The author is deeply confused. This part is simply incorrect:
See, “First to File” doesn’t require that you build
anything at all, that you have an intention to build
something, or that you’re getting a patent on something
that hasn’t, you know, already been invented by someone
who couldn’t afford the patent.
First to file makes no changes to any of that. It all remains exactly the same as it was under first to invent.
All that first to file changes is what happens when two (or more) separate inventors, who meet all of the statutory requirements for patentability (e.g., novelty, utility, non-obviousnous, reduction to practice and enablement), both file for patents.
Under first to invent, you try to figure out when each inventor actually started working on reducing the invention to practice (just having the idea is not enough). You find the earliest time T such that the inventor was working diligently on reduction to practice between T and the time of the patent application. T is that inventors invention date. Lowest date wins.
Under first to file, you look at the date on the application. Lowest date wins.
First to invent is problematic because it comes down to comparing two or more vague dates, and there can be a lot of subjectivity involved, such as deciding whether or not an inventor was working diligently or not toward reduction to practice.
As the author, I do appreciate that someone thinks I'm confused, but this sentence is hardly the point of my argument. My point is that large corporations are at an even larger advantage with this system and that it hurts entrepreneurs. What has already been a tilted playing field will become even more so.
The key with "First to Invent" is that someone could demonstrate that they had invented something but not actually patented it previously. For instance, if you came up with a novel software process, didn't patent it, and went about your business, it should theoretically be harder for another company to come along and patent your work. Since our patent system is broken, it hasn't always (maybe usually) worked that way. With "First to File" all you need to do is be first to file. And at a mere $15k to $50k to file, this is an undue burden on startups who are inventing things.
As the author, I do appreciate that someone thinks I'm confused, but this sentence is hardly the point of my argument. My point is that large corporations are at an even larger advantage with this system and that it hurts entrepreneurs. What has already been a tilted playing field will become even more so.
The key with "First to Invent" is that someone could demonstrate that they had invented something but not actually patented it previously. For instance, if you came up with a novel software process, didn't patent it, and went about your business, it should theoretically be harder for another company to come along and patent your work. Since our patent system is broken, it hasn't always (maybe usually) worked that way. With "First to File" all you need to do is be first to file. And at a mere $15k to $50k to file, this is an undue burden on startups who are inventing things.
edit: Because some people clearly don't want to understand the point here, there are several steps (lawyers, patent searches, etc.) that cost you money before you can file. This is included in the $15k to $50k I'm talking about.
>For instance, if you came up with a novel software process, didn't patent it, and went about your business, it should theoretically be harder for another company to come along and patent your work. Since our patent system is broken, it hasn't always (maybe usually) worked that way. With "First to File" all you need to do is be first to file. And at a mere $15k to $50k to file, this is an undue burden on startups who are inventing things.
Did you publish it? Was it put out into the public space? If so, it's prior art, and evidence that the patent should be invalid. This HAS NOT CHANGED.
If I make a piece of software, put it out for sale, then someone files a patent after my software has been out in the open, the patent is not new and novel. It should be thrown out.
Again, like tzs said, this only applies if two inventors invent something simultaneously. Under First to Invent, each inventor needs to prove that they thought of the invention before the other. Under first to file, the patent is granted to whoever patents it first.
The problem also applies if one inventor invents something, then another individual gains access to it before the inventor distributes it publicly, and files a patent on it.
But in the more general case, I find it difficult to believe that the notion of Prior Art is unaffected by the elimination of first-to-invent. Under the new rules, if the PTO considers two patents determined to cover the same invention, they will accept the one filed first and decline the one filed later, even if the first-filed invention was invented at a later date. If a patent lawsuit can reverse this decision by finding prior art, then how can these be reconciled?
Maybe the idea is simply to shift the burden of determining the earliest invention date from the PTO before the patent is filed to the courts after the fact?
>The problem also applies if one inventor invents something, then another individual gains access to it before the inventor distributes it publicly, and files a patent on it.
No. This would be fraud(because the "inventor" that filed didn't actually invent the thing) or could be theft of trade secrets(if an NDA is involved). Both of which are covered under US law.
>But in the more general case, I find it difficult to believe that the notion of Prior Art is unaffected by the elimination of first-to-invent. Under the new rules, if the PTO considers two patents determined to cover the same invention, they will accept the one filed first and decline the one filed later, even if the first-filed invention was invented at a later date. If a patent lawsuit can reverse this decision by finding prior art, then how can these be reconciled?
If there is clear prior art before either were filed, then the patent would be nullfiied in either a First to File or First to Invent scenario.
I think that you're confusing prior art and what first to file/invent means.
First to File and First to Invent means that if two people file for an identical patent, the patent is granted based on either the filing date or the invention date. The only time that First to File applies is when two independent inventors file for the same patent in the same timeframe. Ovbiously, one cannot try to patent an idea after the patent has been accepted, even if they did invent first.
Prior art applies to all patents, based on the date of the filing of the patents. If I try to patent something that someone else is selling, like a mousetrap design, then it should get thrown out, no matter when I actually invented it, because it's already in production by someone else.
> For instance, if you came up with a novel software process, didn't patent it, and went about your business, it should theoretically be harder for another company to come along and patent your work.
First to file doesn't change this. Prior art is still prior art.
This is slightly simplified, but here's the gist. The only change is when Inventor A comes up with an idea first, Inventor B comes up with the same idea (independently!) later, and they both file for patents, but Inventor B has the earlier filing date. Under the old law, Inventor A wins; under the new law, Inventor B wins.
If you're not in that situation, first to file doesn't change things.
That's ridiculous! Inventor A should obviously always win, since he actually implemented his idea before B did. B could have seen A's idea and filed a patent on it.
I'm not sure that's obvious. (And by the way, we're talking about conception of the idea, not actual implementation. You can patent something that you haven't yet prototyped.) Anyway, the entire rest of the world uses a first to file system, so bear that in mind when you say that first to invent is obviously correct.
Just to make sure no one gets the wrong idea, it should be noted that the law requires:
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,
and shall set forth the best mode contemplated by the inventor
of carrying out his invention.
(35 USC 112). So, even though filing counts as constructive reduction to practice, that doesn't mean you can just come up with a vague idea, write it up with no clue how to actually make the device, and get a patent. (Well, you aren't supposed to be able to--the examiners sometimes slip up!).
You don't actually have to have built a working prototype, but you have to be telling people in the patent how they can actually build the thing.
The examiners often slip up. That is really the nub of the problem. The examiners are incentivized to close cases, not to make the right decisions. So if you badger an examiner enough you can get them to sign off on just about anything, e.g. U.S. patents 6368227 and 7126691.
He said A came up with the idea first, not that he implemented it first. You could have this timeline:
A comes up with the idea, starts working on implementing
B later, independently comes up with the idea, and start
working on implementing
B completes implementations, files for patent
A later completely implementation, files for patent
Under first to invent, A wins even though B implemented first. Under first to file, B wins.
> B could have seen A's idea and filed a patent on it
A could challenge B's patent on the grounds that B was not an inventor.
Thanks for your detailed response! If that's truly the case, then I could support this move to first-to-file.
One worry I have is that the order might often be more like:
A comes up with the idea, starts working on implementing
B later, independently comes up with the idea, and starts
working on implementing
B files patent application describing a grossly simplistic,
untested implementation of the idea
A later completes implementation, irons out the bugs,
files for a patent, starts a business, adds great value to
the economy
B never really does much with his patent, ends up selling
it off to C who sues A for mucho $$$ in patent licensing,
or sells it to A's competitor D for mucho $$$, so they can
copy A's product with impunity.
If that scenario actually happened, then either A's patent application would contain claims that are not in B's application (and in all likelihood, A's dependent claims would prevent B from completing their product), or A was just plain stupid about waiting too long to file or publish. (The other option is that B gets granted an overly-broad patent, but that's a completely different issue.)
Is this situation really better under the current law though? If you change the situation so that B had the idea first, but the rest of the facts are the same, then the current law produces the same problems.
The reason I suspect my situation is more common than the one you describe, if the first two lines are chronologically reversed, is that the time it takes A to go from idea to practical product and filed patent is assumed to be greater than the time it takes B to go from idea to filed patent, since B didn't wait until he had a marketable product before filing. This new system gives B a corresponding advantage over A.
Perhaps another assumption is that by the time the seeds for an idea are drifting in the ether, A would be more attuned to receive them, and would have started before B.
Though of course the situation you describe should be addressed, this change in policy may be an additional check against abuse that we are losing.
> With "First to File" all you need to do is be first to file.
No. You have to invent, and then you can file. You seem to think that "first to file" removes the requirement for invention. It does not. It just changes the priority determination in the case where two independent inventors have both invented overlapping inventions.
It's not very hard to "invent" things if your sole reason for being is to copy entrepreneurs' work before they can afford to file, before they can afford to hire a lawyer, pay the USPTO an $8800 fee to ask for a review, and you've got only a short window (which you probably won't know about) to do so.
Look, we need real patent reform and this (even if you think it's only marginal) tilts things yet again in the wrong direction.
This is a small thing, but filing fees (for a small entity) are $462. The fact that you think USPTO fees are this much indicates to me that you have never had anything to do with an actual patent application, have no real idea what you're talking about, and are basically making shit up. But then, your article conveyed that pretty clearly as well.
Wow. Usually we try to keep things a little more civil here. This filing fee is the smallest portion of what you need to pay in order to file a patent. First, you need to get a patent attorney to help you put your filing together. I don't know what lawyers you use, but this can typically run more than $15k and up to do it right.
Edit: Since I don't seem to be able to reply to your comment, thank you for your sort of apology. You can read at length about this particular bill here: http://en.wikipedia.org/wiki/America_Invents_Act. It's not all sunshine and equivalencies.
Perhaps we can have a discussion about it some time, but as someone who has both been in startups that have pursued patents and been in a company sued by a patent troll, you're just going to have to accept that we have different views on the issue.
Sorry, not really trying to be a dick. But your article reads like you're trying to stir up the (already considerable) patent hatred around here, without really understanding the issues.
And yes, I know the filing fee is insubstantial compared to the legal fees. The point was that you should know that too, if you're expounding on patent law and policy.
The problem lies in the fact that an inventor no longer has a recourse to contest a patent lawsuit against him by showing evidence that the patented idea was implemented prior to the patent filing date (or before the determined invention date). This exposes all inventors to having their ideas copied and patented by others, drastically increasing the incentive to patent their own ideas on a continual basis. This has the effect of significantly increasing the total number of patents being filed.
The earlier invention can still be used as prior art against the later filing, assuming there was some publication or public use of it. See my comment here: http://news.ycombinator.com/item?id=2834229
It's true that the system was broken before and will remain broken after this reform, but I don't see how that makes the author confused.
One might reasonably hope that a reform bill would contain more reform, after all, and one might legitimately criticize a reform bill for failing at its purpose.
Entities like China that are less than assiduous in respecting patents are going to benefit enormously from the hobbling effect of U.S. patent litigation. I don't think it's a complete counterbalance to the lack of free information flow which generates opportunity for corruption, but we really shouldn't do this to ourselves.
I don't see how this isn't a bigger issue or hasn't been brought to light. As I understand it, large corporations (or anyone with the money, really) can simply file patents for future products they predict will come to market and then sue startups and entrepreneurs when they create the actual product.
The original idea of patents was along the lines: Person A develops a self propelled buggy moving with the speed faster than light. Without patent protection, there is a lot of incentive to keep the details of the technology secret. The patent protection would be awarded in exchange of him disclosing the details of the technology, thus enriching the human race's body of knowledge and furthering the technology advancement.
Modern day patent system - a Corporation A files a patent for the idea of a "self propelled buggy moving with the speed faster than light." Whoever comes with real implementation later, would be forced to pay license fee to the corporation or wouldn't be able to bring it to market. This is pretty much the innovation tax (though if a tax isn't extorted by a government then it is more correctly called extortion/racket).
In theory, a patent is supposed to be specific enough that the patent itself accurately describes the object, to the extent that a reasonably knowledgeable person could actually manufacture the object using only the patent as a blueprint. In other words, you would not be able to write the patent application unless you had actually invented the object in question, because otherwise you wouldn't know how to describe the theoretical object well enough for a person to manufacture it. Therefore you cannot patent imaginary future inventions.
In practice, this rule is effective only to the extent that the Patent Office enforces it.
As you point out, the issue is that the UPO doesn't enforce that as rigidly as it should (or maybe, is reasonable to expect). As tjr pointed out in a separate thread, 'This American Life' did a great piece on software patents specifically, definitely worth a look if some of you are interested in learning more. Certainly gave me better perspective.
http://www.fsf.org/blogs/community/tal-when-patents-attack
The key point is that "first to file" doesn't actually honor the idea of invention. It honors the idea of filing the patent first. So, now it's a race to file instead of invent. Perhaps the patent system has behaved this way for some time, but it's codified now.
From wikipedia: "In a first-to-file system, also called "first inventor to file" system, the right to the grant of a patent for a given invention lies with the first person (the first inventor(s)) to file a patent application for protection of that invention, regardless of the date of actual invention."
You aren't entitled to a patent unless your disclosure is specific and detailed enough to be enabling. If it's enabling (and novel, nonobvious, etc.), then why shouldn't you get a patent on it?
On the plus side, it seems like only a small matter of time before this obviously broken system collapses under its own weight, and we're forced to do something slightly more sane instead.
Once the ensuing tidal wave of patent applications is upon us, the USPTO will have no choice but to drastically lower their (already depressingly low) standards for patent acceptance due to lack of resources available to check them, and the growing portion of corporate profits evaporating to legal costs will put pressure on some actual patent reform.
