Here’s why economists hate software patents
washingtonpost.com
washingtonpost.com
For most software patents, the capital cost is almost nothing. Even if it took a programmer a whole year to invent something, that's like 100k--no money at all in the larger sense. So it makes no sense to protect a capital investment that's negligible.
That said, the article really fails to explain why lawyers don't want to have a categorical exclusion for software patents. Usually, you want a legs principle to get to the root of why something is bad, not address bad situations ad hoc. Why are software patents bad? Sequential innovation, as mentioned in the article? If so, that should be the basis for the exclusion, not whether something is software or not.
As a class, lawyers love systems in which bad situations are addressed ad hoc. All of their major incentives are aligned that way:
+ They get to tell their clients what the clients want to hear: "The law doesn't look good for what you want to do, but we've come up with what we think is a powerful argument that your particular situation should be treated differently." That's good for getting hired, and for staying hired by being perceived as a "team player."
+ They can earn the admiration of their peers for successfully persuading a court that the client's situation should indeed be treated differently than the law would otherwise require. That enhances reputation, which helps generate future business.
+ And of course let's not forget the direct economic incentives associated with ad hoc proceedings: Lawyers get to bill for factual investigation; legal research; opinion-writing; seeking discovery from the other side; responding to the other side's discovery requests; filing or responding to pre-trial motions; trial prep; the actual trial; and the inevitable appellate proceedings.
In the same way, lawyers can recognize that special cases make the system worse, while still pursuing special cases for personal benefit.
This is eminently true -- but many, many lawyers would argue fiercely that special cases make the system better: more flexible, more accommodating of the needs of real people, etc. It's hard to say how much of that is due to incentives affecting perception. To quote Upton Sinclair, "It is difficult to get a man to understand something, when his salary depends upon his not understanding it!" [1]
Could it be as simple as software patents are good for lawyers and bad for almost everyone else?
Is infringement contingent upon hardware implementation? Partial hardware implementation? Substantial hardware implementation? If you have an awesome DSP and I virtualize it and then use hardware acceleration to speed up that virtualization, is that a hardware implementation, or software?
Saying "no software patents" as a practical matter is hard to nail down objectively, without just generalizing it to "no patents on digital anything". Which becomes a larger upheaval than people tend to consider.
But, who are we kidding anyway. Software patents cannot e gotten rid off because of the entrenched patent holders. If you let them keep those patents they have an unfair market advantage over new entrants, and if you retroactively cancel all software patents (if you can define such a category) then they'll sue the government's pants off and through the court system have software patents declared legal. The onything i could imagine is gradually shortening the lifetime of software patents to decrease their harm and attractiveness.
The reason many patents, not all and not necessarily the majority, are harmful is because there is no clear definition of what a patent is allowed to be. The definition of hardware patents that was intuitively accepted by patent clerics did not evolve properly to accommodate software. This is by no means a limitation on patents but by the bodies that grant them.
It also isn't a matter of the bar being low for patents applicants. The weakness in the system is that until a clearly worded definition of what a software patent constitutes, we can't pass a judgement on whether patents are useful or harmful.
Software patents are good for lawyers. Software patents are good for monopolists who want to control a market without having to compete for it.
Software patents are not good for software developers, who are frequently prohibited from reading patents by their employers, rarely read patents even when allowed to, in any case don't speak the legal language in which they are written.
Software patents are not good for startups, who have no way of discovering whether they are in violation of patents short of going to market and seeing whether they are sued; a lawsuit can easily bankrupt a startup regardless of the merits.
Software patents are not good for large companies, who have to waste enormous amounts of their employees' time making up bullshit patents just so they can have an arsenal of defensive patents with which to ward off other large companies.
Software patents are not good for open source, which neither needs nor wants the temporary monopoly that is the point of getting a patent in the first place.
Software patents are not good for software users, who suffer from reduced marketplace competition, increased vendor lock-in, and a more sluggish pace of progress overall due to the opportunity cost of the time wasted on patenting things and dealing with bullshit patent lawsuits.
Software patents are clearly harmful.
