The simple fix that could heal the patent system
slate.com
slate.com
The assumption of the 19th century patent system was that everything that should be patented already was patented already; therefore, when you would bring a invention to the office, the correct way to determine whether it is novel is to ask whether it is novel in light of the art already in the office. When asking whether it is obvious, you are essentially asking whether the very semi-skilled form of art knowledge possessed by the patent examiner would have regarded this as a natural extension of the existing art deposited in the office.
In other words, determining a patent comes down to searching the patents offices' own archive. If the old patent texts would be so lucky as to say: "To unlock the cellphone, the user’s finger moves across the touch screen", it won't match when Apple adds the word "continuous". The patent examiner is forced into a position where the least resistance is simply to grant Apple the patent.
This system is not working. The patent office archives do not have every iteration and written description about everything. By certainty, they won't have all the implied descriptions. As such, one should not assume that it is efficient to do an binary determination - 0 or 1, novel or obvious or not - on the basis of the consultation of the patent offices' own administrative records.
The simple change suggested by the article would only remove the ability for companies to brute force a claim phrase so it does not match exactly something in the patent offices' archives. It would not fix the patent system, but it might patch it up so exploiters can't run a brute force method like some script kids attacking web forms and see what string returns a SQL injection vulnerability.
The search and substantive examiners in the UK are one and the same (or were a few years ago) and the majority I knew of had doctorates.
Your example of mismatch according to a word/phrase is laughable. Novelty means novelty of the invention [defined by the claims], no matter how it's claims are worded, not novelty of the claims. If indeed USPTO examiners are searching only on specific word forms then I'd be amazed; that certainly isn't the case in Europe. You give an example of adding "continuous" - the problem here lies in that either that is anticipated or the objection is that it's obvious based on prior art. If an obviousness objection is made the applicant can claim that "if it's obvious why hasn't anyone done it already" - that then forces their hand to leave the analysis to the courts.
The classic UK example (quite old now) of the breadth of prior art is citing a children's comic against a "doggy doorbell". The breadth of related arts used in searches has vastly expanded since then.
In addition to that search and examination is divided up based on the IPC/ECLA classification systems (and additional subdivisions cast over those that are local) - which allows for specialisation.
How is that contrary to the core concept of the assumption that "Government knows everything"? If they do use specialist databases (source?), and subscribe to journals, it is still simply just more files collected in their computer system, which patent examiners use when searching key words/phrases located in the patent application.
A few months, there was a article here on HN about how patent writers carefully choose words in patent application just so key word searching Google and patent databases won't result in prior art. The industry that writes the large mass of patents knows how to exploit the mismatch according to a word/phrase system.
Europe patent offices might have a current better time dealing with this, since the brute force method illustrated in the article do not exist here. Like with SQL injections, its harder to do an exploit without the possibility to try until the claim string matches.
>which patent examiners use when searching key words/phrases located in the patent application. //
You said examiners only search patent documents in their computer systems. That's just wrong.
There are many commercial databases of technology - IBM's Technical Disclosure Bulletins (TDB) that they used for defensive publication for example or compendiums of papers from various scientific journals. All areas of technology have trade magazines that give details of current tech, enough to provide a suggestion of where to search in more detail. Of course the internet is a vast resource too
>how patent writers carefully choose words in patent application just so key word searching Google and patent databases won't result in prior art //
Can you link that. It sounds like nonsense. Claims use language such as "fixing means" to allow breadth to the legal monopoly defined encompassing all relevant/possible forms of fixation - glue, screws, bolts, magnetics, welding, flexible retention. Searches are made for the idea, not the language used to express it.
Again, it's not about string matching. If it were then it would be a simple computerised process.
Example of sources of non-patent disclosures - http://www.delphion.com/search-prior_art#tdb, http://www.questel.com/index.php/en/product-and-services/pri... [questel orbit], http://www.sciencedirect.com/. It used to be that the UK PO had console based access (eg telnet) to create search queries against these (and other) databases. Now of course the web is [often] used as the interface medium.
