Slashdot Mirror
Interview With 'Idiot' Behind Key Software Patent
An anonymous reader writes "Last week, an appeals court ruling opened the door to making it easier to kill software patents. It turns out that the guy whose name was on the actual patent didn't even realize it was at the center of the debate, and doesn't like software patents very much. 'So I was thinking — great they invalidated software patents, lets see what crappy patent written by an idiot they picked to do it — then I realized the idiot in question was me.'"
It just shows that most people have double standards. When they or someone they know do it, it's all good. When it's someone else, it's the root of evil.
There are lots of errors and other assorted silliness in the article. For example: "But because the USPTO focuses much more on "prior art" (i.e., "is this new?" rather than, "is this obvious?") all sorts of obvious stuff gets patented."
In fact, obviousness rejections are extremely common. In my experience they are the most common kind of rejection. Moreover, the obviousness analysis is based on prior art. The analysis is basically thiis: would it have been obvious at the time to put these pieces of prior art together in order to create the claimed invention? Requiring prior art evidence of obviousness is important because it helps avoid hindsight bias. Note that the KSR decision made it easier to find things obvious by invoking 'common sense' and 'common creativity,' often with fairly minimal evidence.
"if a patent doesn’t actually tell you enough information to understand and build the invention, it shouldn’t be valid."
This is already a foundational part of patent law. "The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same." 35 USC 112. I would agree, though, that these requirements (called 'enablement' and 'written description'), are not applied rigorously enough by the PTO or the courts. The PTO's policy on software is especially silly. For example, it prefers flowcharts over pseudocode to describe algorithms. Not only is this not very searchable, it's also decades out of date and captures far less detail about an algorithm than pseudocode.
"End the venue shopping for lawsuits"
The Federal Circuit has been clamping down on venue shopping somewhat. I wouldn't say that the Eastern District of Texas's days are numbered quite yet, but litigants are definitely finding it easier to get out of there. See, e.g., In re Genentech, Inc., 566 F. 3d 1338 (Fed. Cir. 2009); In re Microsoft Corp., No. 944 (Fed. Cir. Jan. 5, 2011).
"Expedited review of new and disputed patents by a panel of experts in the field with a high bar for validity"
Who determines who these experts are? And who would review their determinations? A new appellate court of super-experts in every field? Or would it be back to non-experts? The PTO can barely keep itself staffed with non-expert examiners and board of appeals judges, much less actual experts in every field.
Furthermore, the reality is that litigated patents are already reviewed by a panel of experts: the expert witnesses called by the parties.
From TFA: "... because the USPTO focuses much more on "prior art" (i.e., "is this new?" rather than, "is this obvious?") all sorts of obvious stuff gets patented."
I couldn't agree more.
Except that both the article author and you don't understand what prior art and legal obviousness are. That prior art exists does not mean that something is not new: prior art is any relevant art in the field that was published or publicly available prior to the filing date. RFC 793 (TCP) is prior art for RFC 2460 (IPv6) because it's in the relevant field (networking protocols, albeit a different layer) and was published earlier.
What you and the article author think prior art means - "is this new" - is actually whether an invention is novel under 35 USC 102. An invention is not new if a single piece of prior art discloses each and every element in the claimed invention.
Obviousness, on the other hand, means that while a single piece of prior art doesn't disclose the whole invention, multiple pieces of prior art can be combined to teach or suggest each and every element of the invention. In other words, if the patent application claims "A+B+C+D" and no one has ever done that before, it's new... but if a prior art "A+B" exists, and prior art "C+D" exists, then it may be obvious to combine those two pieces of prior art to get "A+B+C+D" and the invention is obvious.
Why this is important is because, by misunderstanding that art has to be prior, even for the purposes of obviousness analysis, you're asking the wrong question... Essentially, you're asking whether, in hindsight, the innovation was too trivial or minor to be awarded with a patent. But that's improper - almost everything is obvious in hindsight. That's why the patent office has to rely on prior art to show obviousness - they can't just say, "eh, I think this invention is obvious." Rather, they have to explicitly list the prior art references that can be combined to teach each and every element of the invention. If they can't find one - if, "A+B" exists, and "C" exists, but there's no art for D, then A+B+C+D can't be obvious except in hindsight.
OK, so let's trot out the old chestnut of pharmaceutical patents. Let's say I invest $100 million to do the research needed to create a new drug, I create it, and the doses cost only $1 to make. Extrapolating from current demand, I will sell a million doses in the next 20 years while my patent is in force. I set my price at $200 per dose; with the costs being $1 in manufacturing and $100 in repaying my investment, so I'm making $99 in profit per dose. After 20 years, I have $99 million dollars in profit.
If I don't have patent protection, as soon as my drug hits the market someone will do an analysis and make a generic clone of it, selling it for $100 a dose, also settling for a $99 profit margin. They would take over 99% of the business from me, leaving me stuck with the tab for about $99.99 million dollars in research investment.
Would any pharmaceutical company ever do research again? Would anyone be trying to cure anything if they thought it would bankrupt them to do so? My guess is that a few celebrities will be affected by some diseases and create "foundations for the cure" efforts on a one-off basis, but in general, innovation in drugs would die without the patent system.
Of course, this is a simplistic picture, and the real world of shady marketing, fraudulent studies, suppressed side effect reporting, drugs to treat imaginary ailments, and all the other unethical stuff the pharmaceutical companies do certainly complicates things. And there are ongoing costs to the inventors and manufacturers of the drug: lawsuits over side effects, wrongful deaths, etc. But at its core, without the patent system these drugs would never be created.
So if we've established that pharmaceutical patents are necessary to drive research that may benefit us as a whole, then at least part of the patent system should at least be salvaged and not dismantled.
John