Software patents controversy
Posted by davidmcnab 9 years, 9 months ago to Technology
A discussion question about intellectual property - what constraints or qualifications, if any, should be imposed on the patentability of ideas and concepts?
In other threads, people have noticed that I get a bit fired up on the topic of patents.
I'm a software developer, and invention is what I do for 8-10 hours a day. I look at code, data and design notations on my screens and think "How am I going to do THIS? How am I going to do THAT?" "Ok, yes, this should work"...
The USPTO has for years been granting patents for trivial, obvious, mundane and even non-original software programming practices. Every day in my work, just by performing my assigned duties and engaging my intellect, I am breaching hundreds of software patents that I have never even heard of.
It's equivalent to a chef being unable to cook without infringing, because there are patents on ways of cutting carrots with a knife, and patents on methods for sauteeing onions, methods of determining if a piece of fruit is fresh by smelling it and so on. Or if this regime applied to music, a guitarist could patent certain chord progressions, even ones others have played before him. Or a songwriter could patent trivial musical devices like rallentendi (slowing the beat in places for emotional emphasis).
There are literally hundreds of thousands of software patents registered. For a software developer to know every possible patent and its scope is beyond human capability (except possibly in certain cases of profound autism or photographic memory). To search through the patent database when designing and writing code, and amending one's practices to sidestep all patents, would grind nearly all software developers to a complete halt. To submit all my code for review to patent holding firms such as Intellectual Ventures, receive a report of infringements, and pay license fees to cover these, would drive my company (and virtually all companies which develop software) out of business.
Parallel to this, even the most trivial, obvious and non-original of business methods can be patented. Even the practice of a physiotherapist phoning or texting their client 2 days after a consultation to remind them to do their exercises can be patented. Also, the USPTO has switched to a 'first to file' regime, so that 'prior art' can NOT be used to defeat patents on unoriginal inventions.
So how far should we go with patentability? What qualifies an idea as worthy of granting a decades-long legal monopoly? I recognise that we must have incentives for innovation, but the current patents regime has been starting to exert the exact opposite effect.
In other threads, people have noticed that I get a bit fired up on the topic of patents.
I'm a software developer, and invention is what I do for 8-10 hours a day. I look at code, data and design notations on my screens and think "How am I going to do THIS? How am I going to do THAT?" "Ok, yes, this should work"...
The USPTO has for years been granting patents for trivial, obvious, mundane and even non-original software programming practices. Every day in my work, just by performing my assigned duties and engaging my intellect, I am breaching hundreds of software patents that I have never even heard of.
It's equivalent to a chef being unable to cook without infringing, because there are patents on ways of cutting carrots with a knife, and patents on methods for sauteeing onions, methods of determining if a piece of fruit is fresh by smelling it and so on. Or if this regime applied to music, a guitarist could patent certain chord progressions, even ones others have played before him. Or a songwriter could patent trivial musical devices like rallentendi (slowing the beat in places for emotional emphasis).
There are literally hundreds of thousands of software patents registered. For a software developer to know every possible patent and its scope is beyond human capability (except possibly in certain cases of profound autism or photographic memory). To search through the patent database when designing and writing code, and amending one's practices to sidestep all patents, would grind nearly all software developers to a complete halt. To submit all my code for review to patent holding firms such as Intellectual Ventures, receive a report of infringements, and pay license fees to cover these, would drive my company (and virtually all companies which develop software) out of business.
Parallel to this, even the most trivial, obvious and non-original of business methods can be patented. Even the practice of a physiotherapist phoning or texting their client 2 days after a consultation to remind them to do their exercises can be patented. Also, the USPTO has switched to a 'first to file' regime, so that 'prior art' can NOT be used to defeat patents on unoriginal inventions.
So how far should we go with patentability? What qualifies an idea as worthy of granting a decades-long legal monopoly? I recognise that we must have incentives for innovation, but the current patents regime has been starting to exert the exact opposite effect.
Thanks for your reply. I can totally understand your antipathy and incredulity, given that the software patents controversy has received very little if any air time in the mainstream media.
To get started, here's a tiny list of some trivial software patents, by no means exhaustive, but valuable for example sake:
http://en.swpat.org/wiki/Example_softwar...
Here is a paper on the phenomenon of making trivial software inventions look non-trivial for the purpose of patenting, together with evidence:
http://www.ccp14.ac.uk/maths/software-pa...
Evidence regarding 'first to file' doctrine as signed into law by Barack Obama:
http://www.ipwatchdog.com/2013/03/16/a-b...
A ludicrous example of a trivial 'invention' which was granted a patent:
http://www.google.com/patents/US6368227
More information on triviality of software patents:
http://www.ffii.org/Why%20software%20pat...
Further information:
https://www.gnu.org/philosophy/trivial-p...
And overwhelming amounts of further information available on the academic databases and mainstream search engines.
And in answer to your question - I have read Atlas Shrugged 4 times over the years, The Fountainhead 3 times, and seen We The Living once. And yes, I am an admirer of Ayn Rand.
Any questions?
