Patents protect a process, composition of material or apparatus. Most awarded software patents show apparatus and/or process, and if they don't, a challenge will usually crush them.
First of all, I doubt that all program claims can be easily crushed when challenged. Unless of course you take a wording as "a computer program product, characterised in that when loaded in a programmable device a process according to claim X is executed" as a description of a composition of material (the program could be on a CD) a process (although the claim that program claim refers to is the actual process claim).
Next, such a challenge easily costs US$1.5-2 million.
Finally, it is quite trivial to word abstract software ideas in a way that they become a process executed by a computer (after all, that's the only useful thing you can do with a computer program). So that's not really a limitation to what you can patent.
The problem is that in case of software, the product is simply another way of stating the description. It's similar to saying that you don't want to allow patents on stories, but you do want patents on printed books, because a book is a product.
And then you get patent descriptions like "an information carrier characterised in that it contains a story involving a boy and a girl who both have an affair, whereby...". It's not all that different from how program claims in software patents look.
In the end, one could even start arguing for patents on patent descriptions themselves: just like a computer program contains instructions that tell you what to do, a patent description for any physical invention also contains a description of such steps.
Patents on information entities simply don't work, because then the system deadlocks: patents are supposed to encourage sharing of information, but software patents forbid that same sharing if you use a particular class languages (programming languages) to express it. And the reason is simply that the information is the product.
That's also exactly why we have copyright and why it's appropriate for the protection of computer programs: there the way how you express something is what matters and what is monopolised, and not what you express.
It's true that copyright protects more than the literal text (which is a nice counter-argument against software patent proponents who claim otherwise, but that's off topic here:). It does not protect ideas however.
Rowling and her lawyers cannot claim in their cases that they possess the idea of some wizard kid and the things he does, but only on the way Rowling expressed it. If someone else bases his expression on Rowling's expression, then this is a violation of copyright. Not because he used the same idea, but because he used her expression of that idea as basis for his work. There is a substantial difference there, and then you don't have pure original work anymore (but something based on her work).
Where the line has to be drawn between original work and plagiarism is for the courts to be decided on a case by case basis. The fact that there is *also* some original work, is irrelevant. Otherwise, I could just take a Harry Potter book, change all names, switch the word order of some sentences, add a few new sentences here and there, possibly add a chapter of my own and then sell it. Or maybe I could just create a translation of the work and sell it.
You're right, that must happen before. But you simply have to do a spectral analysis and then quantise the coefficients to infringe (regardless of how you do this). That's not really limiting their monopoly either...
If the patent even partly applies to sender ID, you're screwed nevertheless. And it's quite possible that the patent office will consider the combination of those two things non-obvious as well, resulting in another patent on the whole thing.
That said, I have no idea whether their called-ID patent applies nor whether they already have another patent (pending) on the combination of SPF and called-ID.
The base MP3 patent is much broader than just what is needed "to get back some of the money they spent on research". They patented compressing (using any kind of algorithm) sound, as long as you do it in a loop that involves an entropic encoder (such as Huffman encoding), until the sample reaches the desired bitrate and if you store the result together with the bitrate afterwards (or send it over a network).
Dude, he was arguing against patenting languages and compilers in order to enforce the "no patenting stuff you write using this" licence...
Yes, but that's not what I was replying to. Facekhan said that people should be forced to protect the software they write only by copyright, and not by patents, and that this should be enforced somehow by the language creators. He did not say anything at all about using patents for this.
The AC replied that enforcing this would be only possible through patenting and licensing of the programming language, and then said that such a clause would impose a restriction on people's abilities to make money with the software they wrote (because they would not be able to get a patent on what they wrote, since that would go against the license under which the language is used).
So he was claiming that without software patents, you can't make money on the software you write.
GPL keeps the code (and as such the information) free, BSD keeps the individuals free to do what they want to do. It's simply different kinds of freedom. Get over it.
This is not about patents, it's about copyright. And the fact that you give something away for free, does not mean you should be allowed to infringe on copyright (otherwise e.g. Microsoft would be allowed to infringe on the GPL by incorporating GPL'd code in a closed source Internet Explorer, back when they were giving it away for free).
notice how the interviewer seems to think that somehow Apple Hardware and PC Hardware are different?
