Hacker News new | comments | show | ask | jobs | submit login

This John Carmack quote is relevant and shares your sentiments:

"The idea that I can be presented with a problem, set out to logically solve it with the tools at hand, and wind up with a program that could not be legally used because someone else followed the same logical steps some years ago and filed for a patent on it is horrifying."


Except it's even worse than that: usually patent trolls haven't even written the code to solve the problem. Working code is not required to be granted a patent.

And not only you might accidentally break patents - you might not even be able to figure out if you break patents at all without paying a lawyer to translate the patent specifications. Maybe it's just because I'm no native english speaker, but I often don't get patent texts. Coding around patents is like walking blindfolded through a minefield.

I'm a native English speaker and I don't generally understand patent texts. In fact, I've filed quite a few patents and have always been amazed that I give something readable to a lawyer to write up and they turn it into something that I can't understand any more. The idea that a patent is the medium through which an author shares their invention with the world is rather perverse.

So I agree; coding around patents is difficult given that they tend to be so hard to read and understand.

I have seen quite a few people who have filed patents state that they don't understand their own patents and (as a general statement) I find this puzzling. A patent has 3 main sections (a) background and prior art (b) description of the invention and (c) the claims.

The claims in particular are indeed technical and difficult for a lay person to readily understand. But the background, the prior art and the description of the implementation are not. Background and prior art are a summary of the technical context and I have never had any difficulty reading or understanding this section. The description may be rather long winded as it has to detail how to implement the invention so that it may be implemented from the description by someone skilled in the art, but it isn't legalese and should certainly be understood by the inventor! The lawyer can't write this by him/herself.

I can certainly sympathize with those who can't follow the claims (although I have found it to be perfectly possible with determined effort), but I simply fail to understand how a lawyer turns it into something "I can't understand any more." Find alien and unfamiliar - yes. But incomprehensible?

Yet the only things that matters are the claims.

You have to understand, though, that legalese isn't English. It may look like English, and even work like English in ways, but its not English. Saying that legalese and English ought to be equivalent because they share words and syntax is like saying C++ and Java ought to be equivalent because they share keywords and syntax.

For patents in particular, though, a big part of the justification is that they constitute a disclosure of the invention to other practitioners in the field. It's a quid pro quo: we grant you a limited term of protection, in return for you divulging the details.

As part of that arrangement, it seems like we ought to require that the disclosure be in English, not merely in a tongue that shares some English words and syntax.

English? Hah. I'm against the whole patent charade but if we're going to grant them, the inventors should be forced to disclose the entire goddamn source code, development infrastructure, and version control repository which proves that they've actually invented something and not just handwaved in the direction of an idea.

This is presently true, but shouldn't it be possible to define a legalese that can be parsed as English as well?

Yes, but the fact that it's impossible is considered a feature, not a bug, by those who use it -- it keeps the bastards employed.

That's one way to look at it.

Alternatively, you could say that the point of "legalese" is to define and convey ideas more explicitly than they can be defined and conveyed in English. Try writing a contract in "Plain English" - it will necessarily be full of ambiguity and uncertainty, which defeats the purpose of having a written contract.

Similarly, although programmers don't write code in "Plain English" I don't think we should begrudge them the right to keep using obfuscatory programming languages - even if it keeps the bastards employed.

The argument for specific languages for programmers is stronger because programs are intended to be interpreted by machines. I would be perfectly happy with a legalese that is opaque to the layman if it could be fully parsed and implemented by a machine.

I also believe that, by defining one's terms clearly, it should be possible to write unambiguous plain English contracts. Additionally, an English-parsable legalese needn't necessarily lead to an unambiguous interpretation when parsed as English, so long as it is unambiguous when parsed as legalese, and most of the meaning is retained when parsed as English.

I'm guilty too. I have some patents granted and I don't even know what it is about base on the description written by lawyers.

