Cutting Through the Patent Thicket

xzvf writes wrote to mention a BusinessWeek article positing that the overgrowth of patents is harmful to innovation. From the article: "The first problem with patents is that the entire process takes too long: three years on average, often as long as five, and getting longer all the time. So when a venture capitalist invests in a company, its IP 'dowry' remains, at best, provisional. How much would you pay for a company when its assets are hidden from view?"

Comments

His 4th problem with patents

[TubeSteak]

Fourth, and probably most important, few venture-capital-backed companies will ever dare to defend their IP in court. If they do, they'll risk losing customers and squandering anywhere from $1 million to $5 million of their precious venture funding.

I wonder what he's basing this statement on.

I mean... I guess it makes sense. But I don't see how
defending patent = losing customers

The most significant bit of the article

[dillon_rinker]

Frmo the article:

"While at AT&T (T) in the early 1990s, I sponsored two separate ideation sessions around a potential new market, bringing in 50 experts each time to brainstorm for applications. Both groups generated ideas with real commercial value.

Both groups, however, generated more than 95% of the same ideas in common. They were "obvious" in the fullest sense of the word and would have been commercialized with or without the incentive of a patent. But the Patent Office found them "novel," and issued AT&T claims by the basketful."

This, in a nutshell, is everything that is wrong with the patent office. Most patents granted are NOT non-obvious. I would suggest that what the patent office needs is a peer-review process.

False premise

[Anonymous Coward]

For over 200 years, the U.S. patent system has catalyzed economic growth and protected the national interest.

Bovine excrement! This is a fallacy, it can neither be proven or disproven. Repeat after me; "patents are not a metric for innovation".

Re:Another foul use of the word "troll"

[nhnfreespirit]

Except thats not what is happening... What is ususally labeled a patent troll, is someone who buys up patents with no intention of ever doing anything with the invention covered. The simply sit on the patent until someone invents (and markets and promotes) something (often only vaguely) simmilar and then they sue them. This is basically what is happening in the NTP vs. RIM case. NTP has never used these patens to create a product or invest the money the poor inventor lacked to commercialize his idea. This is a parasitic behaviour at best.

One reason for the current state...

[mavenguy]

...of the US patent system concerns the issue of obviousness, and is illustrated by the following quote from TFA:

And much of what the Patent Office sees as invention is merely science applied to a new field by equation or analogy. At AT&T, we took old microwave patents and filed identical claims on optical inventions, which are also radio waves, only 10,000 times smaller. We were able to do this even though it was obvious to anyone who ever picked up a physics textbook that once you have the ability to make things smaller, the physics just translates over.

When I started in the Patent Office (This is before it was renamed Patent and Trademark Office), it was common to use this kind of reasoning to make rejections, and you would be sustained if challenged, because this was the "accepted" view by the Board of Appeals and the Court of Customs and Patent Appeals, the judicial appellate court from the Board, at the time (since then merged into the extant Court of Appeals for the Federal Circuit, who follows the CCPA law). Examiners were accorded great respect in making obviousness judgements and in dicussing the prior art references.

But, about the same time, the patent bar started to become more aggressive in challenging these rejections, demanding that the cited prior art show, or, to use current terminology, "suggest" reasons why the cited prior art references "would" be combined. Say, hypothetically, the applicant claimed a light source, a mirror at 45 deg to deflect the beam 90 deg, a modulator modulating the deflected light beam, and a detector to, well, detect the deflected, modulated beam. Now consider some prior art: Firstly, a light source producing a beam going directly to a modulator, thence to a detector and, secondly, a reference showing a mocrowave source sending microwaves to a microwave reflector, followed by a microwave modulator, that then followed by a microwave detector. In the old days you could combine the two references in an obviousness rejection and, for the applicant to overcome the rejection (beyond adding significant limitations to the claim(s)) he would have to provide some convincing argument, perhaps supported by evidence supported by a "132" oath/affidavit, showing why such a derect analogy was incorrect. But this is no longer true; now the burden is on the examiner to provide prior art that specifically shows that optical and microwave elements can have similar designs. Of course, any such reference, if not exactly showing the source, reflector, modulator, detector combination would be attacked as not applicable to that combinatio; of course, if you had such a reference, you'd have an anticipating prior art and would make the stronger "102" rejection in the first place ("anticipation is the epitome of obviousness")

So, why the change? The answer is, basically, the CCPA and its successor, the CAFC. Attorneys kept appealing and winning reversals, and the Court opinions in those cases clearly kept raising the bar on making obviousness rejections. It's not something that, on the surface, is very stark, but it has greatly increased the burden to make each rejection. Now you might have to search twice (yeah, pulled out my butt but probably in the ballpark) as long, even with modern online search technology, to get the prior art needed to support the rejection. Furthermore, you might not find just the exact "teaching" reference to put you over. The result is claims, and applications, go to allowance that would have never made it under the environment that existed years ago. Couple this with extreme PHB management culture that has developed over the past 30+ years and you have the current mess. And, due to the explosion of filings in recent years, even though the examining corps has increased by a factor of 4 or 5 since around 1980 there is still a 3 year pendency in many arts, yet examiners have even less time, due to lots of tasks, many having no positive impact on examination, heaped on top of them, and the fact that the average time allotted to examine an applcation, has not changed since Commissioner C. Marshall Dann gave a whopping one additional hour per application back in the 1970s.

Re:And The Answer

[yog]

The U.S. patent system is seriously flawed. In my opinion, it's all these vague "business process" patents that have really screwed things up. They get hundreds of thousands of these applications per year clogging up the system, and the net result is that nothing gets done before at least three years, as the article points out.

Even worse is that the business process patents make it nearly impossible to implement anything without violating someone's patent. I looked into patenting an invention that had to do with a linux-powered answering machine and soon discovered that almost everything you can imagine has been patented. Some guy got a patent for "compressing a voice recording". So do I have to ignore his patent and let him sue me, or send him money for something that's a questionable "innovation" at best? Perhaps megacorporations can afford this hassle but not that many individuals, I would think.

The patent system was originally intended to encourage innovation by protecting people's rights to their inventions, and it has now been perverted into a thicket of pointless, indefensible rules that inventors must navigate to get a product out the door.

Probably the solution is to tighten up the definition of an invention and, as so many in this forum and elsewhere have pointed out, invalidate software patents and business process patents. Even Congress is supposedly getting wind of the problem, but I'm not holding my breath until it's solved.