Mac and PC video cards with the same model number actually have been different already several times in the past (e.g. different core/memory speeds). But please don't let this information influence your view of Mac users.
Have you ever developed anything you wished to profit on?
Please stop spreading the FUD that you can't make money on software if you don't patent the hell out of it. Ever heard of Opera Software? They're one of the most vehement high-profile opponents of software patents in Europe.
Besides, software patents allow others to steal your profits, which you made by selling a program you wrote all by yourself. It's really great, you spend several man years on writing a program, bring it to market and then get a ton of nice requests for royalties because auto-updating, RLE compression, plug-ins etc are all patented.
Or maybe one of those patent holders has a similar program himself, and doesn't even offer you a license. Tough luck.
It's not a program, it's a protocol, exactly like the TCP/IP and HTTP you mention. You can both write proprietary and Free Software that uses it. Which license some implementation from Apple uses is irrelevant (or do you consider HTTP to be non-Free as well because Internet Explorer and IIS use it?).
The claims are not what the invention is, but indeed what the "inventor" wants a monopoly on in return for the publication of the description of how his "invention" works. As such, claims are always broader than the invention itself (the reasoning is that otherwise, someone can get around the inventor's patent by just changing one small detail of the invention).
With non-software patents (i.e., where the invention/innovation lies in a novel way of using physical forces/material), how far exactly you are allowed to abstract is mainly limited by three things:
You can't abstract the claims until only the forces of nature you are harnessing remain, because those are not patentable;
You can't abstract the forces of nature you harness out and remain with some generic algorithm/method that could apply to anything, because then your invention (novel way to use those forces) is no longer part of the claims;
You're of course also limited by prior art (you invent a new car, but other cars already exist -> you can't claim all 4-wheeled vehicles etc) and whatever the patent office deems too general (after all, society grants you a monopoly in return for disclosure of an invention, so those two should -in theory- be proportionate).
However, if you look at software patents, then
There are no unpatentable basic "forces of information";
Since what you start with is already some abstract method/algorithm, no matter how much you abstract it further, you can always argue that your invention is still embodied in those claims;
This one is the only thing left.
The net result is indeed that you end up with ridiculously broad claims in pretty much all cases with software patents, even if the innovation itself was not as stupid as in this case. An example is the base patent on MP3 compression, whose claims cover all iterative music compression schemes in which an entropic encoder (such as Huffman encoding) is used in the loop and whereby the loop stops when you've reached the desired bit rate.
You've provided a lot of interesting links. I need to read and digest them before responding (I mean we are having one of the most intellectual discourse I've seen here on/., so I'm not going to muck it up by just blasting off another post). However...
Thanks for the compliment! I'm also glad I finally found someone pro-software patents who wants to read the stuff I point to:)
first-to-market is great, but you still have to fight it out. Copyright buys you virtually nothing (even the SSRN article you provided echoes this) and trade secret doesn't buy you much because you would have to show that your competitor stole your trade secret. If they came up with it clean-room, you've got nothing on them. Combining them doesn't help much.
I disagree with the SSRN article on that point. Copyright is quite strong imho, because you indeed almost need a clean-room implementation in order to evade it (just look at the hassle they went through to re-implement the original IBM PC bios). After all, copyright protects a lot more than literal bit-copies (translation also falls under it and just changing some things here and there is not enough either).
Therefore, the time necessary to reverse engineer it, re-implement it, integrate it in your application and test it will take quite a lot of time (possibly even as much as the original implementors). The reverse engineering component is (virtually) non-existent in case of e.g. business methods and interface elements, but that immediately demonstrates how the "disclosure value" of such patents is (virtually) non-existent either: using the technique is the same as publishing how it works, so there is no reason for society to grant a monopoly for it. Companies love it of course, since they more or less get something (a monopoly) for nothing (a publication that they were going to do anyway).
Nevertheless, even of those people who think patents are completely unsuitable for monopolising software-related innovations, there are some who think that the protection offered by copyright is not strong/deep enough for things like very complex algorithms. Some alternatives are discussed here.