The problem is not the list that you mentioned above. In fact I can refute every one of them. The problem is that there hasn't been a concerted effort to define a software patent. Hardware patents are easier to define. They are that thing you build according to a patent claim. Software is neither a material thing nor a concept. It actually runs in a computer and more or less controls it. It is an invention. But the boundaries of the claim are difficult to discern. Most of that is because software patents are a few decades old whereas hardware patents are more than a hundred years old. Hardware patents have run through their paces and "debugged" so to speak of glitches in the system. Hell, all software is new. So software patents are even newer.
But you still have to define software patents.
Everything you wrote applies to non-software patents as well, with appropriate adjustments of the remaining wording. ("software developers" to "engineers", "open source" to "scientific research", "software users" to "consumers")
How is it so large just because it applies to some would-be hardware patents that are really nothing but hard-coded software patents?
If you have a patent on using gallium arsenide in semiconductor manufacturing, that isn't a software patent. If you have a patent on a fancy piece of glass to separate wavelengths of light from a fiber optic cable, that isn't a software patent. Because it's physics and chemistry rather than math. You shouldn't be able to patent math.
In a hardware sense, I can take apart an object (like a smart phone), and see how it works. In a software sense, it's much harder to decompile something and see how it is working if the scope of the project is large enough.
Software patents are bad (in my opinion) because the things which get patented are the things that form the foundation for what almost every programmer works with in their daily routine. Things that are abstract ideas such as "Connecting to a server to process a payment" might seem like a logical thing to patent as a software idea, but if I transferred it to its real life analogue it becomes unpatentable as a business practice: Person A gives payment to Person B, who then forwards payment to Person C.
All this being said, I'm not aware of any legitimate software patent suit that didn't involve corporate espionage. Can anyone link anything to show me a legitimate use of our court system in order to protect an innovation that should be protected?
For the argument of encouraging investment, the investment should be measured accurately, including risk.
Note: I'm talking about real invention (e.g. relational database, regular expression searching), almost all software patents aren't.
EDIT to clarify: (1) if someone is more capable, more educated, more talented, works longer, works harder etc, you generally must pay them more, hence greater capital investment. (2) To reward success in a risky endeavour is not about rewarding "luck", but making the reward worth the risk, in order to encourage others. This is an economics-and-law argument for patents by analogy with high-risk resource exploration (e.g. gold, oil).
The trouble with that is it's such an unclean line and there is so much money on the table that it doesn't solve the underlying problem of patent litigation proliferating in the software industry. To get rid of the trolls you have to get rid of the ambiguity as to whether what they're doing is allowed.
"No software patents" is useful because it's clean. If your product is a piece of software, it allows you to immediately answer the question of "does it infringe any patents" with "no" and then get on with developing the product, instead of dealing with skittish investors who won't risk a pile of money against the probability that an army of trolls will jump out from under the bridge as soon as they start to recover their investment.
The alternative I favor is adding an element to infringement that requires some evidence of copying, reverse-engineering, or appropriation. Since the plaintiff has to plead the elements of a suit up front, this will shift the fact-intensive part of the litigation over to the plaintiff unless they can offer some evidence of bad conduct by the defendant.
This changes the nature of a patent completely, but I think not in a bad way. If your patent is really worthwhile, independent invention should be negligibly probable.
Maybe. It could be worth trying. The concern is that trolls will just start selling a sham product in order to claim actual damages. And there is still the problem of industry incumbents using bad patents and the threat of ruinous litigation to destroy innovative smaller competitors.
>This changes the nature of a patent completely, but I think not in a bad way. If your patent is really worthwhile, independent invention should be negligibly probable.
If we could easily prove whether something was copied, sure. I think the problem is that it's so difficult for either side to prove that one way or the other, so whichever side has the burden of proof usually loses. If plaintiffs have to prove copying (and "proof" means something like defendants admitting they copied, and not something like the products being similar) then you're making serious trouble for legitimate patent holders when some Chinese supplier rips off their patents but they have no way to prove it wasn't independent invention.
Which is where it comes down to pragmatics: Drug companies (for example) aren't going to stand for anything that would create that much exposure for them, so you either have to solve that problem for them or we're back to creating carve outs for different industries.
That's an interesting suggestion, but what are the odds that such evidence will come to light when one of those things occurred? Is the discovery process reliable enough that we'd want to require patent holders to depend on it? After all, in such a case, the infringer knows they're infringing and can expect a lawsuit -- it doesn't seem like it would be that hard to destroy any evidence of reverse engineering in advance.
Here's another alternative: require that the plaintiff in an infringement case show objective evidence that their invention was nonobvious.
Basically, this means admitting that the PTO is only a first-level filter and removing the presumption that a granted patent is valid; the patent holder would have to defend their patent in order to litigate it. Along with that, we put some teeth into the nonobviousness requirement by requiring some evidence along the lines laid out in Graham v. Deere: commercial success, long-felt but unsolved needs, or failure of others. (In software, publication of a paper in a recognized journal or conference proceedings would probably be a common form of evidence.)
Like your proposal, this would be a huge change that would make most extant software patents worthless. But it doesn't depend on discovery to determine malfeasance by the defendant. What do you think?
I don't know if there is a need to be so skeptical of discovery. There is nearly always a danger of destroying evidence...
It has certainly never been true, in my experience, that the actual software development is the easy part, regardless of the quality of the spec. A spec detailed enough to make the software development easy would, itself, be a piece of software; it is precisely that process of reducing the idea to all of its details and harmonizing all of its competing constraints that is the work of software development.
I have no background in machine vision, navigation, or the like, but I have done some embedded software development, and one of my projects was a wearable device which did sensor fusion, data logging, networking, etc. It was a relatively small project, and the core algorithms were basically obvious as soon as we knew what the product was supposed to do, but it still took almost two years to implement it all, test it, see what real world data looked like, tune our constants, and run more tests, etc. There were many practical tasks involving power consumption, peripheral management, network communication, debugging, manufacturing tests, and the like which contained absolutely no new ideas but still had to be implemented in order for the product to succeed.
There may have been a patentable invention in there somewhere, but who would ever license it? Anyone trying to solve the same problem would come up with something similar, and they'd still have 18-24 months of expensive development work to do getting a product designed and ready for market. If we'd patented that hypothetical idea, we'd have contributed nothing to the public sphere; the only value in such a patent would be to make it more difficult for someone else to follow in our footsteps, which is exactly the opposite of the reason the patent system was established in the first place!
If you have a patent on, for example, the chainsaw, and are the sole manufacture of chainsaws, you can set the price of a chainsaw based solely on how much your tree-cutting-down gizmo is worth to the consumer.
As a concrete example, say you're ARM and contemplating investing $1 billion to put a new quad-core 3 GHz 1W CPU on the market. Because this a new product, you have a certain amount of pricing power to begin with. That pricing power allows you to extract positive economic profits, at least until your design is copied and the market becomes highly competitive. For the whole exercise to be worth while, you have to recoup that $1 billion in R&D before the market becomes competitive and prices approach marginal costs of production. If your Chinese competitor can copy your design for $100 million in 6 months, they have a huge advantage. At that point both companies will together likely still have a lot of pricing power, but because your competitor had much lower R&D costs to get the same product, he can undercut you o price and still recoup his lower R&D expense. In that sort of market, the profitable strategy is to copy, not to do original R&D.
As a practical matter, profit has to drive R&D, because there is no money for R&D without profits. And in perfectly competitive markets, there are no profits and no R&D. The only way to make real money is to produce products that are non-fungible, and differentiable. You can do that in only a few ways: R&D, network effects, marketing, or branding. If you look at the profitable companies in the tech industry, their profits come from one or the other. If you look at the companies that produce fungible products for highly competitive markets (Acer, Lenovo, etc), you see that they their profits are marginal and so is their level of innovation.
The solution to just set a higher price might not work, as people might not be willing to pay that much, or even have that much money.
Stephan Kinsella has written extensively in opposition to the state's intellectual "property" laws.
a) you get to build/use/sell an invention exclusively for a certain length of time, in exchange for b) publicizing how you did it.
A good number of software patents fail the second part of that test, to say nothing of originality etc., because they only declare the behavior and don't provide a reusable implementation - a lot of the time because they're derived from "wouldn't it be cool if..." spitballing sessions (especially in large companies).
Read his comment for some entertainingly wrong-headed views on how software (and, indeed, thought itself) works.
Ten years and we didn't make any progress, how miserable is that?
This is a fact that's widely recognized, even among economists.
Source: I've got an undergraduate degree in the topic, have been continuing to study it over the subsequent decades, and have been specifically researching in depth the foundations of the discipline in the past few months.
One rather striking failure of economics has been its lack of interest in long-term history. Which is peculiar if you've read Adam Smith as he actually goes into depth on the matter. It's the #2 economist, David Ricardo, who drops that view (though there continue to be economic historians), a fact which was criticized quite strongly by Arnold Toynbee (the economic historian, not his universal historian nephew Arnold J. Toynbee), whose Lectures on the Industrial Revolution begin with the observation:
It has been a weakness of the science, as pursued in England, that it has been too much dissociated from History. Adam Smith and Malthus, indeed, had historical minds; but the form of modern text-books is due to Ricardo, whose mind was entirely unhistorical. http://archive.org/stream/industrialrevol00toyngoog#page/n16... (page 1, 18 of the PDF)
It goes further than this. Nassau Senior, a strong early supporter of Ricardo writes: "Political economy is not greedy of facts; it is independent of facts." Serious WTF moment there.
The Austrian School (from which Classical Liberalism, a/k/a Libertarianism takes its lead) is even worse. Speaking of praexology, the interpretive method on which it's based:
"Its statements and propositions are not derived from experience. They are, like those of logic and mathematics, a priori. They are not subject to verification or falsification on the ground of experience and facts." -- Ludwig von Mises http://www.goodreads.com/quotes/846342-praxeology-is-a-theor...
Suddenly the whole "reality-based" vs. "faith-based" divide becomes a tad bit easier to understand.
There's a great deal else out there, and no, not all of economics is complete tosh, but a great deal of it really should be tossed out.
I'm still digging into the history behind Ricardo, but briefly:
- He was an ideological opponent (though apparently close personal friend) of Robert Malthus. Malthus strongly criticized Ricardo's theories, and repeatedly questioned their inclusion in the Encyclopedia Britannica, suggesting they wouldn't stand the test of time.
- Malthus himself became strongly criticized by Ricardo's supporters. Much of that present criticism focuses on Malthus's writings on population, though these largely focused on the poor, and expressed a viewpoint not much different from Ricardo's. But by focusing the attention here, Malthus's own criticism of Ricardo is distinctly underrepresented. Compare this with Helen Keller, whose notability to most people is her role as an advocate for the rights of the disabled. Her actual attentions during her life were in support of socialism. For some reason, that message doesn't make it into elementary school texts....
- Ricardo's principle work was his theory of rent, which focused attention on landowners and laborers, suggesting that capitalists would be left wanting in allocations of wealth.
- Ricardo's own wealth came from stock speculation and banking. He later bought a parliamentary seat, according to Toynbee (I'm not sure how unusual a practice this was at the time), and used his MP position to advocate for his political economy (as the study was called at the time) views.
- Ricardo's papers and letters were collected by John Ramsey McCulloch, and subsequently bought by Lord Overton, another banker. They're now hosted by a number of largely libertarian-leaning organizations including the Ludwig von Mises Institute and Online Library of Liberty.
Again: my working hypothesis is that much of economics is more a too for justifying allocation of resources to benefit a specific class than a dispassionate study of a system. Not a unique view by any means:
https://en.wikipedia.org/wiki/Rajani_Kannepalli_Kanth
Rajani Kanth's major philosophical theories include:
a) that Ricardian Economics was merely policy advocacy in theoretical guise,[2][16][17][18][19][20]
b) that Economics is the crown jewel of the hegemonic ideology of Modernism,[2]
c) that European Modernism, whose derivates are capitalism and socialism, is the real grey eminence that has crafted the alienated misery of our epoch ...
John Kenneth Galbraith has some similarly damning comments:
Leading active member's of today's economics profession...have formed themselves into a kind of Politburo for correct economic thinking. As a general rule --- as one might general expect from a gentleman's club --- this has placed them on the wrong side of every important policy issue, and not just recently but for decades. They predict disaster when none occurs. They deny the possibility of events that then happen.... They oppose the most basic, decent and sensible reforms, while offering placebos instead. They are always surprised when something untoward (like a recession) actually occurs. And when finally they sense that some position cannot be sustained, they do not reexamine their ideas. They do not consider the possibility of a flaw in logic or theory. Rather, they simply change the subject. N one loses face, in this club, for having been wrong. No one is dis-invited from presenting papers at later annual meetings. And still less is anyone from the outside invited in.
(Quoted in Robert Heinberg's The End of Growth).
I agree that much of conventional economics functions largely as apologia and "retroactive continuity" for existing power structures and legacy wealth. While there are obviously spectra of disagreement within the field, as a whole it does seem largely oriented towards received wisdom in its academic culture, but that's from a distinctly outsider perspective.
What's your opinion, if any, of David Graeber's ideas in "Debt: The First 5000 Years"?
That said: his is some of the most original writing on the topic of money I've seen, period, it's not obviously wrong, to the extent I've read it, it addresses a great many misconceptions baked into economics since the time of Smith. Specifically: correcting Smith's descriptions of pre-money economies as being barter based, which Graeber replaces with gift-based (though that's something which works only in communities up to a certain size, most communal communities peaking around 300 members).
Money in general is something that fascinates me (in an abstract, what-is-it and how-does-it-influence-the-economy sense). It seems intrinsically tied to both the fossil-fuel lead growth of the past 200 years. As some have pointed out, the US transition to a central bank, fractional reserve, fiat currency pretty much tracked the milestones of industrialization: railroads (1830s), electricity and industrialization with the start of the petroleum boom (1930), oil imports and Bretton Woods (1950s), and the transition past peak oil with the birth of the petrodollar (1972). Dates somewhat approximate. There's the point that's raised by many (Richard Heinberg and Chris Martenson in particular) that fractional reserve lending and monetary supply growth mandate economic growth, which I haven't fully subscribed to, but that would suggest one primary source of opposition to sustainable, zero-growth economic paradigms.
That's not really a correction; there are pre-currency societies with both barter- and gift-based (and there's not really a clear delineation between the three, there's overlap between barter- and gift-based structures in pre-currency societies, and overlap of all three once currency is in the mix.)
> There's the point that's raised by many (Richard Heinberg and Chris Martenson in particular) that fractional reserve lending and monetary supply growth mandate economic growth, which I haven't fully subscribed to, but that would suggest one primary source of opposition to sustainable, zero-growth economic paradigms.
The human desire for better things demands economic growth; zero economic growth literally means things never get better (currency-value-based measures of economic growth are proxies for changes in experienced utility; zero economic growth means no increases in experienced utility. That's not really a desirable goal.)
Smith doesn't allow for the concept of a gift-based economy. Or for the possibility of an explicitly accounts-based economy, also supported by the anthropological record.
> zero economic growth literally means things never get better
I've argued both sides of that statement myself. I think I'd argue it on an evolutionary basis: the level of supportable complexity is static, but the focus of that complexity, where complexity is used to provide solutions to problems, isn't fixed. As the environment itself changes, or even interests within that environment, changes can occur within the economy. It's bounded but not static.
The material standard of living cannot increase overall, but its application and allocation can be shifted.
And whether or not the goal is desirable really has zero impact on how real the situation is. That's a classic wishful thinking response -- or if you prefer, argumentum ad consequentiam, appeal to consequences, or Diax's Rake. The Universe doesn't owe you your desired lifestyle.
Thats not to say that if they started from scratch there wouldn't be some unforeseen bugs down the road.
Which is sort of like writing code without ever referring to your specification, use-cases, and/or users, or in fact, compiling or running it.
Now, there's the Great Recession that got going in 2008 which economists should be embarrassed about not seeing as a consequence of the crash of 2006, but we all know that economists don't agree much about macro stuff. By contrast, the effect of patents is all microeconomic. And microeconomics is the part of economics that is actually generally understood and agreed upon by professionals.
As Jeffrey Reuben has pointed out, all but one of the post-WWII recessions in the US was preceded by an oil price rise. Including the 2008 recession.
Agreed that microeconomics (mostly) works better than macro, though it too has its issues.