EFF, Public Knowledge and Engine tell the USPTO how to improve patent quality




A common concern about software patents is that applicants are able to use invented terminology to avoid prior art.

Surely the examiners investigating the patents should have enough technical knowledge to be able to do a search on the common terms that are being obfuscated.


The best idea I have heard:
The holder of the patent must demonstrably, clearly use the idea in a product within the first 2 years of issue. A product which has some hope of actually appearing in the marketplace. Otherwise the patent is void.


Quick, patent the idea of a searchable database of patents.


In addition to the patent application, the applicant should also provide two patents that are less deserving of their status than the submitted idea, along with supporting documentation and arguments. If those two patents can be overturned using the information supplied, then the application will be given a fast-track to approval.

Alternatively, the applicant could instead provide only one patent if it is one that has been approved through this fast-track process, ensuring that the new patents will be more heavily scrutinized than the old ones.

Let the bad-patent cannibalism begin!


That's really not so simple. The reason is that for each common term, there's a virtual infinity of possible obfuscations.

Software patents are usually kept in a vague and silly metalanguage with lots of generic terms. That makes it really, realy difficult to search them for anything specific.

The real solution to many (not all) patent-related problems is to ban software patents altogether.


How about: For 1 year after filing (before being disclosed), the patent office sits on the patent. If another patent for the same idea comes in from an independent party, both are considered invalid.

A lot of times the technology just isn't there yet for a lot of these innovations, and once the technology comes along it's suddenly a race to be the first guy to lock in the idea that everybody had.


Should it be understood from the emphasis on fighting "abstract" and "vague" software patents that EFF has given up on having software patents abolished altogether?

Here's a working definition of a "software patent": If you can violate a patent by installing software onto a device, then that patent is a software patent, since it depends only on the algorithms in your program. In that case, the idea should be non-patentable, since it's really just a mathematical statement, which can't be patented.

That's how it used to be in the US many years ago, too. That would solve a lot of the current mess.


Another idea: Let the fee for submitting a bad patent application and have it rejected be four times as much as that of having it approved. That gives the USPTO a reason to act as the control authority it should be rather than as the rubber stamp it apparently is.


Indeed they are. The summary is a bit misleading, in that it suggests that the EFF is asking the USPTO to actually look for prior art. The real request is actually a request to develop a system to make searches for prior art actually turn up useful results.

Prior art is a difficult thing to research, because you're often searching for evidence that someone before the filing claimant thought of a similar idea and considered it just not worth filing a patent on. You're looking for industry periodicals, journals, doctoral theses. In the case of software patents, you can expand that search to textbooks that use the claimed code (or similar) as an example. For the really low-level things that people try to patent, even general mathematics textbooks (under the theory that nothing is more obvious than copying an equation out of a book). Rather than develop a corpus of searchable prior art -- and assuming that Congress doesn't have a sudden epiphany on software patents -- it makes more sense to increase the ability (read: decrease the cost and effort) to challenge a patent on prior art grounds at every steps of the process, and even after a patent is issued.

The examiners know what they're doing. However, the lack the time for in-depth searches for every such claim, or equivalently, the department lacks the funds to hire more people. The government has taken neither of the steps that would help alleviate the problem: pay for more examiners, or wipe away large swaths of new patentable IP (software patents) that are consuming most of the time. The EFF advocates for the latter, but probably won't be able to get it due to the complications of trade agreements.

You see, the U.S. has done a magnificent job crafting trade agreements which bind its own hands. Most of these agreements, such as NAFTA, are crafted to account for what happens when a government no longer wants to "play fair" and tries to tax some business, nationalize some industry, or invalidate some class intellectual property by re-classifying it. How do you have legal protections when government lawmakers are the violators? To start, you create an international system of extra-governmental courts that's beholden to no one that can dole out fines which other member governments can collect directly as tariffs. Then you make a whole slew of things trade violations, including the creation of new classes of "lesser" IP, or the destruction of whole classes of existing IP. Which when you get down to it, is exactly what the EFF proposes. Even if the U.S. realizes it would be a good idea to remove or alter software patent protections, it can't without violating its own trade agreements.


What I would like to see is an additional formal collaborative process to petition for ex-parte reexaminations with minimal formal and financial requirements - crowd-sourced patent hunting, if you will.

Of course with substantially lowered requirements it would be impractical to react to every petition, but it should be possible to set up the system in such a way that the patent office is able to select a certain number of promising candidates.

Unfairly overlooked petitions could still serve as inspirations for conventional (expensive) requests for reexamination.

Of course I am not holding my breath.


The biggest problem with curating the quality of patents is that litigators aren't competing against each other to exclude bad patents. Instead, they just cross-license and acquire, and with a worldwide patent regime they've ended up falling into transnational conglomerates who have no interest in curating the patent system. It's the owners of IP in collusion against the bureaucracy, and since Reagan the bureaucracy doesn't even fight.


"just cross-license and acquire"

Yes and no. "Cross-license and acquire" makes it sound formal and planned. Instead, most companies today have a portfolio of (mostly bullshit) patents. If someone tries to sue them for infringing on a BSpatent and they are a real business that actually makes things, then odds are that the suing company is infringing on someone else's BS patent. If your company (and portfolio) is big enough, the odds of them infringing on your BS patent increase, and then you have leverage. On the one hand, it keeps large, real players more-or-less honest. But on the other hand, that's new incentive to file as many BS patents as you possibly can, in the hopes that some will get through and add to your portfolio. And this also does nothing to deter patent trolls, who don't make anything and thus never infringe on others' BS patents. With them all you can do is let them wet their beaks, or get your lawyers to shoo them off. So patents protect big companies from other big companies, lawyers on retainer protect big companies from trolls, and trolls prey freely on small companies.


Back in July I wrote to my US CongressCritters and asked them to consider action to improve Patent quality. Feel free to mine my letter for ideas. I asked them to consider that the US has made 6 fundamental mistakes in our approach to patents. I believe they are:

  1. More Patents are not better than fewer patents. Patents are not Innovation. Patents are not Progress. Patents are simply grounds to file a lawsuit against an industry. More Patents are simply more grounds for more lawsuits. Reform must recognize that patents are dangerous monopolies. Reform must place hard limits on the number of patents.
  2. Running the US Patent Office as a cost-recovery operation is a mistake. The US Patent Office is a very small, but critical component of the US economy. It's purpose was "..to promote the Progress of Science and useful Arts.." But, once the USPTO became completely cost recovery (Omnibus Budget Reconciliation Act of 1990, Title X, Subtitle B), that primary goal became overshadowed by the more pressing goal of securing funding via patent fees. Reform is painful, but simple. Admit cost recovery is a failed experiment. Revert the funding model to the model used for the first 200 years. The USPTO must be centrally funded by the US government. Any collected fees should be returned to the US Government.
  3. It is a mistake to organize the US Patent Office to create economic incentives to grant poor patents. Currently most of the revenue of the US Patent Office comes from GRANTING patents. See the USPTO FY 2013 President's Budget page 37: "..More than half of all patent and fee collections are from issue and maintenance fees, which essentially subsidize examination activities." A recent study by the Richmond School of Law found that the USPTO's actual grant rate is currently running at about 89% Reform could come in many forms, but the simplest and most reliable would be to eliminate and unify the Patent office fees into a single filing fee. This fee would provide no guarantee of receiving a patent, only a guarantee that your patent would be considered. This would free the Patent Office to be able to deny poor patents. The filing fee should also be high enough to discourage spurious patent applications.
  4. Scaling up the Patent Office to produce more poor quality patents is a mistake. Currently, we expand the number of patent examiners based on demand. See the USPTO FY 2013 President's Budget, page 60, Gap Assessment: "Meeting this commitment assumes efficiency improvements brought about by reengineering many USPTO management and operational processes (e.g., the patent examination process) and systems, and hiring about 3,000 patent examiners in the two-year period FY 2012 and FY 2013.." The underlying assumption is, more patents are better, even if it means decreasing examination, and increasing the number of untrained examiners. Poor quality is an inevitable result of this patent process. Reform must tightly control and limit the number of patent examiners.
  5. It is a mistake to grant all patents that meet minimum standards. A review of the last couple decades changes in the patent approval criteria shows that the minimum standard for granting a patent has consistently shifted downwards. We must abandon the idea that any patent that meets minimum standards is granted. Over time, the courts apply constant pressure to degrade the minimum standard. Reform is easy. You rank Patent Applications according to an agreed measure of quality, and only grant the top few percent. Over time, the pressure will be to improve the quality of patent applications, instead of degrade them.
  6. Finally, it is a mistake to allow patent applicants to modify or extend their patents after submission. This complicates the patent pipeline. It facilitates ‘submarine’ patents. It enables capturing Standards. It also enables gaming the patent system. Reform must simplify and reduce the patent process. Patents should be quickly evaluated. Most should be denied. If an applicant wishes to modify a denied patent, they should alter it, resubmit, and pay a new filing fee.

    The pressures that degrade the patent office are always with us. Our nation has had multiple cycles of patent quality. If we boldly face this problem, we can overcome it.


You're preaching to the choir on that last point. I'm in favour of the old system that required a working model.

But again, if a patent examiner doesn't understand the terms used in the patent application they should be able to reject it on those grounds alone - which I guess is somewhat the point.


One could at least say, that if a patent examiner doesn't understand the terms used in a patent application, they should NEVER grant it - somebody else should. If nobody understands the terms, nobody should grant it.


USPTO should implement an "originality bounty". The way it would work is that anyone filing a patent would post a bond guaranteeing the originality of their solution for each claim made. Say, $5,000 per claim. The patent under review then enters a proof period, in which the claims of the patent are advertised, but not the specific solutions ("embodiments") offered by the patent applicant. Anyone is free to submit a challenge detailing their solution for satisfying any of the claims. A USPTO referee examines each challenge, and if it is substantially similar to the solution offered by the pending patent, then the invention for that claim is ruled non-original, and the posted bounty is paid to the challenger. Only claims which are not successfully challenged during the proof period are granted patent protection, and have their originality bonds refunded to the applicant.

I predict the number of patents granted would drop to a single-digit percentage of current volume, and a number of hungry college students would have their real-world skills honed and their bellies fattened off the forfeited bounties.


This topic was automatically closed after 5 days. New replies are no longer allowed.