Non-Obvious Patents
Home Resources Essays Hobbies Hockey! Projects Resume	(December 20, 2005, by Thomas Fischer, work in progress) I. Obviousness in Patents The patent system in the United States is flawed. There are far too many patents that are granted which are 'obvious'. In seeing all the litigation going on today regarding the issue of patents (specifically software and 'business method' patents), I couldn't help but wonder if there was a rigorous definition of 'non-obvious'. After searching a bit on the internet (see references) I have come up with an additional set of criteria that can be used to judge obviousness, one that is perhaps more broad than the standard definition. But first, the standard definition (4): A patent may not be obtained ... if the differences between the subject matter sought to be patented and the prior art are such that the subject matter as a whole would have been obvious at the time the invention was made to a person having ordinary skill in the art to which said subject matter pertains. Patentability shall not be negatived by the manner in which the invention was made. The main problem with this core definition is that it is quite vague, and thus it is a 'loose' interpretation of obvious. Often times the patent office must utilize 'prior art', which is often time-consuming, and I believe, unnecessary. Even IF a patent examiner can establish obviousness, it may rebutted (6): Ways to Rebut a prima facia case of Obviousness: Show failure of others Show longfelt, unsolved need Show art "teaches away" from the claimed invention Commercial success Unexpected results/properties Copied by others Evidence of state of the art These factors and many others (the whole concept of 'business method' patents for one), all contribute to many 'obvious' patents being issued, though these patents do not fall under the standard 'non-obvious' definition. II. Short-Term Monopolies Before we consider what a better way of definining 'non-obvious' is, we must first consider what the fundamental reason patents are issued in the first place: (1) Per the word's original definition, one theory of patent legislation is to induce the inventor to disclose knowledge for the advancement of society in exchange for a limited period of exclusivity. Therefore a patent is a monopoly right. By this definition, patents are 'short-term monopolies' which are granted to advance society. Accordingly, anything that even remotely comes CLOSE to obvious should not be patentable, since society will not be advanced by introducing monopolies willy-nilly. Since patents are granted for a future term of 20 years, another way of asking the question is, "will this idea be obvious within 20 years?". If the answer is no, then the idea is non-obvious. If the answer is 'probably yes', then the patent should NOT be granted, since for the time duration after which it becomes obvious, society will not be advanced, but rather retarded. This is a very important concept, one which I feel requires another look. I believe that the absence or presence of 'prior-art', while necessary, is definitely not sufficient to assess the validity of a patent. The reason being is that a monopoly is being granted on this idea for a lengthy term of 20 years! How much innovation has happened in the past 20 years? Imagine the waste if an obvious patent were granted, it could stifle creativity and innovation in a particular field for 20 years! This is definitely not what the patent apparatus was intended for. The ONLY reason to grant a patent is for the advancement of society. 'Fairness', 'Ownership of ideas', and all that junk is nonsense. No one has the 'right' to own an idea! It is a ludicrous concept! The question now, is how to repair our broken patent system with as little change as possible? A complete overhaul is obviously not feasable, and not even required. In fact, I believe that the patent system can be fixed with a small modification to the way that 'obviousness' is determined. Thus, I propose a better system for establishing the 'obviousness' of a patent, grounded in reason and logic. III. A Logical System to Determine Obviousness In determining obviousness, we must not only be able to establish an estimate of today's obviousness, we must be able to estimate the obviousness of an invention in the future (up to 20 years in the future, since that is the term that patent-monopolies are granted). In order to do this, I propose a logical system which can be used to help to determine whether or not a patent is 'obvious'. The basic idea is to try to generate a set of preconditions under which the invention covered by the patent would logically follow. If this can be done, then the invention is obvious with respect to those preconditions. That is, if someone were to satisfy these preconditions, then the particular invention would most likely follow. The eventual goal is to show that an invention is absolutely obvious. That is, all of the inventions' preconditions are obvious, and it logically follows from those preconditions, rendering it obvious. A. Preconditions So what exactly is a precondition? A precondition may be: The problem definition. The availability of a tool or resource. A specific environmental condition. A tool may be a software or hardware tool, or a set of knowledge that a person has. In this way, we can expand on the old definition of 'a person having ordinary skill in the art', by defining exactly what skill and tool set would be required to allow an invention to be created. The problem definition establishes the context for an invention. There will generally be very few of these in a precondition set, probably only one. For example, if we were to try to generate a set of preconditions from which a 'one-click' purchase patent application were submitted, we might come up with the following: Had a desire to make money. Ran a shopping internet site. Discovered that most people leave the site immediately upon encountering that lengthy forms must be filled out before each purchase, possibly resulting in lost revenue. Then I would claim that simplifying the purchase process in the form of a 'one-click' purchase mechanism is obvious with respect to these preconditions. And the simplest purchase process is done with a single action on behalf of the user (be it a key press, mouse movement, mouse click, or when voice recognition comes along, with a single syllable, probably 'buy'). B. Logical Integrity In considering this example, two further questions emerge: when does an invention 'logically follow' from a set of preconditions? how 'obvious' are the preconditions themselves? That is, we need a method for determining WHEN a set of preconditions is likely to present themselves to a user of 'ordinary skill' in the art. 1. Of Logical Connectivity In the previous example, it would seem that the first two items are much more general than the last one. The last one is highly specific. This last one is the problem definition, also referred to as the 'converging' element, that which causes the set of preconditions to converge on the invention, in this case, that of 'simplifying of purchase process'. Without this last condition, there are numerous other optimizations that may be done to the website to make it 'better'. None of which may converge on the 'simplification of the purchase process' element. This 'converging' element (or elements) is REQUIRED in order for the invention to logically follow from the preconditions. The best way to show that an invention