"The second technique to getting bad software patents issued is to use a thesaurus. Often, software patent applicants make up new terms to describe things with perfectly good, existing names. A lot of examiners will search for prior art using, well, search tools. They have to; no single patent examiner can possibly be aware of more than (rounding to nearest whole number) 0% of the prior art which might have invalidated the application."
While searches are made with the intention of finding the idea, no human made search engine can search for an concept. Search engines work on words. The semantic web tried to make a web searchable by concept and failed, and the closest second best is google (which only work for popular forms of the English language). Patent examiners could brute force and search for any synonym associated with each term and phrase, including those not making any sense in context, and use that to find prior art. Doubtful to the point of not a chance.
However, going back to core augment here about a system that assumes that the government knows everything, and lets apply the same logic to the justice system. The police do all the evidence gathering, so why do they not simple judge and decide verdict themselves. After all, if we assume they know everything there is to know, why should the state doubt itself in deciding the outcome? Why should the justice system ever doubt the government expertise, and do a costly and slow judgement, hearing all side, and then weight it all as if the government expertise might not always be 100% right all the time?
Is it enough that the police actually go and look outside the officer to collect evidence (like accessing outside databases or do internet searches)? Will that make them 100% right all the time?
That would fix pretty much all the nonsense of garbage software and business method patents overnight. It would eliminate nearly all patent trolling. And not a single patent that actually provides effective incentives to advances the arts, technologies, or sciences would be affected.
I believe patent applicants should have to argue how difficult it would be for a peer in the industry to reproduce their work.
If you develop an expensive new pill to cure a major disease, we'll give you a patent in exchange for you revealing the details of your important invention. Future generations will benefit.
If you put a one click purchase feature on your website, it may be a new idea, but there are tens of thousands capable of reproducing your "invention" on their own. It is a new idea, so grant a short patent (maybe 5 years?) but it hurts instead of helps society to grant a 20 year patent.
This is not only a silly example, but a bad one. If society never learns that it existed before it was lost, society didn't lose anything because it never had it to begin with.
If you can't remember the dream you don't know you had, did you actually forget it?
Said another way, I believe the purpose of patents is to get innovations documented rather than kept secret.
That is, should the quid pro quo of being granted the limited monopoly require the grantee to still provide for the public good by actually doing something with it, rather than just sitting on it or attempting to exploit those who are doing something with it?
So, the public grants you a monopoly on the exploitation of the invention for a limited time in exchange for making the invention public. But purely attacking the public, using the patent as a bat to beat others with, is not part of this exchange because the patent itself is not the invention. Really, shit-or-get-off-the-pot of patent granting.
In exchange for publicizing the invention, you get a limited monopoly on the exploitation of the invention, not on the exploitation of the patent.
And one click purchase is a terrible example to use, because there is no way to use such a thing as a trade secret, since it's a frontend innovation.
I am arguing that front end innovations, especially software innovations, shouldn't be patentable. Or at least grant them a shorter 5 year patent.
I don't believe patenting things that are easily reproducible, like software, benefits society.
- Get rid of license negotiation before the fact and replace it with attribution after the fact. All patented ideas should compete for a bounded slice of the pie of products. Attribution should be incentivized, but failure to attribute shouldn't be penalized excessively.
- Instead of just letting a panel of experts (patent examiners) determine which ideas are novel and which aren't, we should further let them determine how novel they are and award the most novel ideas bigger rewards. The current system of granting a monopoly on an idea allows the patent holder to collect a large potion of the total value of an idea, which would be fair only for ideas that occur to just one person in the world. In reality, almost all patented ideas (the good ones, anyway) occur to a number of bright people, even if they wouldn't occur to a person having ordinary skill in the art.