Like I said, I need to read what you posted before trying to reply. It would do neither side any good if I really responded beforehand.
Let me begin by saying though that the FTC report is meaningless. It is constantly derided as "so what" because the FTC has no say in patent matters as well as the fact that it is easier to criticize than to fix. I can tell a fat person to eat less, but that doesn't really fix the problem, nor is it really an insightful comment (the FTC's comments that is).
Are we talking about the same FTC report? The FTC doesn't simply say that there are too many patents or that there is a quality problem (they mention that as well, but that's not all). They also say that economic effects and effects on competition should be taken into account when extending the scope of patentable subject matter, or in general whenever patent-related decisions are taken. They also interviewed a lot of people from the business and the field to find out what they thought (after all, the patent system is there for them, right?).
They may have no a say in patent matters, but that doesn't mean their report is non-sensical. On the contrary, I think points 6 and 10 of IPO's response to the FTC report to be completely crazy and unworldly, even though they have a lot of (indirect and direct) say in patent matters.
I completely disagree with this statement. While it sounds reasonable, I have yet to be shown anyone that acted this way. Look at Edison (1,093 patents). Look at Hammond (800 patents). These people got patents and kept innovating, often in the same areas of technology. A 20 year patent is great, but why not extend your monopoly by improving your tech and filing a patent on the improvement. *BAM* 25 year monopoly (assuming 5 years between filings). The lazy inventor is a reasonable idea, but not a realistic one.
It is used a lot by parasite companies like EOLAS (browser plug-ins), Forgent (RLE compression in JPEG), Acacia ("video streaming"),... . They have made a whole new business model out of this idea. It would surprise me if other companies did not do this on a smaller scale.
In practice, programmers can barely understand the legalese of software patents.
This is because a) the patent is written poorly or b) the programmer is lazy. Seriously. I have written several patent applications that contain no legalese except for the language in the claims (the enabling detailed description uses all English or trade language). I have also asked programmers what they thought their previous patent applications (filed by another firm) discussed and their repsonse was "I don't know, I couldn't be bothered to read it." The ONLY reason I write claims in "claim-ese" is because certain words mean certain things in the patent world (e.g., avoid "contains" because it is exclusive).
There are many poorly written patents (yes, I have read already a lot). And the claims are also quite important, since not all programmers work in the context of a large company where they can afford a lawyer to sit next to each programmer to check when a design decision infringes on a particular claim. In fact, even large companies don't do this. Companies like IBM simply count on their huge portfolio to get a cross-licensing deal if necessary.
To summarise it as one interviewee in the FTC report said: "There is too much information and it's no longer meaningful". As far as programming is concerned, the patent database is one huge collection of ideas. A programmer in general does not have a problem coming up with ideas. Time is spent designing them into an application, implementing them and testing them.
And in case a patent is about more than some general idea, as in case of e.g. one of the mp3 patents, you end up with claims so broad they include everything but the kitchen sink. I'm still w
Tell me, which is more an incentive to innovate (to the innovator):
A) You come up with an idea and we'll let you fight it out in the marketplace with four other companies OR
B) You come up with an idea and everyone in the U.S. will have to come to you for the solution. And it will be you and only you for the next 20 years.
well? OK.
You are missing several things of the big picture:
Once someone has a monopoly on something, he will be much less inclined to keep on innovating, since the competition is very much restricted in improving what he did. Therefore, this monopoly is only justifiable if without the outlook to this monopoly, the innovation would never have happened. At least in the software world, this is highly unlikely, since competition is the main drive there to innovate (if you don't innovate, you can as well close up shop). That's confirmed by, among others, this study (presentation slides, see especially slides 15-16) by the Fraunhofer and Max Planck institutes and the FTC report on the effects of patents on innovation.
An innovator does not live alone in the world. Once he gets a monopoly, he will without a doubt hinder other innovators with this patent. If generally this hindering effect causes more innovation not to happen than the amount of innovation that happens thanks to the fact that patent protection is available, it's also better that you have no patents. Again, the FTC report notes that this is the case in the software field. The main reason is that innovation in the software field is mainly incremental (improving things that other people have done before) instead of revolutionary (doing completely new things).
