Slashdot Mirror
MSM Noticing That Patent Gridlock Stunts Innovation
trichard tips a column on the editorial page at that most traditional of mainstream media, the Wall Street Journal, arguing the point (obvious to this community for a decade) that the US patent system costs more than the value it delivers. The columnist is L. Gordon Crovitz and here is an excerpt: "New drugs require great specificity to earn a patent, whereas patents are often granted to broad, thus vague, innovations in software, communications, and other technologies. Ironically, the aggregate value of these technology patents is then wiped out through litigation costs. Our patent system [is] a disincentive at a time when we expect software and other technology companies to be the growth engine of the economy. Imagine how much more productive our information-driven economy would be if the patent system lived up to the intention of the Founders, by encouraging progress instead of suppressing it."
From what I understand, patents are not supposed to be granted for ideas, or methods, only for implementations.
Every idea is an "implementation" of a more general idea. There are 1. video games, then 2. puzzle video games, then 3. puzzle video games with falling blocks, then 4. puzzle video games with falling blocks that can be rotated, then 5. puzzle video games with falling blocks that can be rotated and line up x-in-a-row of the same color, then 6. puzzle video games with falling blocks that can be rotated and line up x-in-a-row of the same color to eliminate floating blocks. Nintendo has a patent on 6. So where does "idea" stop and "implementation" begin?
Businesses spend MASSIVE amounts of money either filing patents for offensive/defensive activities or trying to work around them. I would be very interested to see an accounting of a) what percentage of patents actually result in a license b) what the cost in terms of employee and lawyer time was to create the current body of patent work (plus the fees of course), c) the number of decisions to NOT make a competing product due to patent issues, d) the number of patent cases resulting in a patent being invalidated e) the number of patent cases settled out of court for less than the legal fees to challenge the patent f) (this is complex) the number of patents with BOTH cases settled out of court for less than potential legal fees and with existing patent licenses before the litigation and finally g) the ratio of licenses taken out on a patent to the number of observed workarounds (and patents on the workarounds) done by companies to avoid said patent.
If the ratio of a to b is very small, it would mean that there would have to be MASSIVE returns on license fees to justify the money paid to create patents. Otherwise we the customer are footing the bill for the horse and pony show.
c is hard to document, but every instance where it happens is one less competitor and in a capitalistic system that means less pressure to drive down prices on a product. The idea is of course to offer the patent holder a limited monopoly in exchange for publishing the idea, but the fact remains the customer loses on this deal UNLESS the invention would not have been published/implemented WITHOUT the patent system. Impossible to know I suppose, but food for thought.
Every instance of d is a waste of money in terms of all the effort to get the patent, the time of the patent office working on it, and whoever is forced to fight it. Ouch.
e needs to trigger a close examination of the patent - if the settlement is just to avoid going to court, it must mean that either the company doesn't think they'll make more in a lawsuit even if the patent is valid, the patent holder can't afford a battle either, or the defendant is not going to pay out of pocket just to invalidate a bad patent when its cheaper to settle. In the latter case, it is a waste of economic resources.
f is a possible way to get a handle on how often the first possibility for e happens - if they have successfully licensed it (not cross-licensed as part of a stand-down agreement between big players but actually had someone pay for the right to use it) and still took the lesser fees it might be at least a suggestion there could be validity in the patent.
g is simply a waste. Bright, talented minds try to work out a way around some idea, when they might be working on new features, products or inventions. Sometimes you get new ideas working around patents, but a lot of it is just monumental silliness. The consideration is avoiding the patent, not the best engineering solution.
If somehow all of these costs could be totaled up, I would be very interested to see what the end number would be.
"I object to doing things that computers can do." -- Olin Shivers, lispers.org
What is the value that it currently delivers?
"Thanks for all the money you paid to us. We've used it to buy off ISO among other things" -Microsoft
From here:
Staggering numbers.
What is humor if not pain tempered by time?
Considering the massive economic resources that have been invested by big players filing software patents, is there any politically workable way to change the law and make software unpatentable again? It would be (in the eyes of patent holders at least) the same as throwing all the money invested straight out the window. Although this may be (practically speaking) what happens anyway as far as the economy (us) are concerned via paying lawyers to fight, those benefiting from the fighting and having committed the $$ may be a hard sell. Not to mention the difficulties involved in crushing those annoying upstart competitors without having patents to wave around.
The ray of hope for real change may be (oddly enough) the patent trolls and their parasitic activities hurting EVERYONE else, but will they be enough to turn the tide?
"I object to doing things that computers can do." -- Olin Shivers, lispers.org
I agree that the current patent system is broken. However, I don't think the author of the article has all the correct solutions. For example, the following quote would seem to do away with "prior invention" roadblock to (defense against) patent creation
The proposed law in Congress would have reduced potential damages to the value of the technology, not the full value of the completed product. Another uncertainty would have been reduced by moving to the first-inventor-to-file system, instead of our more ambiguous first-to-invent standard.
As programmers and engineers we can whine about the current patent system but nothing will change until we get knowledgable people through law school and working for the patent office.
The trouble is that software blurs the distinction between a device (patent) and a work (copyright). The distinction used to be easy. If you had a new type of engine, you got a patent. If you wrote a book, you got a copyright. But software is kinda like writing a book (so it should be copyright), yet it is used to build the internals of an infinitely modifiable machine (so it should be patented).
