of finding relevant prior art to people who are actually *in* the business concerned. The patent office merely rules based on their own definitions.
That said, I support what they're doing. I don't think that ideas can actually be owned. (*** ducks ***)
Why are you acting like you're saying something heretical?
Ideas themselves cannot be owned.
What can be owned is a temporary, government-granted monopoly on certain uses of that idea. These are called patents and copyrights.
- Neil Wehneman
My legal education, in nifty podcast format
I actually made some comments about intellectual property rights in my blog actually. In a nutshell I am thinking that it should be removed, because having it removed benefits more people than a select few in the big picture.
Archie - CIO-for-hire