Patents increase innovation by forcing you to tell the world about your invention. Once the patent is up, ANYONE can take you patent and solve the problem you solved.
That's the theory. In practice, programmers can barely understand the legalese of software patents. And since software patents do not even include source code, it's even arguable whether they really contain a usable solution in many cases. Finally, (and this goes for all patents) companies are actively discouraged to go looking in patent databases looking for solutions. The reason is that, even if they did not find a solution and came up with something themselves, if they are sued afterwards, they can be ordered to pay tripple damages, because in that case it's considered "willful infringement".
The quid pro quo though is that in exchange for the disclosure, as a "thank you" for innovating and telling the world, you get the right to prevent others from implementing your solution. Which is a bigger thank you? that you get to fight it out in the marketplace, or that you and you alone can practice the invention?
It has nothing to do with "thank you", but everything with "if we didn't give you this monopoly, we would be even worse off, because we wouldn't know how to solve that particular problem". This monopoly can thus only be justified if there is a very low chance of independent rediscovery by other people and if the original innovator would very likely not have done the innovation himself without the incentive of getting a 20 year monopoly.
Please explain to me how the patent system benefits society as a whole, as you've asserted, other than the quid pro quo I stated above i.e., the patent tells anyone how to solve a given problem, in exchange for which, you get the limited monopoly.
The theory is that the disclosure of the innovation benefits society more than that the 20-year monopoly hampers it. This theory may have held back in the 15th century, when you usually had "on
Patents are supposed to be a purely economical tool to encourage investment in innovation. You're right that software patents do not reach that goal. However, patent lawyers make good money writing them and litigating about them, and some big companies like IBM and Microsoft see them as tools which can be very effective when dealing with smaller companies which threaten some part of their market (+ for rent seeking).
The result is that those two classes keep claiming, against all evidence to the contrary, that software patents are indispensable for economic growth and that without them, the end of the world is near. Other proponents are patent offices (they get their income from granting patents) and parasite companies like EOLAS (they don't make any products, they just buy some patents and then go around suing everyone to extract money).
The problem is that it's very hard for many people to understand that the, at first sight logical, rule that "more patents = always better" is a fallacy, as even a VP of IBM admitted. There is also a downside to having more patents in the system, and the balance is completely lost...
Because it would cost a lot of effort and money which can't be spent on more useful things. Besides, not all SME's use open source software (I would guess even only a minority does), and they don't have money for this either (and yet they are perfectly capable of innovating without requiring the incentive of being able to get a monopoly on the underlying principles).
So a better suggestion is to simply abolish software patents, since then you lose a lot of unnecessary and even hampering juridical overhead.
Bottom line: crappy patents will fall quickly if they are crappy enough, barring unclean hands on the part of the challenger.
You are missing the point a lot of patents (especially software patents and the like) are used strategically. As your linked article states, most patents are indeed not used to earn back investments (even though that is what patents were supposed to allow, in order to encourage innovation and economic welfare).
They are mostly just bartering tokens, traded among themselves by the big companies and extorted by them from the small ones and from the mythical lone basement inventors. These large companies don't defend a patent because the patent earns them so much money, but because it gives them control over the other party. They can effectively control who can enter the market and who can't, and that's what the strategic patenting is all about.
And no, that's not just some nutty conspiracy theory of mine, it's part of FTC study on the effects of patents on innovation. And I sincerely hope you, as upcoming patent lawyer, will not dismiss this like your colleagues at IPO who bluntly state that they do "not support the FTC recommendation concerning considering potential harm to competition in deciding upon the scope of patentable subject matter" and do not "support expanding economic considerations in patent law decision making.
".
I really don't understand that stance. Patent law is a purely economical law for X sake, so why shouldn't economical effects have precedence?
This is a bad analogy because it involves screwing the customer over. A patent lets you put your competitors out of business or it makes it significantly hard for them to do what you do, thus giving you the advantage.
