Among technologists, it is commonly imagined that once the USPTO issues a patent then the rights and claims within that patent are ironclad. The argument is thus: did not the government experts examine the claims? Did not the experts at the USPTO make their decision to grant the patent? Will not the government defend its position? The answers are yes, yes, and not exactly. In practice, the courts ultimately determine the validity of a patent and its claims. This latter point is worthy of dilation.
The problem starts with what things are considered patentable. Contrary to even a few decades ago, most patents today are not for palpable inventions like latches, hooks, tools, or mousetraps. Today, the claims often lay upon algorithms, software, processes, methods, and mechanisms, all impalpable and convolved. I have great admiration for the professionals at the USPTO who wade through mountains of documentation and arguments to come to a resolution. Nevertheless, the complexity and abstractness of some of these claims confounds even the greatest of experts. And, experts do often disagree with each other, as every technology wizard knows. In the past, when most inventions were evidenced by a prototype you could hold in your hands (so to speak), the USPTO relied upon its own patent database to weigh against new and unique claims. Now, however, with these intangible inventions, such as those for a process, the USPTO must rely upon other databases. Hence, the USPTO makes its best determination based upon the claims and validation of the filing, but by its very nature, the claims are for untrodden avenues of technology. Also, the rate of change in some areas of technology is exponential. Can one keep the validation databases current? Secondly, there is a trend to open source information, implementation, and engineering. What then, constitutes prior art and ownership? In the final analysis, if the claims are contested, the court will adjudicate.
Thus, a patent is a document of considerable value, a deep stake in the ground, but it is not an anchor against all winds.
In practice, when dealing with algorithms, software, processes, and mechanisms, I have found the greatest obstacle to be what is referred to as “prior art.” In other words, is this invention really unique or would anyone with a reasonable knowledge of the art have come to the same conclusion? And, did they already come to that conclusion in some documented form? This is a difficult question to answer. For instance, I know of a company that owns a patent whose claims include the use of the digital imaging for examining and displaying the inside of a house, such as one for sale.
Now, anyone who has ever even thought about buying a house knows you can go on line and view almost any house on the market by seeing a so-called image “walk-through.” This is not a video of the interior of the house but a digital compilation of distinct views – exactly the things covered in this patent. “Are you kidding me?” you ask. “You mean someone has a patent for that? They must be making a fortune on the patent!” Yes, there is a patent exactly for that, infringed upon every second by hundreds of companies, but not a single dollar has ever been made from it. The problem is this: is the patent defensible in court? Clearly, it was issued by the USPTO but is the invention really unique or will the court determine that anyone with a knowledge of digital imaging, i.e., prior art, would have recognized that as an obvious application and did so recognize it? Is that recognition verifiably documented somewhere? This particular patent to which I refer has never been taken to court so one would conclude that the attorneys who own it likely feel it is not winnable. Hence, the patent was issued but the claims are probably not defensible. Be very careful of this when you seek to patent methods and processes. It is thorny ground that has grown because of the types of patents now issued.
Consider that Jack Kilby worked for TI in 1958 and invented (perhaps co-invented with Robert Noyce) the integrated circuit, for which he received no less than a Nobel Prize in 2000. I worked in semiconductor processing a number of years after Kilby and always wondered at the awe given to his invention which to me seemed almost prior art. If you knew how to lay a silicon substrate for one transistor, which was common at the time, and if you could put two transistors in the same package, which was common at the time, would it not be obvious to anyone skilled in the art that you could put several on the same silicon (or germanium) substrate? Perhaps so, but the clear argument is that Kilby was the first to do it and the USPTO could verify this. It was a tangible invention. and they could validate from their own patent databases that this invention and its claims had not been patented prior. Keep in mind that in 1952, Geoffrey W. A. Dummer had presented the concept of an integrated circuit at a U.S. electronic components symposium. Envisioning the concept does not fulfill the requirements of an invention.
Such is not the case for intangible inventions. If you are considering filing a patent whose claims rest upon processes, algorithms, or similar, this issue of prior art needs to be weighted heavily. You might spend many hours and thousands of dollars, receive a patent, and only then have the court decide against it because of this issue of prior art. And, as patent coverage becomes broader, claims shallower, inventions less tangible, and society more litigious, the need to circumspectly evaluate prior art will be critical.