logically follows from a set of preconditions such as this is probably proof by contradiction. That is, given the preconditions, attempt to find another direction in which the inventor can go and NOT encounter the invention in question. In this case, it is hard to imagine an entity that met these preconditions that would not create the invention in question. Once the examiner is satisfied with the logical integrity of these preconditions to the invention, then he may examine each precondition in turn to determine its 'obviousness'. This involves examining each precondition in turn and determining what set of criteria could lead to it. 2. Precondition Introspection In determining preconditions for an invention, we must also establish the most likely time (in the future, or already present) that those preconditions will be met by a person of 'standard' skill in the art. An invention may have different sets of preconditions, some of which are likely to happen sooner than others. The soonest, most likely set of preconditions is defined to be the minimal precondition set, or MPS. If the MPS is likely to occur within the first half of the monopoly interval, then it is likely to be less cost effective (in terms of gain to society) to grant the patent (see appendix I for a discussion of costs of inventions). However, if the MPS occurs in the second half of the monopoly interval, then the net gain to society will likely be positive, since the term of advancement would exceed the term of retardation. Putting this issue aside, for now, we can investigate in detail the process of analyzing preconditions. If the precondition set is so specific as to imply the invention and the preconditions themselves are obvious, then the invention is absolutely obvious, as anyone who encounters the problem with the set of tools will definitely create the same invention and the preconditions are easily met. Of course, this may not be the MPS, since it may be possible to come up with a simpler set of tools or problem definition. In our example above, the invention is obvious with respect to the preconditions, but the preconditions themselves are not necessarily obvious, particularly the last one. Thus, we have not yet determined that this invention is obvious. In our example, once we are satisfied that the first two points are obvious, we must further examine the last point. I would claim that the last point has the following further preconditions: Revenue growth is slowing, prompting decision makers to optimize their system. The entity must be reasonably successful to have money and time to optimize their system. The realization that there is a non-trivial discrepancy between the number of unique visitors to the site and the number of purchasers, and that significant income can be generated not by increasing the number of unique visitors to the site, but by increasing the number of visitors who BUY something. In this case, the last item may be sufficient to prompt the discovery that most people leave the site immediately before buying, but one would have a tougher time making that argument. Analysis would continue in this fashion until the examiner were convinced that all the points are absolutely obvious, or that they would be absolutely obvious within half of the monopoly interval. If, anywhere in this process, a precondition is found to be NOT obvious, then the whole precondition set, and the patent is not obvious. In this way, it requires a 'flash of genius' to be present. However, a DIFFERENT precondition set may be found in which all the preconditions are obvious. In this case, the patent should still be deemed obvious. 3. Alternate Solutions There may also be various precondition sets, and it may be difficult to determine which is the MPS. It is not necessary to find the MPS, but by definition, it will be easiest to show that the MPS is absolutely obvious. In order to better determine what the MPS is, we can can ask several questions: What IS the most obvious solution for the given precondition set? Eg where does a given preconditon set converge? If there is more than one obvious solution, what are the differences? If we vary the precondition set slightly (either alter the problem definition in likely ways, or add/remove tools), can we create a precondition set in which the invention under question is the most obvious solution? IV. Logistics How should such a system be implemented? Once the precondition sets are created to the satisfaction of the examiner, they can be 'polled' out to 'ordinary workers of the art' to 'test' how obvious the invention may be. Of course, since the preconditions are generated from the invention, care must be taken to ensure that the preconditions themselves (and their particular combination) are obvious, and bias is not introduced at this stage. The management and creation of the MPS can be done via a set of confidential peers, and involving the patent applicant. The initial application must contain what the applicant believes is the MPS, and he must show that it will first occur after half the monopoly interval has expired. Thus, the work on the patent examiner may be reduced. Of course, as indicated above, care must be taken on behalf of the patent examiner to consider alternate precondition sets to see if they may be the MPS. The invention may be submitted to a set of non-competitive confidential peers so that the patent examiner can draw upon their experience in the field. One good starting point is utilizing genetic programming to duplicate an invention with a given set of preconditions. If this is possible, then there is no 'flash of genius', and the solution to the problem necessarily follows from the preconditions. Appendix I. Costs of Inventions Over Time Determining the cost/benefit of granting a patent could be quite complex, but for now, we are assuming a linear cost structure. If this assumption is incorrect, then the estimates of when a patent should be allowed based on its MPS point will have to be adjusted. Thus, we are making the assumption that there is a fixed benefit to society for a particular invention per unit time, and consequently, the total benefit over a period of time will be linearly increasing. Furthermore, the time between the expected invention date and the actual invention date will be refered to as the innovation/disinnovation interval. If the expected invention date is after the actual invention date, then society has benefitted from the extra time that the invention has been existence. Likewise, if the expected invention date is BEFORE the actual invention date, then society has suffered as a result of NOT having the invention available when it was 'expected' to be. Thus, we can see that for any exclusion interval (eg a patent) that covers a time interval t, then if the expected invention date is in the last half (or after), there is a net benefit to society, and if the expected invention date is in the first half, then there is a net loss to society. Since the term of a patent interval is fixed at 20 years, we can see that if the expected invention date for a particular invention is today, and we grant the patent, then we are losing out on 20 full years of innovation that could be done as a consequence of that invention! please email me, zog if you have any questions or comments, I would love to hear them! Especially with regard to problems with this system. References: wordiq.com iusmentis.com inventors.about.com uspto.gov law.udayton.edu cyber.law.harvard.edu lpf.ai.mit.edu