Actually, in this case it involves screwing over society/the economy (since it's society that grants patents to innovators, and which loses when overly broad patents are granted). It's not just competitors, since more competition means lower prices and more incentive to keep on innovating (as long as you are able to recoup your investments, of course; see below), which both benefit society.
A patent is a business tool to put your competitors out of business. Yes it sounds harsh, but a business is about maximizing profit.
This so-called strategic patenting is how businesses use them (especially in case of e.g. software patents), but that's not the idea behind granting patents. The goal of granting patents is to further innovation and the economy at large, thereby providing benefit to society as a whole. The fact that innovating companies profit from this is just a means, the goal is not to let them profit as much as possible and allowing them to screw over the whole system.
Therefore, the parent was right in asserting that a company filing for overly broad claims is trying to steal from society, when it tries to appropriate things to which it doesn't have the right.
You have some interesting ideas in there, but unfortunately you've also lumped in things like polls.
Since that data is primarily intended for governments which have to make a decision on the "directive on the patentability of computer-implemented inventions", I added that poll because it was carried out by the European Commission in preparation of this directive. It's data the EC has supposedly based its conclusion on that software patents are desirable for Europe. I simply demonstrate that it takes a lot of spindoctoring to be able to conclude that from that data.
That in particular is an interesting idea to consider. The slice of the research pie for patent lawyers has grown. However, did Bessen and Hunt also check whether the size of the whole pie had grown, meaning the amount for research was bigger despite its share being smaller?
I would suggest you to read this shorter and non-technical version of their research paper, especially the part about patents and R&D being substitutes instead of complements in the software world. This one is also quite short and informative.
Would you grant that it's possible that patents have just grown the pie of money for software research, even after the money spent on maintaining the patents is subtracted out?
I think the total amount of R&D has grown because the sector has grown. Software patents may not hamper enough for R&D to come to a standstill and maybe (for now?) not enough to outpace the natural growth of the industry, but that's not really the point.
You still haven't answered the question of when innovation actually gets hindered.
Here's my research on that. I only know of *one* study in the entire world which claims that software patents promote instead of hinder innovation, and that's this one. It's an "economic" study carried out by a professor in law connected to a school of law, and uses an interview method (so the results obviously depend a lot on who you interview).
Is there any evidence that innovation in software is being hindered by patents?
See above.
OSes, server applications, programming languages and environments, user interfaces, network protocols, security infrastructure, and other areas seem to be improving as fast as ever
Who knows how much faster it'd go without software patents? Those studies point out that software patents are not required to recoup investments in software innovation (you can easily recoup that using time-to-market, secrecy, NDA's and copyright protection), but that they're almost exclusively used to get a lock on the market (i.e., make sure you have to invest less in innovation, since you can keep milking the old ones) or making sure you don't get locked out of the market yourself (defensive patenting).
And then there's the problem with patent thickets, the fact that software patents pretty much nullify most protection an author gets from copyright (one software patent can completely forbid you from exercising your author's right), the inherent problem of trivial patents, legal uncertainty,...
In fact, the only time I've seen a patent cause real trouble is in KDE, where some people attempted to blatantly copy a Mac OS feature only to find it was patented. In this case the patent worked as intended.
Actually, it probably didn't. The patent system was never designed to allow for monopolisation of ideas/concepts. In fact, these things do not fall under any kind of "intellectual property" regime, and therefore cannot be appropriated nor "stolen". That's the theory anyway, but software patents nicely circumvent this. There is however no economic rationale for allowing this monopolisation of ideas, and in fact many arguments for not allowing it.
Keep in mind that patent law is not a law designed to allow innovators to rightfully profit from their work. This is in contrast with copyright, where there is a moral right associated with the creative work of an author. Patent law is merely an economic law, and is supposed to be only applied if it results in a general benefit to society (after all, it's society that grants a monopoly, so it's only in the interest of society to do that if society will benefit from it, be it under the form of more innovation, better economic situation,...).
No, patents draw a line between creators and imitators.
That's wrong, you can perfectly infringe on a patent without imitating. Copyright protects from imitation, patents give someone an unconditional monopoly. If you independently discover the same thing, bad luck. And in the software world, this happens a lot more than in other fields, concludes the FTC.