Personally, I am angered that inventors will lose a recourse to challenge individuals who file patents on their previous inventions, and I in no way condone breaking the system for honest businessmen in order to make a case for fixing it, but at least I can take some solace in the fact that it will be a quick demise, unless I'm missing something.
> Once the ensuing tidal wave of patent applications is upon us, the USPTO will have no choice but to drastically lower their (already depressingly low) standards for patent acceptance
The US patent system will collapse under its own weight, about when the US health care system collapses under its own weight, the US education system collapses under its own weight, and while we're at it, the entire US collapses under its own weight.
I can imagine that but I don't think we should, uh wait, till then, considering doomed systems show a distressing ability to just keep going when people do nothing but wait for them to collapse...
Now if they would combine expiration date with file date we'd be done.
The reasoning is fairly simple, patents expire and the idea becomes public domain, that is why, starting in 2015 as the great patent binge of late 90's starts expiring, through 2025, all of that 'crap' that you see as being patented will become public domain. Unlike copyright there will be no restriction on you using it. (like you can write your own RSA encryption code now without fear of being sued)
Also during that time there has been some progress made on getting better examiners into the office since we're starting to see examiners who actually used computers in high school versus the ones in 1995 who had just heard about them and never actually used one.
So you file as soon as you can, and if its a spurious filing you have exactly 20 years (if we could have them expire on file+20) to make the product work and to ship it with patent protection. Alternatively, we go back to a requirement that the invention must have an exemplar which the PTO can examine with the application. That however was originally eliminated because it was slowing everything down. (although some of the model work is simply stunning, check out the museum some time).
In the mean time I'm gonna watch all my old Star Trek episodes and file patents on all that stuff :-)
I think we have generally mixed up "value" and "cost".
Patents are (or should be?) a legal construct providing a temporary monopoly on an invention to incentivize investment into (costly) research. I.e. protect the "cost" of an invention.
Cost, however, is not necessarily directly related to value.
Patents (imho) should not protect "value". If you have a great idea one morning, that might indeed be very valuable.
But you would have had that idea regardless of whether its patentable or not, and society as a whole is not better served by granting monopolies on such ideas.
Instead the market decides who builds the best product.
Other IP protections such as Copyright are still crucial for this obviously.
In the end I think that (case) law usually reflects general trends in society.
This is not a society of altruism (anymore) but a society of the mantra "take the money and run". Once or if that changes the law will eventually match that.
One way out of this mess is to document, publicly, every idea - however insignificant it might be. That way there will be prior art to everything that is obvious.
What is the state of patent law, especially regarding software patents, across the world? If patents really are going to be a murder on innovation, then the least broken system should enjoy an advantage, no?
Keep in mind that the standard for "First to File" is actually present in much of the rest of the world. Large US corporations have argued that this switch will make our companies more competitive. What's happened is that American companies see a unified global patent system eventually taking hold and they want to use their resource advantage to dominate a future marketplace. They're primarily concerned with establishing market leadership lock in. Innovation as a desired outcome for the advancement of mankind couldn't be further from their thought process.
I've often wondered what's the commercial and political mechanism that controls patent legislation and the lifetime of a patent. I mean, we're still at a mere 17 years with patents but Mickey Mouse is at roughly 95 years by now. Why is it worth lobbying for copyright extensions but not patent lifetime extensions?
Just a guess, but doing what you describe to patent law would seriously and obviously harm the economy and isn't politically feasible, whereas insane copyright terms just give the assholes at Disney a license to print money, some of which finds its way back to the very corrupt bastards who extended the term in the first place.
Put another way, copyright as it is currently conceived degrades the creative space, but it isn't directly taking food out of people's mouths (it is, but in a less obvious way, and no one gives a shit about artists anyway).
It also means everyone knows patents are bullshit and collectively the companies don't want any of it but since we've got it established already and it's really hard to shut down the companies know they need to play ball yet do nothing in favor of patents.
> America’s most quintessential inventor did without them and thought they were foolish.
Right, but he also didn't care about selling his inventions to feed his family, having other sources of income. It's easy to give away something you don't need.
All that first to file changes is what happens when two (or more) separate inventors, who meet all of the statutory requirements for patentability (e.g., novelty, utility, non-obviousnous, reduction to practice and enablement), both file for patents.
Under first to invent, you try to figure out when each inventor actually started working on reducing the invention to practice (just having the idea is not enough). You find the earliest time T such that the inventor was working diligently on reduction to practice between T and the time of the patent application. T is that inventors invention date. Lowest date wins.
Under first to file, you look at the date on the application. Lowest date wins.
First to invent is problematic because it comes down to comparing two or more vague dates, and there can be a lot of subjectivity involved, such as deciding whether or not an inventor was working diligently or not toward reduction to practice.