Non-Obvious Patents

Home
Resources
Essays
Hobbies
Hockey!
Projects
Resume

(December 20, 2005, by Thomas Fischer, work in progress)

I. Obviousness in Patents

The patent system in the United States is flawed. There are far too many patents that are granted which are 'obvious'. In seeing all the litigation going on today regarding the issue of patents (specifically software and 'business method' patents), I couldn't help but wonder if there was a rigorous definition of 'non-obvious'. After searching a bit on the internet (see references) I have come up with an additional set of criteria that can be used to judge obviousness, one that is perhaps more broad than the standard definition.

But first, the standard definition (4):

A patent may not be obtained ... if the differences between the subject matter sought to be patented and the prior art are such that the subject matter as a whole would have been obvious at the time the invention was made to a person having ordinary skill in the art to which said subject matter pertains. Patentability shall not be negatived by the manner in which the invention was made.

The main problem with this core definition is that it is quite vague, and thus it is a 'loose' interpretation of obvious. Often times the patent office must utilize 'prior art', which is often time-consuming, and I believe, unnecessary. Even IF a patent examiner can establish obviousness, it may rebutted (6):

Ways to Rebut a prima facia case of Obviousness:

Show failure of others
Show longfelt, unsolved need
Show art "teaches away" from the claimed invention
Commercial success
Unexpected results/properties
Copied by others
Evidence of state of the art

These factors and many others (the whole concept of 'business method' patents for one), all contribute to many 'obvious' patents being issued, though these patents do not fall under the standard 'non-obvious' definition.

II. Short-Term Monopolies

Before we consider what a better way of definining 'non-obvious' is, we must first consider what the fundamental reason patents are issued in the first place: (1)

Per the word's original definition, one theory of patent legislation is to induce the inventor to disclose knowledge for the advancement of society in exchange for a limited period of exclusivity. Therefore a patent is a monopoly right.

By this definition, patents are 'short-term monopolies' which are granted to advance society. Accordingly, anything that even remotely comes CLOSE to obvious should not be patentable, since society will not be advanced by introducing monopolies willy-nilly. Since patents are granted for a future term of 20 years, another way of asking the question is, "will this idea be obvious within 20 years?". If the answer is no, then the idea is non-obvious. If the answer is 'probably yes', then the patent should NOT be granted, since for the time duration after which it becomes obvious, society will not be advanced, but rather retarded.

This is a very important concept, one which I feel requires another look. I believe that the absence or presence of 'prior-art', while necessary, is definitely not sufficient to assess the validity of a patent. The reason being is that a monopoly is being granted on this idea for a lengthy term of 20 years! How much innovation has happened in the past 20 years? Imagine the waste if an obvious patent were granted, it could stifle creativity and innovation in a particular field for 20 years!

This is definitely not what the patent apparatus was intended for. The ONLY reason to grant a patent is for the advancement of society. 'Fairness', 'Ownership of ideas', and all that junk is nonsense. No one has the 'right' to own an idea! It is a ludicrous concept! The question now, is how to repair our broken patent system with as little change as possible? A complete overhaul is obviously not feasable, and not even required. In fact, I believe that the patent system can be fixed with a small modification to the way that 'obviousness' is determined. Thus, I propose a better system for establishing the 'obviousness' of a patent, grounded in reason and logic.

III. A Logical System to Determine Obviousness

In determining obviousness, we must not only be able to establish an estimate of today's obviousness, we must be able to estimate the obviousness of an invention in the future (up to 20 years in the future, since that is the term that patent-monopolies are granted).

In order to do this, I propose a logical system which can be used to help to determine whether or not a patent is 'obvious'. The basic idea is to try to generate a set of preconditions under which the invention covered by the patent would logically follow. If this can be done, then the invention is obvious with respect to those preconditions. That is, if someone were to satisfy these preconditions, then the particular invention would most likely follow. The eventual goal is to show that an invention is absolutely obvious. That is, all of the inventions' preconditions are obvious, and it logically follows from those preconditions, rendering it obvious.

A. Preconditions

So what exactly is a precondition?

A precondition may be:

The problem definition.
The availability of a tool or resource.
A specific environmental condition.

A tool may be a software or hardware tool, or a set of knowledge that a person has. In this way, we can expand on the old definition of 'a person having ordinary skill in the art', by defining exactly what skill and tool set would be required to allow an invention to be created.

The problem definition establishes the context for an invention. There will generally be very few of these in a precondition set, probably only one.

For example, if we were to try to generate a set of preconditions from which a 'one-click' purchase patent application were submitted, we might come up with the following:

Had a desire to make money.
Ran a shopping internet site.
Discovered that most people leave the site immediately upon encountering that lengthy forms must be filled out before each purchase, possibly resulting in lost revenue.

Then I would claim that simplifying the purchase process in the form of a 'one-click' purchase mechanism is obvious with respect to these preconditions. And the simplest purchase process is done with a single action on behalf of the user (be it a key press, mouse movement, mouse click, or when voice recognition comes along, with a single syllable, probably 'buy').

B. Logical Integrity

In considering this example, two further questions emerge:

when does an invention 'logically follow' from a set of preconditions?
how 'obvious' are the preconditions themselves? That is, we need a method for determining WHEN a set of preconditions is likely to present themselves to a user of 'ordinary skill' in the art.

1. Of Logical Connectivity

In the previous example, it would seem that the first two items are much more general than the last one. The last one is highly specific. This last one is the problem definition, also referred to as the 'converging' element, that which causes the set of preconditions to converge on the invention, in this case, that of 'simplifying of purchase process'. Without this last condition, there are numerous other optimizations that may be done to the website to make it 'better'. None of which may converge on the 'simplification of the purchase process' element. This 'converging' element (or elements) is REQUIRED in order for the invention to logically follow from the preconditions.

The best way to show that an invention logically follows from a set of preconditions such as this is probably proof by contradiction. That is, given the preconditions, attempt to find another direction in which the inventor can go and NOT encounter the invention in question. In this case, it is hard to imagine an entity that met these preconditions that would not create the invention in question.

Once the examiner is satisfied with the logical integrity of these preconditions to the invention, then he may examine each precondition in turn to determine its 'obviousness'. This involves examining each precondition in turn and determining what set of criteria could lead to it.

2. Precondition Introspection

In determining preconditions for an invention, we must also establish the most likely time (in the future, or already present) that those preconditions will be met by a person of 'standard' skill in the art. An invention may have different sets of preconditions, some of which are likely to happen sooner than others. The soonest, most likely set of preconditions is defined to be the minimal precondition set, or MPS. If the MPS is likely to occur within the first half of the monopoly interval, then it is likely to be less cost effective (in terms of gain to society) to grant the patent (see appendix I for a discussion of costs of inventions). However, if the MPS occurs in the second half of the monopoly interval, then the net gain to society will likely be positive, since the term of advancement would exceed the term of retardation. Putting this issue aside, for now, we can investigate in detail the process of analyzing preconditions.

If the precondition set is so specific as to imply the invention and the preconditions themselves are obvious, then the invention is absolutely obvious, as anyone who encounters the problem with the set of tools will definitely create the same invention and the preconditions are easily met. Of course, this may not be the MPS, since it may be possible to come up with a simpler set of tools or problem definition. In our example above, the invention is obvious with respect to the preconditions, but the preconditions themselves are not necessarily obvious, particularly the last one. Thus, we have not yet determined that this invention is obvious.

In our example, once we are satisfied that the first two points are obvious, we must further examine the last point. I would claim that the last point has the following further preconditions:

Revenue growth is slowing, prompting decision makers to optimize their system.
The entity must be reasonably successful to have money and time to optimize their system.
The realization that there is a non-trivial discrepancy between the number of unique visitors to the site and the number of purchasers, and that significant income can be generated not by increasing the number of unique visitors to the site, but by increasing the number of visitors who BUY something.

In this case, the last item may be sufficient to prompt the discovery that most people leave the site immediately before buying, but one would have a tougher time making that argument. Analysis would continue in this fashion until the examiner were convinced that all the points are absolutely obvious, or that they would be absolutely obvious within half of the monopoly interval.

If, anywhere in this process, a precondition is found to be NOT obvious, then the whole precondition set, and the patent is not obvious. In this way, it requires a 'flash of genius' to be present. However, a DIFFERENT precondition set may be found in which all the preconditions are obvious. In this case, the patent should still be deemed obvious.

3. Alternate Solutions

There may also be various precondition sets, and it may be difficult to determine which is the MPS. It is not necessary to find the MPS, but by definition, it will be easiest to show that the MPS is absolutely obvious. In order to better determine what the MPS is, we can can ask several questions:

What IS the most obvious solution for the given precondition set? Eg where does a given preconditon set converge?
If there is more than one obvious solution, what are the differences?
If we vary the precondition set slightly (either alter the problem definition in likely ways, or add/remove tools), can we create a precondition set in which the invention under question is the most obvious solution?

IV. Logistics

How should such a system be implemented? Once the precondition sets are created to the satisfaction of the examiner, they can be 'polled' out to 'ordinary workers of the art' to 'test' how obvious the invention may be. Of course, since the preconditions are generated from the invention, care must be taken to ensure that the preconditions themselves (and their particular combination) are obvious, and bias is not introduced at this stage.

The management and creation of the MPS can be done via a set of confidential peers, and involving the patent applicant. The initial application must contain what the applicant believes is the MPS, and he must show that it will first occur after half the monopoly interval has expired. Thus, the work on the patent examiner may be reduced. Of course, as indicated above, care must be taken on behalf of the patent examiner to consider alternate precondition sets to see if they may be the MPS. The invention may be submitted to a set of non-competitive confidential peers so that the patent examiner can draw upon their experience in the field.

One good starting point is utilizing genetic programming to duplicate an invention with a given set of preconditions. If this is possible, then there is no 'flash of genius', and the solution to the problem necessarily follows from the preconditions.

Appendix I. Costs of Inventions Over Time

Determining the cost/benefit of granting a patent could be quite complex, but for now, we are assuming a linear cost structure. If this assumption is incorrect, then the estimates of when a patent should be allowed based on its MPS point will have to be adjusted. Thus, we are making the assumption that there is a fixed benefit to society for a particular invention per unit time, and consequently, the total benefit over a period of time will be linearly increasing.

Furthermore, the time between the expected invention date and the actual invention date will be refered to as the innovation/disinnovation interval. If the expected invention date is after the actual invention date, then society has benefitted from the extra time that the invention has been existence. Likewise, if the expected invention date is BEFORE the actual invention date, then society has suffered as a result of NOT having the invention available when it was 'expected' to be.

Thus, we can see that for any exclusion interval (eg a patent) that covers a time interval t, then if the expected invention date is in the last half (or after), there is a net benefit to society, and if the expected invention date is in the first half, then there is a net loss to society.

Since the term of a patent interval is fixed at 20 years, we can see that if the expected invention date for a particular invention is today, and we grant the patent, then we are losing out on 20 full years of innovation that could be done as a consequence of that invention!

please email me, zog if you have any questions or comments, I would love to hear them! Especially with regard to problems with this system.

References: