Software patents are one of the most contentious areas of intellectual property. Check out this famous article or this website or this NPR piece, or just google the words “software patent” for a small sampling of ire towards the monopoly that is the software patent. No other type of patent draws quite the criticism that software patents draw, and it’s rather shocking how often the topic is discussed relative to how infrequently other sorts of intellectual property debates enter the public discourse.
So why is it that our society, one which generally loves the concept of property (though the music and movie industries might disagree), has such a strong aversion to patents in this specific realm? I have a few ideas:
Culturally, the creators of the subject matter covered by software patents are not considered “inventors” in the classic American sense.
When you think of American inventors, images of Henry Ford and Thomas Edison come to mind. Maybe Alexander Graham Bell, even if you forget that he was Scottish. Regardless, the cultural perception of “inventors” tends to direct our focus to, well, white men who make physical objects. They are invaluable, nonfungible assets. The adjective “inventive” implies creativity, and the concept of the inventor as a creative entity is actually captured in various parts of patent law itself. The things inventors create are tangible, useful, and sometimes amazing.
Software, on the other hand, is basically a meaningless word to most people – essentially everything that happens on a computer screen is a result of this mysterious software. And the software engineer is perhaps the most commonly misunderstood profession to the majority of Americans, with few people outside the practice really understanding what C++, Java, or Ruby on Rails even are. Ask yourself – Who are the famous software engineers of our time, and are they considered inventors? Mark Zuckerberg is hardly referred to as a famous inventor, despite Facebook’s 46 patent filings. Bill Gates isn’t usually described as an inventor either. Steve Jobs, perhaps, but only in the sense that he designed rather than coded.
And this sentiment isn’t just from the uninformed masses outside of Silicon Valley – software engineering is frequently an outsourced project even in the heart of software engineer utopia. Many entrepreneurs opine that if they could just get an engineer to perform their vision, their companies could achieve great feats. This popular perception of software engineering as a tool, as a non-creative task that can be outsourced without suffering much in terms of quality, lends itself to a view of software engineering as something less than inventive. And that the product of this work is undeserving of the title “invention”, and thus, undeserving of a patent. (Note that I don’t think software engineering is uncreative or uninventive – it just seems that many people do.)
Practically, software patents are extremely valuable, software is increasingly where the world is headed, but nobody knows when they infringe a software patent and they don’t seem to further innovation in the slightest.
There are two major sides to the practical aspect of the software patent debate, the side that sees patents as critical to covering the subject matter that is becoming central to innovation on the planet, and the side that sees patents as a hinderance to that subject matter, with the patents themselves completely disconnected from the actual innovation.
The pro-patent side of that debate would note that a problem with software patents is really a problem with patents generally, an argument I’ll credit to Paul Graham (first link of this post). Further, patents have covered innovative technologies for the history of our country, and generally they have been seen as a mechanism to encourage innovation and investment into valuable areas of research. Additionally, as software becomes more and more central to our lives as we carry more and more computers with us everyday, continuing without patents opens the industry up to the risks that threaten all fields without IP protection. Patents are credited with preventing big companies from ripping off the innovations of smaller companies and individual inventors without costs, and a world without patents would theoretically favor established companies and hurt the entry of new competitors into the market.
The anti-patent side of this debate would point to a large body of empirical work that has shown software patents in particular to be ineffective at inspiring actual innovation. Technology companies tend to load up on software patents not because they are inherently valuable to the company, but because they are tangible assets which can be shown to investors in fundraising and because they can be used defensively to ward off patent suits by competitors. Software engineers tend to see software patents as evil and unnecessary, and generally unable to capture the essence of what they created anyway after being filtered through a legalese translator. There is also a general feeling that, as fast as technology moves, rewarding a decade monopoly to the first inventor of an idea is undue compensation when the second inventor was probably just weeks or months behind.
Regardless of which side of the debate a person falls on, big companies have invested hundreds of millions of dollars each in their patent activities, be it through litigation, R & D, or simple patent purchases. Eliminated software patents entirely would be very unpopular with these companies, even if each individual in the company disagrees with the practice as a whole.
Legally, software patents tetter on the fringe of being unpatentable algorithms.
It’s been a general principle of patent law since the inception of patents that our country won’t award a patent to a mathematical formula or to an abstract idea. The extension of this principle determines algorithms to be unpatentable as well. The basic problem with software patents, then, are that they are a combination of algorithms (which are not patentable) and written computer code (which are not patentable, but are copyrightable), yet the combination of those two unpatentable concepts yields a patent, and has for the last 20 years. This is hardly well-settled law, even after all this time, and the Federal Circuit is still struggling with what makes software patentable compared to concepts, which aren’t patentable. Two recent cases, Ultramercial and Cybersource, essentially contradicted each other on whether software instructions were simply mental steps or whether they were more tangible, so we really aren’t close to a consensus on where software patents fit in the messy legal framework of patentable subject matter.
The way software is moving, maybe we don’t need patents in this area as much as we might have in the past anyway.
Two major trends in software are lessening the justification for patent protection in software – for one, software is moving to the cloud, and two, software is adding social layers. Both of these trends support the notion that patents are either ineffective in the space, or at least becoming less necessary.
For one, software served over the internet (or the cloud, or whatever metaphor you choose to use to describe a system where applications reside on the internet rather than the hard drive) is less prone to blatant copying, because the source code can’t be accessed by the user. Somebody seeking to replicate Salesforce’s specific technology has their work cut out for them as compared to somebody seeking to do the same with a program that runs on their own hard drive, with physically accessible code. If patent law is designed primarily to promote development of new technologies by assuring the inventor that there will be adequate protection, then that justification is less convincing in areas of research where the end product is never fully in the hands of the user or the public. The details of a web-based service like Salesforce or Facebook are only really known to the engineers who invented those details in the first place, and trade secret law provides adequate protection for those companies without the costs of patenting and the deadweight loss on society that patent law likely inflicts.
Second, many areas of software, including the cloud-based business solutions above, see success through their reliance on network effects. Network effects is the term used to describe the phenomenon where the power of a network (and its value) is positively correlated with the number of people on that network. It’s why we all have Facebook accounts but aren’t rushing to sign up for Google+. It’s economies of scale but with regards to social interaction more than economic efficiency. If I was to try to copy Salesforce or Facebook, patents wouldn’t be my primary obstacle – My problem would be that no consumers would want to use my network without other users. In fact, only when I became a big threat to either company would they consider leveraging their patent portfolios against me, at which point I would probably have patents of my own to countersue with, and I could spend money on litigation fighting to invalidate their patents. Only in very rare situations, then, do patents become relevant as protection for a software business model, and when they do they tend to be economically inefficient at working to provide that protection.
What’s more, our current policy on software is one of protection, and what we have seen is nothing less than an explosion of trolling organizations and patent nuclear war between major technology companies, draining millions of dollars away from actual innovation in the space. So any sort of reasonable change is probably worth a try.
Potential solutions? Not likely: Are there answers? Sure. Our country could do anything from categorically prohibiting patent protection for software, to simply lessening the amount of time the monopoly is good for, to instituting a software-specific rule where the obviousness analysis is made from a much later point in time to accomodate for the fact that most software “inventions” are obvious just a few months after invention. But with Congress just passing a patent reform bill that did little to address software patent issues, the only changes to the system are likely to come from the Federal Circuit, who probably can’t find a way to work those changes into patent law.