It is to be expected that part-time amateurs come up with ideas years after professional innovators invented them and file them with USPTO.
See, here you admit yourself it's not about imitation. And you probably meant "some time after professional lawyers filed for the patents paid for by big companies, who can actually afford this filing and litigating about them afterwards".
The patent system indeed does not have any possibility for differentiating between "truly non-obvious things" (which are only obvious in hindsight) and plain obvious things. That's why some people argue that the patent system is simply unfit for pure abstract advances, because advances there happen a lot more than in the real world, where your ideas are hampered by this pesky physical material that doesn't always do what you want it to do.
Alternatives, which do take into account this problem, are presented here. Basically, you present the problem you have solved, claim it is really non-obvious, promise to give anyone who solves it within a month $5000 and if nobody can, then you get your monopoly.
This solves several problems of the patent system: filing for monopolies on trivial things is very much discouraged and people who spend time invalidating stupid applications get automatically reimbursed for their work.
Otoh, the UK is pretty much the only EU countries that almost invariably holds up software patents in court, so they sure aren't that patent unfriendly...
Slashdot Mirror
Next, such a challenge easily costs US$1.5-2 million.
Finally, it is quite trivial to word abstract software ideas in a way that they become a process executed by a computer (after all, that's the only useful thing you can do with a computer program). So that's not really a limitation to what you can patent.
And then you get patent descriptions like "an information carrier characterised in that it contains a story involving a boy and a girl who both have an affair, whereby ...". It's not all that different from how program claims in software patents look.
In the end, one could even start arguing for patents on patent descriptions themselves: just like a computer program contains instructions that tell you what to do, a patent description for any physical invention also contains a description of such steps.
Patents on information entities simply don't work, because then the system deadlocks: patents are supposed to encourage sharing of information, but software patents forbid that same sharing if you use a particular class languages (programming languages) to express it. And the reason is simply that the information is the product.
That's also exactly why we have copyright and why it's appropriate for the protection of computer programs: there the way how you express something is what matters and what is monopolised, and not what you express.
Rowling and her lawyers cannot claim in their cases that they possess the idea of some wizard kid and the things he does, but only on the way Rowling expressed it. If someone else bases his expression on Rowling's expression, then this is a violation of copyright. Not because he used the same idea, but because he used her expression of that idea as basis for his work. There is a substantial difference there, and then you don't have pure original work anymore (but something based on her work).
Where the line has to be drawn between original work and plagiarism is for the courts to be decided on a case by case basis. The fact that there is *also* some original work, is irrelevant. Otherwise, I could just take a Harry Potter book, change all names, switch the word order of some sentences, add a few new sentences here and there, possibly add a chapter of my own and then sell it. Or maybe I could just create a translation of the work and sell it.
You're right, that must happen before. But you simply have to do a spectral analysis and then quantise the coefficients to infringe (regardless of how you do this). That's not really limiting their monopoly either...
If the patent even partly applies to sender ID, you're screwed nevertheless. And it's quite possible that the patent office will consider the combination of those two things non-obvious as well, resulting in another patent on the whole thing.
That said, I have no idea whether their called-ID patent applies nor whether they already have another patent (pending) on the combination of SPF and called-ID.
The base MP3 patent is much broader than just what is needed "to get back some of the money they spent on research". They patented compressing (using any kind of algorithm) sound, as long as you do it in a loop that involves an entropic encoder (such as Huffman encoding), until the sample reaches the desired bitrate and if you store the result together with the bitrate afterwards (or send it over a network).
The AC replied that enforcing this would be only possible through patenting and licensing of the programming language, and then said that such a clause would impose a restriction on people's abilities to make money with the software they wrote (because they would not be able to get a patent on what they wrote, since that would go against the license under which the language is used).
So he was claiming that without software patents, you can't make money on the software you write.
GPL keeps the code (and as such the information) free, BSD keeps the individuals free to do what they want to do. It's simply different kinds of freedom. Get over it.