One could argue that you shouldn't code around patents in the first go round. Because if you demonstrate that you knew about the patent and then proceeded to infringe on it, you could be liable for 3X the amount.

If you don't know about the patent, then you're only liable for 1X the amount.

Moral of the story: don't read patents, you can only get into trouble. If you get sued after the fact, then you can change your code to work around it.

Yup. The quasi-official policy at some places I've worked is that engineers are not to browse the patent database, for exactly that reason.

Ironic, given one of the original primary reasons for the patent system.

I'm not sure if that really lowers your overall risk. Finding out afterward that you accidentally broke some patent which you could have avoided easily isn't necessarily better even if you don't have to pay treble damages.

And you get fined triple if you're looking for mines.

And they could file for the patent after you have done the work. You then have the hassle and expense of proving that your work came first and even then such cases often don't go the way they should.

I great example of the trouble someone had to go to prove they were first is the ABC computer. It's an important case if you consider that without John Atanasoff the personal computer would never have existed. At the time IBM and some other corp I can't remember the name held some important patent on memory and other things that would prevent anyone from creating a personal computer without paying IBM. Those patents which had come from the creators of ENIAC were later found to have been heavely (well in reality stolen) from the ABC computer. After years of battles, IBM lost the patents and right after that the PC came to be.

For a worse horror story, just take a look at Farnsworth trouble with RCA about the TV patents.

You could still lose that fight, even if your work came first, if you never published it or filed a patent for it. First to invent only holds if you reduced it to practice (either by actually inventing the thing, or boiling it down so someone reasonably skilled in the art could then invent it) within that year. You lose rights to it after that if someone else wants to patent it. There's also first to conceive stipulations. The intellectual property class I took last semester would have been a lot easier if the U.S. just went first to file (like basically everywhere else).

The US is trying to make that transition. The bill passed the Senate, and Committee in the House, and goes to a full House vote in June 2011. But it will likely be difficult for the two bills to match and be agreed upon because of all the other crap in the bill which has good stuff and bad (much like the overwrought SBIR reauthorization bill that can never seem to pass and it just extended ad infinitum). I like the "first to file" idea, but it puts the small inventor at a disadvantage. The main benefit is that it will hopefully lubricate university tech transfer offices. No more "oh I'll put in a disclosure and wait and see, I'm safe." it'll incentivize the inventors to file a provisional as soon as possible. But of you're building (or hacking) something in your garage you'd better get a patent attorney quick. And be patient, because response times on a first-to-file system will be horribly slow even compared to today's glacial system.

In theory that shouldn't be patentable. What John Carmack is describing is something that is obvious to a craftsperson in the field.

OTOH it does still preclude John from being presented with a problem and then doing someone quite ingenious to do a great solution that seems actually quite "novel", to learn that someone else did it years before.

But for patents, if you follow logical steps from a given problem then you shouldn't be granted a patent.

In fact the way patents should work is that the patent officer spends a week trying to solve said problem themselves. And only after that attempt, read the actual patent. If they're struck with, "Ahh... yes, good idea, wouldn't have thought of it mysefl" then its potentially patentable. But if the patent officer says, "Yeah, that was my second attempt." or "that's largely derivative of my general line of logic in my other attempts" then it shoudln't be patentable.

> In fact the way patents should work is that the patent officer spends a week trying to solve said problem themselves.

Does that really sound at all scalable (or even reasonable) to you?

I imagine that many of the proposals that move through the patent office are highly contextual and may be simple ideas wrapped up in daunting language. It seems reasonable that it would take far more than a week to even get your arms around some of the problems themselves, let alone the solutions.

Besides, if patent officers were such able critics, wouldn't they be more useful as engineers?

There's a very nice counter argument to the objection of obviousness: if a given invention was obvious to the skilled man at the time of the application, and the invention is useful then why can't you find it in the prior art?

Basically the argument, which I find quite compelling on one level, is that truly obvious stuff is known and citable. Very simply novel inventions can seem obvious once they are presented to you. I found this a lot examining patents - they would often seem obvious but nonetheless be absent entirely from the prior art.

