Software and business method patents take a hit

A blog article by Dana Blankenhorn provides some links, including to Groklaw. I'm curious what other people have been reading, and whether they think this could be the beginning of the end of information/mental process patents?

Question: If software is distributed unbundled with any specific hardware (ie: "not tied to any machine"), then are the methods it implements patentable? I have no problem with a patent regime that applies to those shipping hardware/software bundles, but not to those simply shipping/sharing software. This would in my mind largely solve the incompatibility between software patents and FLOSS.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Answer

Yes, they can be. This decision restricts the scope of what is considered patentable subject matter, and I agree the world is a better place without many of these granted business method patents.

Patents do not need to specify an exact, specific, purpose-built machine to carry out the method in order to be valid. Even if the software is designed to be run on a general purpose PC, it can still qualify for patent protection.

The inventive step of the patent is not the specific machine on which it runs, but the method that it performs. It doesn't necessarily matter which machine is used to perform the method - it could be a purpose-built FPGA, designed to perform exactly this one function, or it could be a general purpose CPU, with properly written software.

There's no doubt that this may help clarify patentable subject matter, but there's no question that software patents are here to stay. Good software patents will qualify for protection around the world, even in Europe.

I attended the European Patent Office Patent Information Conference last week in Stockholm, and spoke with practicing professionals and European patent examiners. I can tell you that this isn't even really a significant issue at the moment. Nearly everyone accepts the validity of well-written, meaningful software patents.

Nearly everyone

"Nearly everyone accepts the validity of well-written, meaningful software patents."

I guess I'll always be proud to be excluded from this "nearly everyone" *smile*

I'm also suspect of the statistical validity of this suggestion, and the entire posting that was built upon it. Time will tell, and I am not someone who believes that information/mental process patents are here to stay any more than I believed those who told me that the USA could never elect a black president.


Free/Libre and Open Source Software (FLOSS) consultant.

Touche

Good point there - I was speaking specifically of the context of patent professionals, not the general public. Patent examiners, licensing professionals, etc. are not concerned about software patents per se. And frankly, Slashdotters concerned that Microsoft is trying to patent the Page Down key don't know what they're talking about, so they don't (and shouldn't) get taken seriously.

There is a huge concern about patent quality, in all technical categories. Overloaded examiners, duplicitous applicants, insufficient training and access to prior art, and a system being gamed by those who know how - these are real concerns to everyone.

Please don't lump software patents in with 'mental process' or business method patents - they aren't the same thing. I also applaud the recent Bilski decision, and hope it serves to limit the number of these BM patents.

Software patents really are here to stay, and they play an important role in our economy. They serve as a purpose for disclosing innovations to the general public. Without patents, why would a private corporation disclose their groundbreaking new compression algorithm, or communication system? They would rely on trade secrets instead, and not disclose their inventions at all. Patents are preferable to a system of trade secrets.

There are some proposed changes to the patent system that probably have more chance than banning software patents. For example: abandoning the 'one-size-fits-all' approach to patent life, allowing the community to submit prior art (already in a pilot program), mandatory licensing instead of the threat of an injunction, increase cooperation between patent offices, etc.

I'd be interested in hearing your opinion on other approaches to software patents...

Areas of discussion

"Please don't lump software patents in with 'mental process' or business method patents - they aren't the same thing."

Here we disagree. Unless we are using the word "software" in some different way, they are all information/mental process patents. While there is finally a recognition of the problem with business method patents, the same will eventually be true of software patents (Which is different than patentable processes that happened to include software).

I also disagree when people take the general justification for patent law, which I agree with, and apply it to software. What makes patent law useful in hardware (tangible manipulations of nature) or pharmaceuticals does not translate directly to software. When we talk about those so-called "grey areas" like FPGA's, I don't see as grey. I don't believe this "grey" area lends support to patentability of software, but subtracts from the justification for patents on FPGA programming.

The very things that separate tangibility from intangibility (like a natural zero or non-zero marginal costs) are also the same things which make the economics of patent law eithor an incentive or disincentive to innovation. Where the natural marginal cost is non-zero (rivalrous, etc), the patent bargain works relatively well, but where it is zero then the patents serve as a disincentive.

That said, I realize that there are people who have an ideology that supports patents in anything that can be described as if it were patentable subject matter ("anything under the sun made by man", with a very subjective meaning of the word "made"). I then turn to other practical arguments.