This is not about patents, it's about copyright. And the fact that you give something away for free, does not mean you should be allowed to infringe on copyright (otherwise e.g. Microsoft would be allowed to infringe on the GPL by incorporating GPL'd code in a closed source Internet Explorer, back when they were giving it away for free).
Besides, software patents allow others to steal your profits, which you made by selling a program you wrote all by yourself. It's really great, you spend several man years on writing a program, bring it to market and then get a ton of nice requests for royalties because auto-updating, RLE compression, plug-ins etc are all patented.
Or maybe one of those patent holders has a similar program himself, and doesn't even offer you a license. Tough luck.
It's not a program, it's a protocol, exactly like the TCP/IP and HTTP you mention. You can both write proprietary and Free Software that uses it. Which license some implementation from Apple uses is irrelevant (or do you consider HTTP to be non-Free as well because Internet Explorer and IIS use it?).
With non-software patents (i.e., where the invention/innovation lies in a novel way of using physical forces/material), how far exactly you are allowed to abstract is mainly limited by three things:
- You can't abstract the claims until only the forces of nature you are harnessing remain, because those are not patentable;
- You can't abstract the forces of nature you harness out and remain with some generic algorithm/method that could apply to anything, because then your invention (novel way to use those forces) is no longer part of the claims;
- You're of course also limited by prior art (you invent a new car, but other cars already exist -> you can't claim all 4-wheeled vehicles etc) and whatever the patent office deems too general (after all, society grants you a monopoly in return for disclosure of an invention, so those two should -in theory- be proportionate).
However, if you look at software patents, then- There are no unpatentable basic "forces of information";
- Since what you start with is already some abstract method/algorithm, no matter how much you abstract it further, you can always argue that your invention is still embodied in those claims;
- This one is the only thing left.
The net result is indeed that you end up with ridiculously broad claims in pretty much all cases with software patents, even if the innovation itself was not as stupid as in this case. An example is the base patent on MP3 compression, whose claims cover all iterative music compression schemes in which an entropic encoder (such as Huffman encoding) is used in the loop and whereby the loop stops when you've reached the desired bit rate.Therefore, the time necessary to reverse engineer it, re-implement it, integrate it in your application and test it will take quite a lot of time (possibly even as much as the original implementors). The reverse engineering component is (virtually) non-existent in case of e.g. business methods and interface elements, but that immediately demonstrates how the "disclosure value" of such patents is (virtually) non-existent either: using the technique is the same as publishing how it works, so there is no reason for society to grant a monopoly for it. Companies love it of course, since they more or less get something (a monopoly) for nothing (a publication that they were going to do anyway).
Nevertheless, even of those people who think patents are completely unsuitable for monopolising software-related innovations, there are some who think that the protection offered by copyright is not strong/deep enough for things like very complex algorithms. Some alternatives are discussed here.
No problem, and thanks again!Are we talking about the same FTC report? The FTC doesn't simply say that there are too many patents or that there is a quality problem (they mention that as well, but that's not all). They also say that economic effects and effects on competition should be taken into account when extending the scope of patentable subject matter, or in general whenever patent-related decisions are taken. They also interviewed a lot of people from the business and the field to find out what they thought (after all, the patent system is there for them, right?).
They may have no a say in patent matters, but that doesn't mean their report is non-sensical. On the contrary, I think points 6 and 10 of IPO's response to the FTC report to be completely crazy and unworldly, even though they have a lot of (indirect and direct) say in patent matters.
It is used a lot by parasite companies like EOLAS (browser plug-ins), Forgent (RLE compression in JPEG), Acacia ("video streaming"), ... . They have made a whole new business model out of this idea. It would surprise me if other companies did not do this on a smaller scale.
There are many poorly written patents (yes, I have read already a lot). And the claims are also quite important, since not all programmers work in the context of a large company where they can afford a lawyer to sit next to each programmer to check when a design decision infringes on a particular claim. In fact, even large companies don't do this. Companies like IBM simply count on their huge portfolio to get a cross-licensing deal if necessary.
To summarise it as one interviewee in the FTC report said: "There is too much information and it's no longer meaningful". As far as programming is concerned, the patent database is one huge collection of ideas. A programmer in general does not have a problem coming up with ideas. Time is spent designing them into an application, implementing them and testing them.