His new book will be coming out in the Spring, "The Source of Economic Growth." there are many posts on this site which take on your arguments. you can search for them.
In software, I support copyrights. I support patents for any truly non-obvious inventions. Given that software changes so quickly, however, I do question the appropriateness of the usual circa 20 year timespan, and suggest that software as a discipline would advance far more quickly with a much shorter term for the more trivial patents, even in some cases as little as 5 years. In software, 5 years is like 20 years in more tangible disciplines anyway. If patent timings and thresholds were better tuned, it would have a net positive effect on the industry.
I for one would love to sweat for months over a difficult software algorithm then license it out for a lucrative income stream. But as an individual, I am disincentivised against this, given that I don't have the financial resources to go after infringers in the courts.
There is also the pragmatic aspect of international competition. Part of China's massive surge into economic supremacy (now greater GDP than US economy) is that they pay only the barest lip-service to intellectual property, and only honour it if their trade routes are pressured. What if software patents, especially in their present state, are steering us into a future of economic serfdom under China?
With regard to the 'first to file' issue - this goes to patent trolling - the practice of threatening businesses with inherently invalid patents, knowing it will cost the business far less to pay a license fee than to pay the legal costs to get the patent rightfully overthrown in the courts.
Copyrights are very narrow and protect the artistic not an invention. Finally, look at the empirical evidence. Which countries have thriving software industries and look which countries have the strongest patent systems? The macro economic evidence does not support your point of view. I don 't even understand your conclusionary remarks about China. The situation with China snd respecting IP is getting better than it was previously. It 's still a problem. Finally, life of a software patent. If things changr quickly patents become obsolete. If its important tech it 's important 20 years later.
The software industry is plagued with tens of thousands of trivial patents for simple, obvious common techniques, the kind that thousands of people would easily think up for themselves independently, and do. The exact opposite of Rearden Metal in AS, an invention which nobody thought of in millennia of metallurgical practice.
Why software is subject to this kind of abusive practice is its abstract nature. You can't see it. You can't touch it. There is no single way of objectively visualising it. A simple piece of logical sequence can be rewritten to something far more complex, dressed up for patentability, then used to attack the thousands of other developers who independently "discover" the same logical sequence, or something completely isomorphic to it.
In industries such as mechanical or chemical engineering, it is practicable to perform patent searches for discoveries or inventions, because these professions have a much lower 'density of invention'. However, software development is virtually *all* invention. Yes, there are certain algorithms that are standard, like searching, linking, sorting and so forth. But the bewildering diversity of problem spaces means that 99% of a software developer's time is spent inventing.
My point remains - software development is vulnerable to the morally questionable business practice of dressing up very obvious, basic techniques and presenting them as complex original works for patenting. Several jurisdictions around the world have recognised this, and have specifically legislated exemptions such that pure software is not patentable itself, unless it is part of a larger system including both software and hardware. The overwhelming consensus among individual software engineers is that patents pollute their work with ridiculous toll gates at every turn.
If you don't believe me, I strongly suggest you talk to other software developers.
In my younger plant and project engineering days, I was faced with replacing a 40 some year old mechanically driven and controlled adjustable hydraulic valve used to control and limit several tons in free fall or drop for annealing aluminum aircraft metals. The original valve had originated on early 1940's aircraft carrier elevators and no one could make or replace parts. I found an integrated op amp-servo valve that could be coupled to valving large enough to handle the load, but there didn't exist at that time, any off the shelf controllers. So I took an Intel 8088 board and a lot of input/output devices and built the system and programmed, with a lot of trial and error, the eprom utilizing the code supplied with the 8088. I was able to effect the modernization and since I had incorporated programability into the controls, also permit increased efficiency in the operation as well as safer structural shock loading saving a 40 year old production installation.
As best I could determine, the application, integration of components, and programming were unique and original at the time, but I didn't consider it patentable since every component existed and the programming was enabled by Intel and it's existing code. However, using that example, I've found many in the patent arena that argue that the system as a whole, the application, and the software were absolutely patentable. To be honest, I didn't see it as such at the time, and I still don't. Had I developed the code that enabled the programming, I would have considered that patentable, but utilizing logic and speed inherent in the code, just waiting for me to discover and apply it didn't qualify.
I had a client once tell me we can't measure the resistance of a heater to estimate its temperature b/c that's patented. You have to have a separate RTD thermally coupled to the heater. That doesn't ring true. IANAL but I always imagined if they went a head an measured heater resistance, no one would be able to sue them successfully.
I don't think you get what is measured when you determine a resistance. The reason you use thermocouple devices separate from measurements of the voltage and current through a resistance heater element is a matter of determining actual vs. estimated heat output. As far as I know, it has nothing to do with a patent. The voltage and current through the resistance element are still measured or diode (or other current reacting device) limited for safety purposes.
The LASER was a ruby rod laser, a flash lamp, and an etalon... These all existed before.
Your line of investigation whether something is an invention and therefore patentable is nonsensical. You have to start with reality and inventors cannot violate the laws of physics.