- The patent database should be more like the collection of peer reviewed academic knowledge. Ideas should be rewarded based on impact, rather than just who technically had the idea first. We should reward high quality exposition of ideas instead of our current system that mostly rewards obfuscation.
Of course, all of this is even less likely to happen that the idea in the OP. Abolishing patents or at least software patents would be preferable to the system we have now.
When someone goes to the government and asks for financial support, they don't get a "one size fit all" cash bag. It matters how great in need you are, and how great an idea you have for a youth center. But when it comes to get an government enforced monopoly, no argument is even needed. No one will even ask if 3 years would be enough, or what the extra 17 years are there for. It's 20 year or nothing.
We would have gladly taken a shorter term, like 8 years, in exchange for getting the patent granted faster. We didn't need to hold onto the stuff forever, but we needed some breathing room.
When they did work however there was no liability up until the patent becomes granted an unpublished submarine couldn't be used to make money until it surfaced. The idea was that you'd keep it hidden to allow the industry to make use of the invention and then when the invention was established you'd surface the patent and say "stop or pay up".
That's been [largely] fixed now with "early" publication (18 months I think it is in the USPTO).
It's actually even worse now since the US switched to first-to-file. Suppose you invent something and show it to your buddy, Fred. Fred can then turn around and patent your invention, and then sue you for using it. And unless you took the precaution of patenting it yourself or formally publishing it to establish prior art, Fred will win. The fact that you actually invented it doesn't matter.
There is no requirement for "formal publication". It needs to be disclosed in public, that could be a photo on your Facebook, showing a person in the pub with the public present, anything.
The fact that you invented it gives you absolute indemnity in the law and a public disclosure invalidates any subsequent patent application.
But as always you can't protect against being sued for anything.
No, this is not enough. You have to be able to prove in court of law that you disclosed it when you said you did, and that your disclosure was in fact a disclosure of the invention in question. So a blog post that says, "I just invented this cool widget" is not enough. Even a blog post with a detailed description may not be enough unless you have some way of proving that you did not go back and edit the post after the fact.
And in my scenario where you don't make a public disclosure but just show the thing to your friend Fred who then decides to stab you in the back, you will almost certainly have no recourse whatsoever. In theory, yes, the fact that you invented it and Fred didn't protects you. In practice, to avail yourself of this protection you have to prove that Fred didn't invent it independently, and that will be nearly impossible for you to prove if Fred is even a remotely competent backstabber. (And, of course, in actual practice it won't be Fred, it will be FredCo, Inc.)
I did point this out in the post but it's kinda implicit within the topic of law that courts require proof.
For the avoidance of doubt - any public disclosure which you can show beyond reasonable doubt to have been made prior to the priority date of the application may be used to invalidate the patent. [Jurisdictions vary but AFAIK this is now true both before the USPTO and EPO and member state's offices].
Is it really necessary to point out that it has to disclose the invention. Or that it has to be public within the legal definition. Do you want to go in to how "prior" should be interpreted?
Of course, there's the tremendous problem of the manipulability of what constitutes "an entity" - shell companies and what have you.
This would abolish the patent trolls as well as stop companies from doing preemptive filing because its "good practice".
But maybe that's a too simplistic view of the situation.
The thing here is that patent terms are too long for the rate of change of technology nowadays. About 7 years seems like a reasonable term to me.
An alternative to limiting years would be having a vastly increasing rate of increase on the patent maintenance fees per year for granted patents; particularly if it could be related to the revenue of any products that it is in. Then at least the public purse would benefit - from the monopoly the public are giving - in proportion to the benefit the private company is receiving..
in essence i feel they convince us that they have a good reason to not do their jobs - they can do this because they are great sales people - which makes sense because we select them by popularity contest.
remove patents is a simple solution for instance, its not hard to implement either and it will save running costs etc.
the consequences might be complex... i'd like to think it would just help everyone. its not like, e.g. legalising marijuana where you need to think about previous convictions. you really could just invalidate every patent overnight.