We need to come up with more objective criteria for quality, and get past the presumption of validity of patents in courts. We then need to divide up the very different subject matter areas granted patents and analyse each separately. For areas where the vast majority are of poor quality (and that will "coincidentally" include any subject matter areas with a zero marginal cost), then patents in that subject matter should not be offered legal protection. This will provide the proponents in that subject matter area to work to improve the system. I happen to believe it is a lost cause for any area of zero marginal cost, but at least the costs of figuring that out would rest with those ideologues rather than the sector or the taxpayer.

Another option is to deal with specific business model incompatibilities. Marginal prices should not be able to be demanded in cases where a zero marginal cost is possible. I already described in many places an example of zero marginal priced software (FLOSS) unbundled from hardware as being offered a "fair use" exception in patent law.


Free/Libre and Open Source Software (FLOSS) consultant.

Dancing on the head of a pin

the same will eventually be true of software patents (Which is different than patentable processes that happened to include software).

I don't know exactly what you are defining as a 'software patent' then. To me, a good 'software patent' would simply be an invention that may be best implemented in software. There is nothing wrong with patenting a novel process dealing with information, so long as there is a real, tangible, and technical benefit.

You may dislike both business method and what you are calling 'software patents', but there is a significant difference between them.

the patent bargain works relatively well, but where it is zero then the patents serve as a disincentive

You're going to have to back that up with some credible research, because I fundamentally disagree. I don't see a problem with splitting up the patentable technical areas, and treating the sections differently. For example, perhaps the ideal term of software patents is only 5 years, instead of 20. I think it would be very difficult to implement, but I think it might be an idea worth pursuing.

We need to come up with more objective criteria for quality, and get past the presumption of validity of patents in courts.

Agree with the first point, couldn't disagree more with the second. Corporations and individuals alike require greater certainty in patents, not less. Eliminating the presumption of validity reduces the purpose of the patent office to a registration system, with no place for any kind of substantive examination. If you don't like junk patents being issued, then this is precisely the most dangerous path we could take.

Marginal prices should not be able to be demanded in cases where a zero marginal cost is possible.

Sorry, Russell, but I like capitalism. The price of anything is, and should remain, relative to the value to the purchaser, not to the cost to the manufacturer. If I can make a thousand widgets for a penny, but they're worth $1000 to consumers, then I am, and should remain, entitled to charge anything between $0.01 and $1000. This is just as true for software as it is for hardware.

There is a very valuable place in the economy for FLOSS, but I'm afraid it doesn't get a pass from the regular rules of the existing economy.

I like capitalism

"Sorry, Russell, but I like capitalism."

And thus signals the end the thread, as nothing is realistically left but the references to Hitler ;-)

I think you have confused copyright (pre 1996 WIPO), which is what allows a full spectrum of business models across competing implementations of a given method (Which would allow FLOSS and non-FLOSS to co-exist) and patent law which does not facilitate that competitive co-existance. Your ability to charge royalties is dependant on copyright, not patents, so please don't try to confuse the conversation by incorrectly suggesting that all PCT is the same.

We'll disagree on whether it is FLOSS that might "get a pass from the regular rules of the existing economy", as it is non-FLOSS that has been getting that free pass from the regular rules of the economy, through government granted statutory (and other) monopolies. "Real supporters of capitalism" (IE: not those into government handouts/bailouts/monopolies) recognize that the government isn't good at picking successful businesses or business models, and that offering government intervention in the marketplace (a realistic definition of patent policy) needs to be done extremely cautiously and minimalistically.

Presuming that exclusive rights should exist in all subject matter and then map our understanding of the economy backwards from that is, well, backwards. We need to start from recognizing that statutory monopolies exist for a clear public policy purpose, and that advances in the economy must lead to changes in the scope of these statutory monopolies. The growth of business models which do not need as strong statutory monopolies feeds directly into analysis of the purpose of patents, and thus the necessity and validity of having patents in a given subject matter.

I'm curious: Have you read Yochai Benkler's The Wealth of Networks yet? A smaller version of the thesis is in Coase's Penguin, or Linux and the Nature of the Firm which puts the methods behind FLOSS in their appropriate economic context. Suggesting that the laws shouldn't modernize as a result of these new methods of organization would be as silly to me as suggesting that we should not have created the concept of a corporation as we organized ourselves in firms.


Free/Libre and Open Source Software (FLOSS) consultant.

what's "tangible" about software ?

To me, a good 'software patent' would simply be an invention that may be best implemented in software.

The sounds like an algorithm to me - something that could just as easily (if much slower) be performed entirely in a human brain. You mention compression algorithms as examples of good software patents, but they're just abstract mathematics - the very thing that the patent system was intended *not* to protect.

To my mind, anything that manipulates *information* shouldn't be patentable. Why? Because "manipulating information" is also called "thinking".

As for disclosure, in my experience most software patents go out of their way to try to monopolise as much as possible while revealing as few actual details as possible.

Like Russell, I have no problem with patents that happen to include software (industrial process control machines spring to mind as very software-intensive, yet clearly not just algorithms).

In my experience, the people who are most keen on software patents are those that don't write software, while those that do are generally against them, and that speaks volumes (and unlike Russell, I don't work in the world of FLOSS).

Going to have to disagree with you there...

The sounds like an algorithm to me - something that could just as easily (if much slower) be performed entirely in a human brain. You mention compression algorithms as examples of good software patents, but they're just abstract mathematics - the very thing that the patent system was intended *not* to protect.
To my mind, anything that manipulates *information* shouldn't be patentable. Why? Because "manipulating information" is also called "thinking".

Compression algorithms are certainly not abstract mathematics. They are applied mathematics. Just because something can be done by humans doesn't mean that a machine that can perform the same task is therefore unpatentable.
Is there a real, tangible result of compression algorithms? Sure - MP3 players is an obvious answer. Sure, I *could* carry around a stack of paper representing the entire .mp3 file in hex code, and decode it in my head (actually, I really couldn't). This is obviously impractical.
This type of innovation is the very thing that the patent system was *intended* to protect.

As for disclosure, in my experience most software patents go out of their way to try to monopolise as much as possible while revealing as few actual details as possible.

While there are certainly such patents, particularly from the USPTO, I would again question your conclusion. It is certainly not *most* software patents - it is at worst a small minority. Did you take the time to try to understand the limitations provided by the claims?

Like Russell, I have no problem with patents that happen to include software

That is a difficult distinction to make, and I have trouble seeing the difference between an "industrial process control machine, heavily dependent on software", and an MP3 player.

MP3 players aren't compression algorithms

I have trouble seeing the difference between an "industrial process control machine, heavily dependent on software", and an MP3 player.

You weren't talking about copyrighting an MP3 player, but the compression algorithm used to create MP3 files - those are very different things. I can see lots of things about an MP3 player that might be perfectly reasonable to have as claims in a patent :
- storing the music in the player itself, rather than on external media
- possibly various things related to form factor, user interface, etc, if they're novel enough.

I can't see anything that should be patentable about the MP3 compression algorithm - at its most fundamental it's just a way of taking a big number and converting it into a smaller number.

Patents, not copyright

You weren't talking about copyrighting an MP3 player, but the compression algorithm used to create MP3 files - those are very different things. I can see lots of things about an MP3 player that might be perfectly reasonable to have as claims in a patent

I wasn't talking about talking about copyrighting anything, but patenting. Those are indeed very different things.
Storing the music in the player itself wouldn't pass a simple obviousness test. When it comes to playing stored MP3 files, the primary inventive step is in the algorithm, not in where or how the files are stored.

I can't see anything that should be patentable about the MP3 compression algorithm - at its most fundamental it's just a way of taking a big number and converting it into a smaller number.

I can take any number and make it a smaller number. The trick is to do it in such a way that something approximating the original number again can be calculated. MP3 was unquestionably innovative.

At its most fundamental, an LED simply converts electrical energy into light. Be careful not to oversimplify - the details are important.

I still see nothing fundamentally different between a "industrial process control machine, heavily dependent on software" and an MP3 player. In both cases, the inventive step lies in the software.

Understood, but disagree.

I still see nothing fundamentally different between a "industrial process control machine, heavily dependent on software" and an MP3 player.

One is a process used to manipulate tangible forces of nature, and the other is a process used to manipulate intangible knowledge that follows from an understanding of nature (in this specific case, sound). It isn't the 'software' that is the determining factor for whether the subject matter should be patentable, but what is being manipulated.

Many software patents aren't even that close to what I would consider patentable subject matter. In the case of patents on interface methods (file formats, communications protocols, etc) they aren't even based on an understanding of natural forces (sound, video, etc), but purely on manipulating digitally stored/encoded information.

If information/mental process patents (IE: where the manipulation in the process is of information and not of tangible forces of nature) are to have patent protection (and I'll never concede this is a reasonable thing, even if the law says otherwise), then we clearly need to have a "Fair Use" regime that deals with the obvious interoperability/competition issues, and incompatibilities with business methods based on pure intangibles.

Patent policy is, after all, economic policy and it would be wrong to have policy which fundamentally harmed rather than helped the economy. While I realize the pro-sw-patent people tend to base their economic analysis on faith, nearly every independent study I've read over the years on information/mental process patents have indicated that they are a friction and not an incentive in these subject matter.

The reason for the failure in some subject matter of policy that works well in other subject matter seems obvious to me, and is tied in the economics of tangibles vs intangibles. Unfortunately, those who don't see a difference between tangibles (naturally rivalrous) and intangibles (naturally non-rivalrous) will also be blinded to the very different economics.


Free/Libre and Open Source Software (FLOSS) consultant.

Agree to disagree

It isn't the 'software' that is the determining factor for whether the subject matter should be patentable, but what is being manipulated.

In the first case, we have a machine that makes widgets, and in the second, we have an audio player. Neither concept is patentable, as there is nothing novel here. In both cases, any inventive step lies in some software manipulating electrical signals input from one or more sources, and generating electrical output signals in response. In both cases, it is not the machinery around the software which is novel, but the software implementation.

Patent policy is, after all, economic policy and it would be wrong to have policy which fundamentally harmed rather than helped the economy.

On that, we can agree, except that I would go even farther to say they should help society as a whole, not necessarily just the economy. Intellectual property rights are only defensible if they serve the public good.

While I realize the pro-sw-patent people tend to base their economic analysis on faith

LOL - generally speaking, those in the patent profession tend to have the same opinion of the anti-sw-patent crowd.

Unfortunately, those who don't see a difference between tangibles (naturally rivalrous) and intangibles (naturally non-rivalrous) will also be blinded to the very different economics.

Of course I see a difference between tangible and intangible products. I don't see the clear distinction that you do between 'software' patents and 'non-software' patents. Sure, there are patents that plainly exist on one side or the other, but it's the fuzzy middle that concerns me.

If information/mental process patents (IE: where the manipulation in the process is of information and not of tangible forces of nature) are to have patent protection (and I'll never concede this is a reasonable thing, even if the law says otherwise)

And there we'll have to agree to disagree, because I'll never concede that this is anything but reasonable. I'm afraid that trying to get 'software' excluded from patentable subject matter is simply not going to happen. There are several other potential methods of addressing the issues, including 'fair use' exceptions and changing the term of patents. Perhaps there is some middle ground there, but it will take some significant effort to even get these options on the table.

PCT laws

One of the interesting things with PCT law is that there is so much diversity, that people who agree on one of these dissimilar areas of law (IE: copyright) can disagree on another (IE: patent). Or even more focused, I run into people who agree with me about information/mental process patents, but it then turns out they are opposed to patentability in general which I am not.

In a report I did for Industry Canada a few years back I had a number of specific recommendations. In the following years I have come to realize that adopting a "Fair Use" defence under patent law is the most politically feasible in the short term. This should handle not only situations involving interfaces which was the subject of Maureen A. O'Rourke's paper, but also in situations which would result in patent failure (IE: royalty-free unbundled software distribution cannot pay patent royalties). You keep speaking of an MP3 player as a hardware/software bundle, which is an example outside of what most of us are concerned with which is software unbundled from any specific hardware.

That hardware/software bundle would not lead to patent failure, although there are many potential competitive market failures that could lead to an abuse of exclusive rights. This is an area of law not adequately enforced, with our old-economy competition bureau largely ignoring competitive problems in PCT (See IPEG -- Intellectual Property Enforcement Guidelines of the Competition Bureau).

Information/mental process patents run into some of the same logistics problems that has lead many countries to exclude non-commercial tinkering/etc from patent enforcement (IE: a private citizen tinkering with their own car would not be found guilty of infringing patents involved in the manufacturer of cars). In the case of unbundled software it is not some "manufacturer" that is carrying out the process, but the end user of the software. This is another example of the obvious distinction between the tangible/intangible separation -- well, at least obvious to me :-) I'll be blunt and say that I'm not an educator and can't articulate something that is too obvious to me, so I'll have to leave that to someone else.


Free/Libre and Open Source Software (FLOSS) consultant.

Diversity

I'll agree with you there, Russell. I couldn't agree with you more on the copyright file, and yet we seem to be on different planets on the software patent issue.

Let me address a couple things from your last post:

This should handle not only situations involving interfaces which was the subject of Maureen A. O'Rourke's paper, but also in situations which would result in patent failure (IE: royalty-free unbundled software distribution cannot pay patent royalties). You keep speaking of an MP3 player as a hardware/software bundle, which is an example outside of what most of us are concerned with which is software unbundled from any specific hardware.

I haven't read the paper, so I can't comment on it. I do take exception to your comment on 'patent failure'. There is no failure here. A patent provides the patent owner with the right to exclude others from practicing their invention, not the right to charge royalties. The owner may, at their discretion, accept royalties or other forms of payment to allow the other party to take a license to the patent. If the other party cannot meet the terms required by the patent owner, then they cannot practice the invention. This is by design, not a failure. It's a feature, not a bug! ;)

You keep speaking of an MP3 player as a hardware/software bundle, which is an example outside of what most of us are concerned with which is software unbundled from any specific hardware.

Ah, but you hit my point on the head. With the MP3 player, the inventive concept is not in the hardware, nor in the hardware/software combination, but in the software itself! So, if you accept the MP3 player hardware/software combination as being patentable subject matter, then I'm afraid that this same MP3 algorithm, running on any hardware, is also patentable. This includes a PC.

a private citizen tinkering with their own car would not be found guilty of infringing patents involved in the manufacturer of cars

This would probably qualify under the private experimentation limitations of patent law, similar to fair dealing for copyright. Now if the private citizen created a business (for profit or not) using this technology, that's another matter.

Patent failure

There is no failure here. A patent provides the patent owner with the right to exclude others from practicing their invention, not the right to charge royalties.

You aren't taking exception with me, but the way the law is intended to work. When a patent isn't being licensed, the legal system calls that "patent failure" and it is discussed extensively in the legal literature. Using patents to exclude competition, rather than receive moral and material rewards for the invention, is considered anti-patent in the literature: not a feature as you claim. You seem to be speaking from the perspective of the desires of some vendor, not the perspective of the policy makers and legal system that this law exists within.

I didn't invent (smiles) the term "patent failure", or "anti-patent" -- I'm just using them as I have read them.

So, if you accept the MP3 player hardware/software combination as being patentable subject matter, then I'm afraid that this same MP3 algorithm, running on any hardware, is also patentable.

You seem to be looking at patent law as technologist, rather than as a person analyzing economic public policy. While from a purely technical perspective they may appear similar (innovation is in software, not hardware/software bundle), they do not have the same economic analysis and thus are not the same from an economic public policy perspective.

I also don't accept the MP3 hardware/software bundle as patentable subject matter, but for practical public policy purposes I focus my critique on the economic problems associated with unbundled software. If annalists agree with you that bundled and unbundled software is the same, then this should lend less patentability to the bundled software, not more patentability to the unbundled software (an information process).

I believe those looking at hardware equivalences are looking at the public policy question backwards. As you move from a patentable technique moving from being implemented in hardware to being implemented in software (including FPGA's), then the economic analysis moves us away from justification of the patent monopoly for this type of hardware, not lends patentability to information processes.


Free/Libre and Open Source Software (FLOSS) consultant.

"Patent failure"?

You aren't taking exception with me, but the way the law is intended to work. When a patent isn't being licensed, the legal system calls that "patent failure" and it is discussed extensively in the legal literature.

I have never heard the term before today, and I can find few references to this term beyond a recent book by Besser and Meurer. It is certainly not 'discussed extensively in the legal literature'. The 'right to exclude' is a fundamental part of the patent system, not something desired by a vendor, and is not considered to be 'anti-patent'. This is precisely the way the law is intended to work.

The exclusive right granted to a patentee in most countries is the right to prevent or exclude others from making, using, selling, offering to sell or importing the invention.

http://en.wikipedia.org/wiki/Patent

For example, in the US, the ITC has the authority to block importation of infringing products, but not to award damages. This is a classic case of 'exclusive rights' awarded to inventors.

Nice illustration of why I disagree

So, if you accept the MP3 player hardware/software combination as being patentable subject matter, then I'm afraid that this same MP3 algorithm, running on any hardware, is also patentable. This includes a PC.

Or somebody's head, right ? So we have patents acting as "thought control" - it's now illegal to think these thoughts without paying some patent holder. That end result is what shows me that something went wrong somewhere along the way. Monopolies on particular ways of thinking are bad public policy, IMO.

No thought control patents

Or somebody's head, right ? So we have patents acting as "thought control" - it's now illegal to think these thoughts without paying some patent holder. That end result is what shows me that something went wrong somewhere along the way. Monopolies on particular ways of thinking are bad public policy, IMO.

This type of argument comes up often on Slashdot. It's just as superficial and incorrect there. Let me put your mind at ease.

If you look at the claims of any of these patents, they include in the preamble a reference to some kind of machine or apparatus on which the algorithm runs. This is the current standard, even in the US. Pure algorithms are not permitted.

In the MP3 example, the claim might start off:
'An audio playback device, comprising of: ' ..,

So, therefore, if I were to think the entire MP3 decoding algorithm through in my head, type the results into a binary file, and then play the file, I would not be violating any patents on MP3 decoding.

Let me be clear: there are no patents restricting any train of human thought. If any do get past the (low) quality bar of the USPTO, they do not hold up in court. Patents are often invalidated for precisely this reason - insufficient ties to hardware.

NewSpeak, and ways of changing my mind.

If you look at the claims of any of these patents, they include in the preamble a reference to some kind of machine or apparatus on which the algorithm runs.

The reference to what amounts to general purpose computing equipment being used to manipulate digital information is a matter of wordsmithing the claims in a way to confuse the patent system, and isn't what I would consider a legitimate activity. Calling a cat a dog doesn't make it bark.

When the only difference between an algorithm carried out in someone's mind manipulating information and in a general purpose computer manipulating information is speed, that is an information/mental process. You don't seem to like this categorization, but I don't yet understand how it could be considered inaccurate.

Information/mental processes are different from, say, manufacturing processes in what the process is manipulating: information or physical artefacts. Because these processes are manipulating 'things' (And I use that term loosely if we are to include information) which are extremely dissimilar, it should be obvious that both the economic and legal analysis should be considered dissimilar.

Proponents of information/mental process patents have gone so far as to suggest to me that promoting different thoughts makes electrons move differently in the brain, which is a manipulation of a tangible force of nature. While a cute way to try to make everything into a barking dog, I believe it is ludicrus to suggest an equivalence between changing someones mind via reasoned debate and changing someones mind via a sledgehammer.

MP3 compression is an information/mental process that is based upon learning about tangible forces (sound waves), but is not a manipulation of tangible forces. As I suggested above, it is closer to what I would consider patentable subject matter than many information/mental process patents, but is still very easily distinct from processes which manipulate tangible forces of nature (applied natural sciences, manufacturing processes, chemical compositions, etc).

Interesting conversation, and one that is important to have, but unless a 4'th person joins in I think we've written all each of us can at the moment without just repeating (as I largely just did).


Free/Libre and Open Source Software (FLOSS) consultant.

Good conversation

Proponents of information/mental process patents have gone so far as to suggest to me that promoting different thoughts makes electrons move differently in the brain, which is a manipulation of a tangible force of nature.

LOL - that is quite a stretch. Even I wouldn't go there. :) And it's quite unnecessary, I might add. :)

The reference to what amounts to general purpose computing equipment being used to manipulate digital information is a matter of wordsmithing the claims in a way to confuse the patent system, and isn't what I would consider a legitimate activity.

And there we simply disagree, and I agree we're just arguing past each other at this point. We agree quite a bit on copyright issues, but we aren't going to get any farther on this issue, I believe.

You can always look forward to a healthy debate when you try to slip some more anti-'software'-patent talk up there, I'm afraid. I won't let you get away with it without a debate. ;)