This is going to get worse as home 3d fabrication like RepRap becomes more common. Software is now being used to build a physical object, thus eliminating the patent/copyright distinction. At a TED conference, an MIT professor talked about a fab method they apparently have going in the lab, where computation is done by arranging molecules; in theory, you could compute yourself a new car. Just imagine what that will do to the patent/copyright distinction.
The end result is that a new form of IP will have to be developed that will combine copyright and patents. In the US, this is probably going to take a constitutional amendment, which almost dooms the effort from the start.
Not a typewriter
Well said!
At the end of the day the real test of whether something should be patentable or not should be related to the reason patents were instituted in the first place...to incent investment in R&D by rewarding that investment in innovation. The reward, in the form of artificial protection from competition for a limited time, is enough to ensure the investor(s) profit from the investment. Obvious or not, if a company or individual has invested significant time/money in a program aimed at solving a problem and come up with a new and unique (even if obvious by hindsight) solution they should be rewarded not for the idea, but for the investment, thus incenting investment in innovation.
The fundamental problem with the patent system today is that it has been warped over the years into something it was not intended to be. Remember, the patent system is not something that has to exist; it is something that we as a society agree to have in order to incent individuals and companies to perform activities that are of benefit to society. Patenting of business processes, software patents and incidental patents (my own personal winner for least deserving) are all the result of this move away from the original intention. Combine this shift with the allegations of overworked and wrongly incented employees and the patent system certainly looks broken
There appear to be two basic uses for the patent system that unfortunately are sometimes at odds with each other.
[Aside: When I worked for a large s/w company we were encouraged to regularly trawl through our developed code for potentially patentable algorithms, this is clearly a case of (2) not (1)]
Surely the only useful purpose for a patent system is to incent companies to make investments that would otherwise not have been made. If a company got a clear benefit from an investment and would continue to benefit whether granted a patent or not then there is no point in society (i.e. the rest of us) granting them a patent! What they have is a trade secret that should be protected by other laws (copyright?); it should not be a patentable innovation. Other companies should have the right to make a similar investment to develop a similar solution (or license the technology/solution from the original company if that is agreeable and makes more economic sense)
Today, if a company has a trade secret that they feel they could make money off they typically have to patent the trade secret (even if only defensively) and then license it. This behaviour (licensing developed solutions) should be incented but not using the same system as that which incents investment in innovation.
So how about taking this approach...
Now, what can we say about the abstract? Well, there are (a) generalizations, and (b) there are specifics missing, without which the specification cannot be converted into a narrow set of possible implementations. Abstract data types, for example, say nothing about the language they would be written in, how they are to be implemented, or even what the actual programmatic interface will be.
Let's say we narrow some things down. We've defined implementable data types, we've defined the primary programming language and (if need be) dialect, we've defined a style (eg: procedural vs. functional vs. OO vs. 5th Generation), we've defined at least one target architecture (be that a specific JVM or a specific piece of hardware), we've defined the exposed API and we've defined some means of testing compliance to these requirements in a computable, programmatic fashion.
You now have something you can white-box test. That's close, but I don't think it goes quite far enough. Let's add one more requirement: A sufficiently large range of externally-used functions, internal APIs, data types and invariants are also defined such as to produce a high level of confidence through testing that what has been written is indeed what was designed.
THEN you have something that's as solid as, say, a car. You can always add extras to a car, so that is still "abstract" in some sense, but it's solid enough. You can test the controls within the car, and perform basic observations on things like whether the engine is running, to establish that it is indeed a car and not a pile of scrap. I would argue that software could be considered "implemented enough" once it had reached the same level of solidness and reality as a model of car from the manufacturer.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
Attempting to fix this would draw legislative fire from congress.
I've been thinking about that too.
For the sake of discussion, lets say that some event happened that forbade patenting the 'method of' anything.
What happens then? There are so many highly ambiguous patents already issued which companies paid between $10,000 and $25,000 to obtain, not counting the costs of litigation.
Would those just instantly become null and void? I doubt it unless the government bought them all back under the idea of eminent domain (as patents are considered to be property). What would be the compensation? Each one will be argued to be worth millions and a company asset. I can hear the argument now .. 'if you take our patents you seriously devalue our company'.
It would be a happy day to hear that the madness will end, but the next day would bring litigation as usual. I don't think we'll hear the end of these things until they stop issuing them and all that are issued have expired.
A lot could happen in that time.
Maybe I'm wrong, perhaps the patent office would review every single software patent and begin invalidating bogus ones.. but the lawsuits that follow would tie that up for years.
I think if anything we'll see that they can take your house to build a highway .. but they have incredible difficulties taking your patent to unplug innovation.
All in the name of progress.
I studied this "Patent" problem at length. I ran baseline study after baseline study. I compared and contrasted fully patented systems (like the US at the moment) against fully unpatented countries (Turkey and China as examples) against sparsely patented countries 2-5 year patent terms, 5-7 year copyright terms. Optimal systems come in with the sparsely patented countries. There is enough of an incentive to innovate, but also the 'shit or get off the pot' idea that being granted a right is a limited thing. You get rewards for hard work, but not forever. The American system (the system they want to use to stifle innovation everywhere else) is deeply flawed.