Otherwise, you're co-opting language about things that are precise and provable to make your subjective point.
</rant>
And par for the course, this article of his has several flaws as well:
1. Giving examiners the power of irreversible rejection is pretty bad, if you think about it. Valid patents are regularly rejected for the very same reasons that people argue invalid patents are allowed. Contrary to popular belief, if examiners are overworked and lack enough time, their default reaction is usually to reject the application. I have seen many a BS rejection issued because the examiner just wanted to meet his quota. And I'm not even a patent agent or lawyer!
2. (Nitpicking) Continuations are not the only way to continue examination after a rejection. You can also file for a Request for Continued Examination (RCE) and continue prosecuting the same application.
3. Continuations are a very valuable tool and not just fodder for abuse. Like TFA says, it allows applicants to claim multiple inventions off a single patent application. But that is often what happens! A single invention could have multiple facets that are inventions in their own right. As a random example, a lightbulb that uses tungsten filament and bulb with an inert gas could actually be three inventions: a) the material being used for the filament, b) the use of an inert gas in the bulb, and c) the combination of both that creates an even longer lasting bulb! What may happen is you file a patent for the combination but may realize each aspect is valuable by itself, and you may later want to claim them separately.
And sometimes, this is not in your control. A patent is allowed to only claim a single invention. An examiner may look at your claims and decide that they cover two (or more) separate inventions, and issue a "restriction", which essentially forces you to choose one invention to continue. If you still want to protect the other claims, you have to file a continuation.
And even further: Companies regularly dump hundreds of pages of technical specs (or a professor dumps a dozen papers) on to a patent lawyer and ask them to file on all the inventions covered in there. Sometimes this happens under a pretty tight deadlines (e.g. the product has already been on sale, or the papers published, for almost a year, which puts a bar on when a patent can be filed). In that case, lawyers write up one gigantic spec, slap on claims for whatever they think is the invention, and file it. Over time, they can then sit the inventors down and hash out the real invention and cover them in continuations.
4. I looked at the file wrapper for the Apple patent. "continuous" was not the only word added. Almost a third of the claim was amended to make it more precise of what they were claiming. I personally don't think this is a stellar patent, but such misinformation must be pointed out.
5. Strangely (or maybe not so strangely) enough, academics who write so much about patents have no idea how examination actually happens, or even how patents even work. To show invalidity, an examiner must show a one or more references that completely cover each and every element of the claim. If they cannot, no matter how obvious it seems to us, they have to allow it. The reason for this is that the decision must be an objective one, and all our opinions are inherently subjective. Supporting a rejection with previously published information, which is a recorded fact, is the only currently known way of doing this objectively.
6. Bessen implies other countries don't have patent litigation problems because their examiners can issue an absolute rejection. First of all, I highly doubt that is the case - there always avenues of appeal. Secondly, he ignores the legal environment in other countries, such as "loser pays" in EU, which tend to discourage lawsuits in general, not just frivolous ones. Thirdly, this makes no difference in the quality of patents - I have seen a ton of US patents and their international counterparts, and they pretty much all have the same claims. And they can all be just as each other. Nokia actually prevailed in some lawsuit in the EU over a patent that claims, without (much) exaggeration, pausing downloads when something more important is to be downloaded.
7. Rambus, as underhanded as their FRAND patent shenanigans were, was not a patent troll. By this standard, all fabless semiconductor companies would be patent trolls.
1. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2091210
2. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2117421
3. http://blog.patentology.com.au/2012/06/29-billion-us-troll-t...
4. http://gametimeip.com/2012/07/30/patent-scholars-challenge-b...
And yes, I realize that this submission directly contradicts my earlier vow not to submit direct links any more. What I didn't realize at the time I made that vow was that it directly contradicts explicit HN posting guidelines. I think that particular part of the guidelines is counterproductive, but as long as the guidelines are what they are I will respect them. But I will also keep pointing out why I think they should be changed.