And in case a patent is about more than some general idea, as in case of e.g. one of the mp3 patents, you end up with claims so broad they include everything but the kitchen sink. I'm still w
You are missing several things of the big picture:
That's the theory. In practice, programmers can barely understand the legalese of software patents. And since software patents do not even include source code, it's even arguable whether they really contain a usable solution in many cases. Finally, (and this goes for all patents) companies are actively discouraged to go looking in patent databases looking for solutions. The reason is that, even if they did not find a solution and came up with something themselves, if they are sued afterwards, they can be ordered to pay tripple damages, because in that case it's considered "willful infringement".
It has nothing to do with "thank you", but everything with "if we didn't give you this monopoly, we would be even worse off, because we wouldn't know how to solve that particular problem". This monopoly can thus only be justified if there is a very low chance of independent rediscovery by other people and if the original innovator would very likely not have done the innovation himself without the incentive of getting a 20 year monopoly.
The theory is that the disclosure of the innovation benefits society more than that the 20-year monopoly hampers it. This theory may have held back in the 15th century, when you usually had "on
The result is that those two classes keep claiming, against all evidence to the contrary, that software patents are indispensable for economic growth and that without them, the end of the world is near. Other proponents are patent offices (they get their income from granting patents) and parasite companies like EOLAS (they don't make any products, they just buy some patents and then go around suing everyone to extract money).
The problem is that it's very hard for many people to understand that the, at first sight logical, rule that "more patents = always better" is a fallacy, as even a VP of IBM admitted. There is also a downside to having more patents in the system, and the balance is completely lost...
So a better suggestion is to simply abolish software patents, since then you lose a lot of unnecessary and even hampering juridical overhead.
They are mostly just bartering tokens, traded among themselves by the big companies and extorted by them from the small ones and from the mythical lone basement inventors. These large companies don't defend a patent because the patent earns them so much money, but because it gives them control over the other party. They can effectively control who can enter the market and who can't, and that's what the strategic patenting is all about.
And no, that's not just some nutty conspiracy theory of mine, it's part of FTC study on the effects of patents on innovation. And I sincerely hope you, as upcoming patent lawyer, will not dismiss this like your colleagues at IPO who bluntly state that they do "not support the FTC recommendation concerning considering potential harm to competition in deciding upon the scope of patentable subject matter" and do not "support expanding economic considerations in patent law decision making. ".
I really don't understand that stance. Patent law is a purely economical law for X sake, so why shouldn't economical effects have precedence?
Therefore, the parent was right in asserting that a company filing for overly broad claims is trying to steal from society, when it tries to appropriate things to which it doesn't have the right.
And then there's the problem with patent thickets, the fact that software patents pretty much nullify most protection an author gets from copyright (one software patent can completely forbid you from exercising your author's right), the inherent problem of trivial patents, legal uncertainty, ...
Actually, it probably didn't. The patent system was never designed to allow for monopolisation of ideas/concepts. In fact, these things do not fall under any kind of "intellectual property" regime, and therefore cannot be appropriated nor "stolen". That's the theory anyway, but software patents nicely circumvent this. There is however no economic rationale for allowing this monopolisation of ideas, and in fact many arguments for not allowing it.Keep in mind that patent law is not a law designed to allow innovators to rightfully profit from their work. This is in contrast with copyright, where there is a moral right associated with the creative work of an author. Patent law is merely an economic law, and is supposed to be only applied if it results in a general benefit to society (after all, it's society that grants a monopoly, so it's only in the interest of society to do that if society will benefit from it, be it under the form of more innovation, better economic situation, ...).
Alternatives, which do take into account this problem, are presented here. Basically, you present the problem you have solved, claim it is really non-obvious, promise to give anyone who solves it within a month $5000 and if nobody can, then you get your monopoly.
This solves several problems of the patent system: filing for monopolies on trivial things is very much discouraged and people who spend time invalidating stupid applications get automatically reimbursed for their work.
Otoh, the UK is pretty much the only EU countries that almost invariably holds up software patents in court, so they sure aren't that patent unfriendly...