However, to answer some of the proposed solutions there is also the aspect of presentation of the invention to the public domain. That is, even if it were possible for an engineer ("the notional skilled man in the art") to create a close enough solution if they are set to work on a problem this does not provide a reason not to provide some form of patent protection.

The deal is that for disclosure of a workable, industrially applicable invention that a limited monopoly is granted. The monopoly does not only reward but also encourages. Whether others could solve a problem is not necessarily relevant to rewarding those who have addressed it and solved it. Do you see. You want to motivate progress in the relevant arts by rewarding those who solve problems in those arts by the sweat of their brow.

Re ability of patent officers: a judge can watch ice skating and tell you if the people fall over, if their toe-step is dramatic interpretation of the music or them tripping, if their Salchow is wobbly, if their outfits breech regulations, etc., but possibly barely ice skate themselves.

That said in the UK at least there is a good deal of art-specific expertise so much so that the patent office makes a substantial income hiring their examiners out to large companies to do "commercial searches". Mind you when we assessed international searches for the US ones were nearly always useless and had to be re-done (WO patents entering the national phase would sometimes have pre-existing examinations by foreign examiners, usually US or EPO).

>"more useful as engineers"

I don't think people stop being engineers because they become patent officers. That said they probably would be more use to the human race using their skills in a more directly constructive way but this assumes a lot about the world that just isn't going to happen. When there's no greed and everyone works for the greater good then we'll have no need for patent officers.

A scalable proposal:

When a suit is filed, an engineer is given the problem and told to build clean-room implementations. If he gets it in under a month, it's obvious and overturned. If he gets it in a day, it's extremely obvious and the plaintiff is fined three times the asking damages. If he gets it in under an hour, thirty times.

This disincentivizes completely bogus patent claims, and should bring the total rate down to something manageable.


When the proposal is submitted, the engineer attempts the recreation. If he fails, the patent office pays him. If he succeeds, the attempted patenter pays for a month of his salary. If he succeeds in under a day, the patenter pays a year.

This gets up-front examination, but weakens the incentive to hire good engineers and doesn't guarantee to make trolling unprofitable.

Unfortunately, this doesn't work if there's any possibility that the engineer has seen the patent (not terribly likely in itself) or been influenced by it indirectly (much more likely: maybe she's talked to a friend who read a web page on the subject, written by someone who had a friend at the company that produced the patent).

Once something is patented, the idea is out there for others to hear about (not just because they can read the patent; the original inventors can now talk about it, too). And ideas that get out into the wild tend to spread.

On the other hand, after all these years of having patents on obvious things granted, it might be no bad thing to try throwing out some genuinely inventive patents instead for a change :-).

(My own preference: either do away with software patents completely, or substantially shorten their term. [EDITED to add: the latter probably only makes sense if you can greatly reduce the latency between initial filing and granting. Which would be a good thing anyway, for all patents.])

> In fact the way patents should work is that the patent officer spends a week trying to solve said problem themselves. Does that really sound at all scalable (or even reasonable) to you?

I imagine that many of the proposals that move through the patent office are highly contextual and may be simple ideas wrapped up in daunting language. It seems reasonable that it would take far more than a week to even get your arms around some of the problems themselves, let alone the solutions.

Then spend more than a week on it. I didn't use a week as "cap", but rather a starting point. If it takes a month, sure then spend a month. You'll get no arguments from me. Especially since patents regularly take 5+ years to get approved.

Either you're trolling, or you're unaware that there were close to 500,000 patent applications filed in 2009 alone[1]. Let's say you have 500 able critics; that puts us at 1,000 patents to review per officer per year. Let's assume any given officer can review an application within one week with reasonable efficacy; it'll take ~20 years to review one year's worth of applications (1000 / 50; I excluded two weeks for vacation).


Not trying to be an ass... But I would feel stupid if I didn't correct you. I'm sure it was just a typo though.

A year has 52 weeks (plus one or two days), if 2 of these weeks are vacations, 50 weeks remain. Therefore 1000 / 50 = 20.


The obvious rhetorical question here is: Who seriously believes that there have been 500,000 inventions worthy of being patented in 2009?

The sheer volume makes it clear that the vast majority of patent applications are for trivial stuff. The system is broken, it's that simple.

Can you imagine there is one new invention that pertains to your job each year? Now consider the number of different jobs there are. Now consider the vast number of other activities outside of work.

Or think of it this way, a mobile phone: it could have novel programming, novel chips, novel screen design, hinge, case plastic, button return springs that last longer, easier to use touchscreen, stronger glass, more flexible, brighter, etc., any of these things might give it the market advantage. How many other things do you use each day, each of them could potentially be improved and that improvement be registered in a patent.

The parent spoke of applications for patents anyway. Granted patents in the US are apparently about 40% of this.

Your position is not consistent or just poorly thought out, and poorly researched. Or maybe I'm just a troll.

> In fact the way patents should work is that the patent officer spends a week trying to solve said problem themselves. Does that really sound at all scalable (or even reasonable) to you?

So clearly you were saying that a week is too much time to spend on a patent. OK.

I imagine that many of the proposals that move through the patent office are highly contextual and may be simple ideas wrapped up in daunting language. It seems reasonable that it would take far more than a week to even get your arms around some of the problems themselves, let alone the solutions.

But in a week you can't understand the problem, much less a solution. So you seem to be saying that having a system where examiners approve patents they don't understand is preferred. I say that if the examiner can't understand the solution, much less the problem, then the patent is rejected. This will really get people to focus on writing clear patents, rather than the legalese that passes for a patent today.

But lets get to numbers. There are not 500 examiners. There are more than 6,000. So at one week per patent it's now less than two years per patent. That's actually about on par with the time today. That's extremely reaonable. And I'd argue that making patents more expensive, with less protection time, you'd reduce the number of SW patents, and thus probably can turnaround SW patents more quickly.

Regardless of whether or not it is feasible to have 6,000 patent examiners perform this process, it still seems remarkably subjective: a given patent examiner, with all of his/her individual strengths, weaknesses, imperfections and idiosyncrasies, determines if s/he alone can solve the proposed problem and, if not, essentially says, "Aw shucks, this is patentable!"

And to me that is one of the many problems with (software) patents: it relies on an inherently biased, subjective process to determine whether this objective, rigorous piece of thought is worthy of a monopoly.

If subjectiveness in judging is your concern, I think there are bigger fish to fry than patents.

And also recall with patents you are guaranteed another appeal to a different body if ever sued, the courts, where you can invalidate a patent. And if MS wins the i4i case the burden of proof will be even lower. And note, there is legislation going through congress right now to make it harder to win a suit, and also expands patent reexamination.

My point? The patent examiner is an important first step, but a patent that clearly has no legs should be defeatable -- if not now, in the near future.

You guys are funny.

The conclusion that you should be reaching is software patents are simply not a viable mechanism, period. In other words, even if you like the idea of patents in theory, in practice they do not work because we will never have enough patent examiners with the necessary expertise much less time to make valid decisions on patent applications.

> So you seem to be saying that having a system where examiners approve patents they don't understand is preferred.

Not so. I'm not defending the current system, just trying to demonstrate that your suggestion isn't feasible.

> There are not 500 examiners. There are more than 6,000.



Edit: found the data on Wikipedia: http://en.wikipedia.org/wiki/United_States_Patent_and_Tradem...

All the people who are saying evaluating obviousness of an invention is impractically hard to analyze, are failing to see that that applies to non-software patents as well.

The real problem with software patents is not that the rules (including non-obviousness constraint) are flawed, but rather that the execution of those rules is flawed. Fix the right thing please.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact