 |
PageBox |
| Support | Map |
|
|
|
|
|
|
|
|
A business view of software and business method patents
The patentability of software and business method inventions has been widely questioned on legal and economical grounds. Software and business method patents have been accused to protect abstract ideas and to impede competition, increase costs and reduce offer. In this article, we consider the problem of software and business method inventions under another angle. We see how and why companies and inventors file software and business method patents. We find that this is not a widely adopted practice and that those who spend time and money to try patenting software and business method inventions do so to avoid unfair competition or to have the same weapons as their competitors. We also compare software and business method patents to other patents to find that they not less legitimate or tangible. Software and business method patents differ from other patents by the amount of prior art, the number of potential infringers and frequently by the difficulty to identify infringements, which are technical but not legal issues. We eventually find that the software industry faces the complex challenge of managing its innovations and that the public interest is to help this industry to address this challenge through adequate protection of inventions.
In this contribution to the debate about software and business method patentability, instead of considering what the patent system enables or should enable from a legal perspective, we analyze business practices and motivations in four steps:
we define what we mean by software and business method inventions;
we compare software and business method patents with patents that work in a domain, the drugs, whose patentability is not disputed; taking the example of Vicuron patents;
we explain how software and business method inventions are made and patented, stressing on differences with other domains in inventive step, motivation and context;
we present the current situation of software and business method patenting stressing on cost and adoption issues;
we consider the motivations of people who file software and business method patent applications and see in the different cases whether the public interest is to grant patents for such applications.
We use the rules of the Patent Cooperation Treaty (PCT) procedure though national rules can be less strict. We consider:
the search requirement; though an international search report is a non-binding opinion;
the feasibility requirement;
the industrial applicability requirements.
Then we give definitions to software and business method inventions.
An invention should be patentable only if the International Searching Authority is able to perform a prior art search for said invention. Therefore an invention has to conform to the PCT and notably to Article 17 and to Rule 39.
This principle is enforced by the European Patent Office that considers that a search report is an examination prerequisite.
Rule 39 excludes the following types of subject matter:
scientific and mathematical theories;
schemes, rules or methods of doing business, performing purely mental acts or playing games;
mere presentations of information;
computer programs to the extent that the International Searching Authority is not equipped to search prior art concerning such programs.
The three first exclusions relate to the “useful art” limitation in the American Constitution statement “Congress shall have power ... to promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries.” The Article 52 of the European Patent Convention includes the limitations of Rule 39 and explains that these limitations “shall exclude patentability of the subject-matter or activities referred to in [these limitations] only to the extent to which a European patent application or European patent relates to such subject-matter or activities as such.” An invention whose subject matter is computing or a method of doing business is not patentable but
the use of computer programs by an invention does not bar patenting;
a process or system that uses computer programs to implement a method of doing business is tangible and therefore can be patented.
The three first exclusions of rule 39 also relate to statutory subject matters. In USA an invention is found to have a statutory subject matter if said invention is in technological arts and if it produces a useful, concrete and tangible result [Toma test.]
Technological arts relate to the “useful arts” of the American Constitution. The technological art limitation is defined by the Article 101 of the United States Code (USC) that says: “whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefore” and aims to limit patent protection to the field of applied technology. So the first step under the Toma test was to check if an invention was either a process or a machine or a manufacture or a composition of matter. The second step under the Toma test was to check whether the invention was applied to physical elements (whether the invention physically transformed the subject matter, data for a software invention.) So the stress was on concreteness and tangibility. To further clarify:
Software and business method inventions are processes or machines.
Said processes (processes, arts, methods) are acts or series of acts performed upon the subject matter to be transformed and reduced to a different state or thing.
We use the term “invention”. In the literature you may find the term “system”. Toma made clear that what is tested is the claimed subject matter. An examiner rejects claims and not inventions. The claimed subject matter comprises what is described in the claim and the description matter supporting this claim.
The Toma test was almost abolished by the State Street decision, in which the court wrote "the question of whether a claim encompasses statutory subject matter should not focus on which of the four categories of subject matter a claim is directed to (process, machine, manufacture, or composition of matter) but rather on the essential characteristics of the subject matter, in particular, its practical utility." The stress was on utility, which also impacted the second step of the Toma test. An invention manipulating data into a form that can be relied upon by subsequent users or for other business processes was patentable.
The Toma test was restored in part by the Bowman decision. Invention no longer had to belong to a process or machine or manufacture or composition of matter category but they had to use technological means. Inventions claimed in Toma and State Street were patentable because they claimed operations performed by computers. So examiners used a “mental steps” test, which consisted in checking whether the invention steps could be carried out in the human mind.
The “mental steps” doctrine was recently rejected by the USPTO bureau of appeals in the Lundgren decision. This decision was reported on a patent law blog. After Lundgren examiners checked:
whether the invention steps could be carried out in the human mind (“mental steps” test determining whether the invention manipulates technology); and if this is the case
whether the invention transforms or reduces a tangible matter into another state of things (concrete and tangible result.)
An invention that can be performed manually by a human and that does not transform or reduce a tangible matter into another state of things does not have a statutory subject matter. This practice is further described in the Manual of Patent Examining Procedure, section 2106 (Patentable Subject Matter - Computer-Related Inventions), part IV (B) (2) (b) entitled “Statutory Process Claims” that explains that:
a claim that requires one or more acts to be performed defines a process;
to be statutory, a claimed computer-related process must either: (A) result in a physical transformation outside the computer for which a practical application in the technological arts is either disclosed in the specification or would have been known to a skilled artisan, or (B) be limited to a practical application within the technological arts.
New guidelines of USPTO reject the “mental steps” test. An invention is patentable “if the claimed invention physically transforms an article or physical object to a different state or thing, or if the claimed invention otherwise produces a useful, concrete, and tangible result.” An invention produces a useful invention if the invention’s utility is specific, substantial and concrete. A concrete result is a repeatable result. A tangible result is a result that is not abstract.
This story illustrates the difference between the law for which every invention is patentable unless it is nothing more than an abstract idea, a law of nature or a natural phenomenon and the patent office need which is to provide a test that can be applied consistently and be understood by the public to decide when an invention is patentable.
The fourth exclusion of rule 39 introduces a feasibility concern. Search feasibility is implied by the definition of invention. An invention being an invention if it is novel and non obvious it should not be possible to grant patents to inventions whose novelty and non-obviousness cannot be checked. But search feasibility does not mean that search should be feasible at any cost but that the sum of the search cost, of the cost of search failures and of the costs related to the uncertainty of search results has to be lower than the benefit generated by patents.
Patents can be granted only to what can be perceived as a progress by the public. There should be enough people interested to learn about an invention or about its effects, to motivate a written publication. Had an invention been invented before, either said first invention had to be necessarily disclosed in a written publication or, if said first invention was kept secret, its advantageous effects had to be described in a written publication. This part of the feasibility requirement follows from the usefulness patentability condition.
It would certainly be difficult to find traces of search cost concern in the Paris Convention for the Protection of Industrial Property or in the American Constitution or in the British Statute of Monopolies of 1623. But the very idea of trading correctness for practicability was also absent in the industry up to 1940s. There were things that could be found or calculated and things that had to be decided in a rather empirical way. Then operation research and computers created an approximate knowledge domain in which one get the accuracy one can pay. If something cannot be determined at a reasonable cost with useful accuracy the industry does not deny the existence of said thing, it just ignores said thing.
As said in Article 27 of the Agreement on Trade-Related aspects of Intellectual Property rights (TRIPs) “patents shall be available for any inventions, whether products or processes, in all fields of technology, provided that they are new, involve an inventive step and are capable of industrial application.”
An invention capable of industrial application is an invention that can be made or used (in the technological sense) in any kind of industry. "Industry shall be understood in its broadest sense and shall apply not only to industry and commerce proper, but likewise to agricultural and extractive industries and to all manufactured or natural products” as explained in the Paris Convention for the Protection of Industrial Property.
We define a software patent as a patent granted to an invention that uses computer programs to produce a technical result and for business methods we take the “modern business data processing” definition from the USPTO class 705: “class for apparatus and corresponding methods for performing data processing operations, in which there is a significant change in the data or for performing calculation operations wherein the apparatus or method is uniquely designed for or utilized in the practice, administration, or management of an enterprise, or in the processing of financial data. This class also provides for apparatus and corresponding methods for performing data processing or calculating operations in which a charge for goods or services is determined.”
The difference between a software invention and business method invention is that the former produces a technical result whereas the latter produces a business result. An applicant aims to protect its inventions. If an invention produces a technical result the applicant file a software application; if an invention produces a business result the applicant file a business method application; but the purpose and the drafting process are the same in both cases. Therefore in a business perspective these two types of applications do not need to be distinguished.
On June 17, 2005 Paul Elias reported in a paper entitled "Giant Drug Maker Pfizer To Pay $1.9 Billion for Biotech” that Pfizer Inc.'s “would pay $1.9 billion in cash for a tiny company that makes a new breed of antibiotics.” Paul Elias explains in this paper that drug makers facing the expiration of their patents acquire other drug companies owning patents for drugs nearing the market. This news illustrates the fact that in a mature industry Intellectual Property is not only a means to be awarded substantial damages in a lawsuit or a settlement but also:
a motivation for acquisitions and mergers (see for instance General Electric) like economies of scale and brand;
a source of revenue through licensing (see for instance Pilkington brothers for float glass.)
We use the example of Vicuron patents to show that claims of drugs’ patents are in no way more legitimate than claims of software and business method patents. At the time of writing Vicuron is granted four patents by the US Patent and Trademark Office (USPTO):
Patent # | Title |
|---|---|
Injectable formulations containing ramoplanin | |
Succinate compounds, compositions and methods of use and preparation | |
Urea compounds, compositions and methods of use and preparation | |
Methods of administering dalbavancin for treatment of bacterial infections |
The patent has forty four claims. Only the first claim is independent. This claim relates to a composition.
The patent has forty eight claims. Only the first claim is independent. This big claim relates to a composition and extensively uses Markush groups ("R.sub.9, R.sub.9a and R.sub.10 are independently selected from the group consisting of..."). It first claims a molecule:
Then it defines the composition of R1 to R4, R6 and R7 and Y using Markush groups. So claim 1 actually claims a family of molecules.
The patent has twenty six claims. Seven claims, 1, 12, 13, 14, 18, 23 and 24 are independent. Said independent claims are about compositions and extensively use Markush group ("R.sub.10 is selected from the group consisting of hydrogen, ...) Some dependent claims are method claims:
"16. A method of treating a prokaryotic microbial infection in a mammal which comprises administration to said mammal of a therapeutically effective amount of a compound of any one of claims 1 to 14.
17. The method of claim 16 wherein the microbial infection is a bacterial infection."
The patent has twenty two claims. Two claims, 1 and 17 are independent. Claim 1 reads:
"A method for treating a bacterial infection in a human in need thereof, the method comprising:
administering initial and subsequent therapeutically effective doses of dalbavancin in a pharmaceutically acceptable carrier to the patient,
wherein each dose is separated by five to ten days and wherein the amount of the initial dose is about 500 mg to about 5000 mg, and wherein the amount of the initial dose is at least about two times the amount of dalbavancin contained in the subsequent dose."
Claim 17 reads:
"A method for treating a bacterial infection in a human in need thereof, the method comprising:
administering initial and subsequent therapeutically effective doses of dalbavancin in a pharmaceutically acceptable carrier to the patient,
wherein each dose is separated by about one week and wherein the amount of the initial dose is about 1000 mg and the amount of each subsequent dose is about 500 mg."
The invention claims a method of administering a known antibiotic in a reasonably convenient way.
Drug patents of Vicuron differ from software and business patents in that they rely mainly on composition claims in which a family of molecules (a genus in USPTO wording) is defined. Because of the use of Markush groups molecule families can be very large. Manuals of examination define rules to determine:
whether a genus can be accepted (definiteness);
whether the description supports such generic claims.
The tangibility of such patents does not make prior art search easier:
Given the size of a molecule genus how to demonstrate that:
None of the billions of plants and animals produces a molecule in claimed geni?
No healer ever used such molecule to cure a related disease for 50,000 years?
As far as drug patents are concerned the prior art of healers is out of the scope. Discoverers and potential infringers are companies with the resources:
to analyse molecules;
to produce molecules in an industrial way;
to perform tests according to defined protocols.
Such criteria do not exist for software and business method patents. Every person of the art is a potential infringer, can be the first inventor and can describe an invention in essentially the same way as largest corporations.
Method claims of drug patents (6,852,752 and 6,900,175) are rather more obvious ("administering initial and subsequent therapeutically effective doses of dalbavancin in a pharmaceutically acceptable carrier to the patient", "A pharmaceutical composition comprising a therapeutically effective amount of a compound of any one of claims 1 to 14 and a pharmaceutically acceptable excipient”) than business method claims. Again it does not matter because few companies define how to administrate drugs.
Drug and software/business method patents differ
by the population of potential inventors/infringers, tens for drug patents, thousands for software/business method patents;
in that drug claims (composition and administering) are closely related to the deliverables of the drug maker (drug and notice). Such link does not necessarily exist for business method and software patents. The issue is not a matter of concreteness. Though it is concrete, a patented process for producing tyres that cannot be deduced from the deliverable (tyre) analysis would face the same enforce-ability issue as software patents if there were more than a dozen of tyre makers.
A patent can be enforced
either if the result of practicing the patented invention is public or;
if the invention is useful only when practiced publicly and said public practice is easy to identify.
As illustrated by the Vicuron example, manufacturing, chemistry and drug patents usually exhibit at least one of these properties:
There are relatively few companies able to produce a drug or a car [1]. In manufacturing, chemistry or drug industries companies know their competitors and analyze what said competitors are doing. Patents (innovation) is only one part of the mix to monitor, mix that also comprises the expression (the way innovations are combined with known art in the product or in the way the product is produced) and the way the product is marketed [2]. In these industries companies can effectively identify competitors’ patent infringements.
Patents in these industries protect the composition of the product, the way of using the product or the way of producing it. To identify infringements companies can buy competitors’ products, read their manuals and so forth.
whereas software and business method inventions do not necessarily exhibit any of these properties.
A product is divided in three parts:
The expression, what is original without contributing to the product function. Colours and font sizes in a user interface, column names in a database, web service names and parameters are expression.
Scenes a faire. What would be made by any person of the art given the product requirements.
Inventions. What is novel, useful for everybody and could have been made in another way.
As depicted below developers create products
with a motivation that follows from requirements;
with skills and knowledge that follow from their education and experience;
using a development method, within a managed project;
trying to reuse existing components and known patterns;
trying to make a product better than what already exists, able to be operated in conjunction with existing means, while designing around known patents and reproducing means that prove to be effective (external world.)
The development methods and processes, and the developers’ education and experience
allow selecting effective solutions;
facilitate communication in and between involved teams;
minimize development risks and allow estimating the development effort.
A project, properly planed and funded, involving adequately educated and experienced developers and using appropriate development methods and processes should produce a minimum of inventions, said inventions following from requirements, for which no adequate solution was found so far. Other inventions are either failures [3] of the project planning:
to reuse existing components and design patterns
to minimize development risk
to estimate the development effort: it is impossible to estimate the cost for implementing the invention before making said invention,
or arise because, depending on the project size, there will be more or less cases in which:
someone find that a certain apparatus is more convenient or flexible, yielding surprising, unexpected results;
a task implemented with a known art solution is too slow or uses too much resources; causing someone to search and find a better solution.
Companies decide about the amount of inventions they can generate with product development. More predictable a development effort is, less inventions this development is likely to generate. The number and importance of inventions depends as much on the predictability as on the size of the project.
If a company choose to develop product A whose development effort is highly predictable this development can generate a couple of minor inventions. If said company choose to develop product B whose development outcome and cost is uncertain this development can generate several major and minor inventions.
This is a difference between the software industry and industries with a mature patent protection. In industries with mature patent protection companies also have methods and procedures, promote reuse and hire knowledgeable and experienced people but they accept high percentages of complete and partial failures.
The software industry suffers from hasty comparisons with other industries in which it was found that product development almost always experienced overspending, delays and quality issues. Developing a program is not like building a house or making a car. Software industry is an industry that produces prototypes and in any industry (aircraft, car...) that makes prototypes said prototypes either have defaults, cost more than expected and do not work properly on time, or are not technically innovative. However software industry faces the challenge of recognizing when a development is ambitious when a poorly planed and managed development of a traditional product faces the same difficulties as a carefully planed and managed development of a leading-edge product.
Companies expect from governments that they provide adequate protection for their products: “We spent this amount of money and time to develop and sell this product and want to exclude others from copying it.” Addressing such request would require a suis generis protection, a sort of extended copyright that governments are reluctant to provide.
The copyright law grants to the copyright owner an exclusive right to reproduce the expression of an original work of authorship but does not protect the discoveries, ideas, concepts, principles or potentially inventive procedures, processes, systems, methods of operation that helped producing said work of authorship. Whatever in the original work of authorship can be deduced from its purpose, subject... is also excluded from copyright protection. Rules and doctrines help identifying what is excluded from copyright protection:
An originality rule: what can be protected by copyright is what was independently created by the author and possesses some degree of creativity.
A merger doctrine: when the expression is inseparable from the idea it expresses, the expression and the idea merge, and the combined idea-expression is not protected.
A scene a faire doctrine: "scenes a faire are sequences of events necessarily resulting from the choice of setting or situation, or incidents, characters or settings that are indispensable in the treatment of a given topic." For computer programs it means that the expression is not protected "as it is dictated by such factors as (1) the mechanical specifications of the computer on which a particular program is intended to run; (2) compatibility requirements of other programs with which a program is designed to operate in conjunction; (3) computer manufacturers’ design standards; (4) demands of the industry being serviced; and (5) widely accepted programming practices within the computer industry."
So copyright protects only the part of the expression that is original and does not follow from:
ideas,
requirements,
standards and programming practices (methods, procedures, tools), and
compatibility issues (environment).
Therefore a product developed from customers’ requirements with known methods, procedures and tools to be used in a given environment still contain original expression protected by copyright but this protection cannot exclude others from developing a competing product addressing the same requirements in the same environment using the same methods, procedures and tools.
Patent law provides a means to get an exclusive right on inventions belonging to technological arts, that produce useful, concrete and tangible results, and are novel and non obvious. Patent law is not designed to grant a protection to everything that is excluded from copyright. Real inventions contain a core idea, come from a kind of “flash of genius”, and practicing them usually yields surprising and unexpected results.
Developers and patent writers have two options:
Claim the combination of ideas, concepts, principles, procedures, processes, systems, methods of operation practiced by the product. Said combination can be patentable because (1) only the combination has to be novel; each of said ideas, concepts, principles, procedures, processes, systems, methods of operation can be known; (2) the usefulness of the product demonstrates the usefulness of the combination (3) the combination being an original expression is usually not obvious. The patent granted for such combination has a narrow scope and hence a little value.
Identify and patent real inventions in the product. The scope of such patents is wider and such patents can have a higher value.
In practice most software and business method patents are between these two options. They are infected to some degree by expression stuff. This is probably the price to pay for the lack of suis generis protection. Because this expression stuff reduces the value of said software and business method patents this is hardly a ground for banning software and business method patents, even if this practice certainly increases the number of patents.
A software or business method invention:
solves a tangible problem;
using abstract means such as algorithms and models;
using computation means, for instance an operating system, a database, a CPU, a memory or disks;
operating in conjunction with other components, part of the environment such as programs or hardware devices.
This is in no way specific to software and business patents (a drug cures a disease using chemical means according to theories and explanations about how the body works and is designed to interact with said body environment.)
However software and business method patents are harder to understand than other patents because:
software and business methods use artefacts (computing means) and operate in conjunction with artefacts (environment) like inventions in mechanics, and
those artefacts are of increasing complexity.
The increasing complexity of those artefacts is a consequence of progress and notably of past inventions. Each invention, each discovery makes our world more complex and becomes a building block for other inventions and discoveries. Mankind used to only have sticks and stones, then came the wheel, etc. Now artefacts can be as complicated as a revenue management system. The increasing complexity of building blocks affects all patent domains but has more impact on software and business methods patent because they pertain to the domain where progress is the fastest.
Developers who make software and business method inventions belong to a class of skilled workers who are given autonomy and from who is expected ingenuity. But this class is not given a status of scientist like chemistry inventors or a status of artist like architects. Companies do not expect that developers publish articles and prefer that developers remain anonymous. We must stress that this is a matter of social status and not of richness. This class may enjoy safer and higher incomes than scientists or artists. This situation is in no way new. We would like to quote The Labour Theory of Knowledge & Its Corollary: The Knowledge Theory of Capital by Harry Hillman Chartrand:
“From before Elizabeth I and the Statute of Artificers of 1563, [...] labour, especially skilled labour, tended to move where wages were highest. To maintain feudal control, the Statute of Artificers established a statutory apprenticeship system replacing the crumbling medieval one in which journey men increasingly left their masters for greener pastures. The Statute set maximum wage rates, established residency requirements, and was intended to instill a sentiment of subordination. [...]In 1814, however, it was only government enforced restrictions on the movement of labour that were replaced when a ‘voluntary’ apprenticeship system was introduced. [...]In short order, the guild system collapsed and the labour market became flooded with unskilled workers. The successful innovation of the industrial factory system meant, however, that Britain quickly became the lowest cost producer of many goods, especially textiles. The top end of that market, however, continued to be dominated by mills in Lyon, France and Munich, Germany. By 1835 the quality of ‘high end’ British textile production had declined to the point that the British Board of Trade appointed a Select Committee to investigate the problem and recommend remedies. It called for the marriage of art and industry to enhance English competitiveness with European rivals. The result was creation of the first school of design in South Kensington in 1836. It is interesting to note that the curriculum of the new school was designed to insure that no craftsman would ever aspire to become an artist. Artists were gentlemen who attended the Royal Academy of Art (Savage 1985). This distinction between the crafts and the arts continues a much older schism in Western culture between the Mechanical and Liberal Arts.”
Developers are craftsmen in the same way craftsmen were craftsmen before or at the beginning of the Industrial revolution. They pay lip service to patents because they are marginally rewarded for their inventions and because their inventions like their publications do not facilitate their move to better positions. Indeed a patent protecting what they know the best to the benefit of their current employer may reduce their chances to get a better job elsewhere. This attitude does not follow from a technical-only background. Developers are highly reactive to Society proposals, which are the same as in past centuries:
No incentive: move to another job. Consequence: shortage of skilled labour; loss of competitiveness.
Statute: job security, recognition.
Freedom: move where wages are the highest.
Therefore staff developers do not carry their inventions. Instead inventions are noticed by their hierarchy. This has an adverse impact on application quality but this does not mean that there are no inventions in software industry. Problems are as complex as in other domains and the motivation for solving them is the same.
Few companies involved in software development file software and business method patents and check whether they infringe existing patents.
A first reason is that software and business method patents are still new. Rules are still evolving on key aspects such as statutory subject matters and re-examination procedures. People who run companies involved in software development were never taught about patenting and cannot easily find external help. They wait because they do not have the resources to go forward and because the outcome of the current discussions and lawsuits is uncertain.
Beyond this temporary problem there are issues that will remain:
How to be sure that your patent application was not anticipated even if (1) you did not find prior art (2) the patent office did not find prior art either and granted the patent? A patent invalidated as soon as you try to use it has little value.
How to identify infringers? There are two difficulties: (1) the number of potential infringers [4] (2) and for each potential infringer the determination of the infringement. The former difficulty is the same as for trademark infringement. This is a story about monitoring Internet and listening customers and local representatives. The latter difficulty is harder to overcome. When the subject matter of the patent is apparent and useful only when public this is a time-consuming expert work to determine whether a process or method infringes a patent. When you need a product license to practice the process or method of the potential infringer or when the subject matter of the patent is not apparent (this is the case for instance for caching systems that provide a service to more users at a lower cost) you may depend on mistakes of the potential infringer, a very uncertain means.
How to deal with identified infringers? Because settlements and license agreements are not public, it is difficult to evaluate the proportions of settlements, legal actions, cases in which plaintiffs give up, licence agreements, and IP-motivated acquisitions or mergers for software and business method patents [5]. In a sane IP market like the drug industry IP generates license revenues and improves the company valuation (Vicuron example), and litigations are the exception.
The following scenario explains the fear of companies involved in software development that do not manage their intellectual property:
The software company identifies a development opportunity. It analyzes existing products, asks customers for the functions and modus operandi they would like to see in a new product. Then it consolidates the result of this study into a requirement document.
The software company designs and implements a product fulfilling said requirements, using well-known methods and tools, making conservative design choices in order to meet an aggressive schedule.
The software company markets the product.
A patent holder identifies a patent infringement in the ads, in a public presentation or in the web site of the software company and notifies the software company. A patent provides the right to exclude from using the patented process. Except if the software company is a direct competitor the patent holder will not exercise this right (which can be limited by the public interest) and will just ask for a license fee.
The software company has three options, (1) abandon the product (2) pay the license fee or (3) get sued with the risk to have to pay even more.
This scenario is credible because customers want what they have seen for instance in a conference or in a demonstration. They express needs and ask for solutions addressing said needs. From an intellectual property point of view a need is a motivation for developing a certain thing whereas the solution’s process or method can be protected by patent. If it does not make the difference between needs and solutions, the software company produces requirement documents containing patented matter and because the software company uses the same methods and tools and make the same conservative choices as the patent holder its product infringes one or more patents.
Most companies involved in software development are probably not ready to properly handle software and business method patents. This does not mean that patent handling is expensive. This means that said companies lack training and experience.
Software companies already:
properly protect their trademarks and domain names, which are also part of Intellectual Property;
know their competitors;
analyze customer requests;
identify the most appealing features of their products to support their marketing and sales forces.
Only someone who knows the company business and its competitors’ offers can identify and properly appreciate patents of interest. The hard work in processing customer requests is in collecting, understanding and evaluating them, not in finding whether a request is for a need, a public solution or a patented solution. Identifying patentable innovations made throughout a development should be easy: an innovation sells and is an inventor’s achievement. This is not to say that a company does not need external help, for instance to put the patent process in place or for tasks like patent drafting or prosecution work but this work complements what is already done. Furthermore a good patent processing improves the whole development process. For instance patent databases are reliable sources of business intelligence and companies that learnt to identify inventions can put in place policies aiming to increase the number of inventions.
Though it takes time and efforts to properly handle software and business method patents, this is not even clear whether this is an extra cost or a means like accountancy that software companies need today to assess and stimulate their inventiveness.
Drug makers explore several preparations and molecules in parallel knowing that each studied molecule has little chance to lead to an approved drug that sales. This R&D is distinct from the analysis of market needs and from the definition of the industrial process.
In the software industry things go differently. Inventions usually do not follow from a separate R&D effort but occur throughout product development. The time spent to invent is included in the project schedule in the same way as administrative work, coordination, analysis, code writing and technical tasks related to the use of computers.
Therefore, though the inventive work is probably of the same order of magnitude in the software industry as in other industries, the extra effort for filing software and business method patent applications is small (essentially identification and application writing). Therefore software and business method patents look cheap.
A patent is a right to exclude others from making, using, offering for sale, or selling an invention. This right can be exercised in three ways:
exclude others from making, using, offering for sale, or selling said invention;
sell the right of practicing the invention (license fee);
get a compensation for past practice of the invention through settlement or lawsuit (damages).
Universities usually have the policy to file a patent only when they found customers ready to buy a license (2) for practicing the invention. Large companies file patents to exclude their competitors from practicing the invention (1). Risky business entrepreneurs may file patents to get license fees but frequently get more revenue from compensation for past practice of their inventions (3).
A forerunner can anticipate the development of a new business and file a software or business method patent solving a problem of this new business. If other companies developing products for this new business do not identify this patent or believe that they do no infringe the patent or that the patent is invalid, said forerunner can sue these companies seeking damages based on the revenue generated by the use of the patented process or apparatus. When the alleged infringer is a company like Microsoft or eBay damages may represent a sizable amount of money.
This practice is broadly and loudly criticized. However there are a couple of aspects to consider.
Software industry is a zero marginal cost industry: A program costs the same to develop regardless of the number of users. License fees have a direct impact on the lowest price at which a program can be sold and on the development of new versions. Furthermore no free open source product can be developed for a patented process or apparatus. Such products play an important role in a broad adoption of a new technology and broad adoption is as important as intrinsic value because it conditions communication between organizations and creation of knowledgeable workforce. Decades of price falls taught customers to wait rather than pay the high price. Usually it is possible to design around an invention to provide a marginally inferior solution and market happily trades optimal convenience for inexpensive, broadly adopted solutions. Therefore software companies always design around patents that they believe to be valid. Software and business method patents addressing mass market needs are fences that deter users and providers from adopting the protected system or method, and hence make said patents worthless [6]. This fact seems to be acknowledged by companies like IBM that makes some of their patents free for open source developers or even for anybody.
Internet and the Web have been invented with public money (DOD, CERN, universities.) This public funding created a sane market place, in which providers could compete and users could choose. This market expects free, neutral (no actor should expect a much greater benefit from a new direction or standard than others), and simple directions and standards with a long term commitment of their promoter. We cannot expect simplicity from a consortium, or long term commitment of a company, or that a company will give away something which does not favour it, whereas public institutions are good at providing such directions and standards. This is not to say that governments favouring their own citizens or international institutions working like private consortiums can provide directions and standards. Directions and standards can only come from the competition between organizations such as universities using public money.
Software and business method patents addressing mass market needs are not the majority of software and business method patents. The public interest commands that a number of forerunners never stop experimenting new ideas and thus implies a proper incentive for forerunners who succeed. For instance why would corporations acquire small innovators if said innovators were not granted adequate protection for their inventions?
There were innovations in the software industry despite lack of protection when incumbent companies ignored what new entrants were making. Now companies that survived do not allow anymore innovators to consolidate market shares and make profits. Without an adequate patent protection the winning strategy would be to let someone pull the chestnuts from the fire with a first product and develop a second product providing the same function and addressing the shortcomings of the first product. If innovative new entrants were not rewarded there would not be innovative new entrants.
The goal of incumbent companies is to keep and increase their market shares, which implies matching third party innovations. The goal of incumbent companies is certainly not to innovate in a way that could kill or harm some of their milky cows. So innovation primarily depends on the number of new entrants.
So public interest generally commands to provide an adequate protection to software and business method inventions.
Between 1995 and 2000 patents were issued that did not deserve to be issued either because they were anticipated or because they would have been obvious to any person of the art. They were ignored by forerunners because they generally ignored intellectual property issues (software and business method patents were new) and even by incumbent companies that either:
were not ready to analyze so many patents, or
thought that said patents were not valid, or
thought that they were not infringing said patents.
Then juries found in a couple of cases that defendants were infringing these patents and awarded damages based on the revenue generated by the past use of their patents. Though widely regarded as unfair these decisions follow from two principles of patents litigations:
the presumption of validity of the patent examination, and hence
the burden of proof on the defendant.
These cases were just the tip of the iceberg: companies refrained from developing products that could infringe dubious patents to avoid costly settlements and the bad publicity associated with court cases and called for patent reform.
Mistakes were the consequence of a combination of factors. For instance the Congress decided that the size of the U.S. Patent and Trademark Office's budget would depend on the number of patent applications it receives. It is possible to size a help desk according to the number of calls because
other metrics, notably the average time to process an incident, are available;
help desks cannot incite customers to call more..
The patent office was in a different situation:
A general problem in patent business is that consequences of any writing and decision become apparent only a long time after. There is no easy way to measure examination quality.
To accept to disclose some of its secrets and make the effort of writing a patent application a company has to believe that it will get a patent in return. So granting patents to more applications increases the number of applications filings as far as company believe that granted patents are valuable.
The problem was made worse by:
the high turn over experienced by the patent office in late nineties;
the first reading of the State Street decision. According to this reading a patent could be issued for any method producing useful, concrete and tangible results.
Now the examination process is improving:
after Bowman a second reading of the State Street decision applies; business method inventions have to be within the technological arts;
examiners check unity of invention and enablement;
examiners have better databases.
This is not to say that the software and business method problem has been fully addressed. The very use of patents to protect innovation in the software field is still in question.
A legal system faces the nightclub dilemma, which is:
If you let everybody enter, nobody will come but if you let nobody enter you will not make money.
[The second alternative is more complicated. Potential customers will not come if they feel that they may not be allowed to enter.]
If the patent office grants patents to software and business method applications fulfilling a couple of presentation rules, courts will not presume that these patents are valid – de facto if not de jure. Because such patents have little value companies will not buy that sort of legal paper and hence make the effort of filing applications. On the other hand if the patent office rejects most patents, companies will find that they have little chances to get patents and will not make the effort of filing applications. Our analysis entitled “Evolution of the number of software and business method patents granted by the USPTO” suggests that this may be or become the case for business method applications but not for other software applications.
Markets react to changes of rejection rate with a delay. This delay equals or is higher than the time needed to get informed about changes in rejection rates and grounds. For patents this delay is probably three years or more. And it takes more time to get customers back than to drive them out.
Examiners are not bouncers and the patent office just enforces rules that follow from the laws and case laws. The nightclub model may help to predict the number of applications filed given a set of rules but cannot help to answer questions such as:
Is it the public interest to provide adequate protection for software and business method inventions?
Is it the public interest to incite inventors of software and business method inventions to protect their inventions?
What kind of framework and boundaries should be given to such patents to reward innovation while not impeding fair competition?
Today claims are accepted in other domains (drugs as in the Vicuron claims, semiconductors...) whose equivalent would be rejected for software and business method patents when software and business method applications are already among the hardest to draft for two reasons:
The drafter of a drug application can follow guidelines such as claim first the chemical composition, then the pharmaceutical composition and the treatment. The drafter of a software or business method application does not have such guidelines.
The inventor of a drug studied the properties of a molecule genus, found that said molecule genus could be prepared in a certain way, had a useful effect and could be administered in a certain way. So the layout of the application is somewhat similar to the inventor work. Software and business method applications require a construction notably because software and business method inventions do not follow from a planned R&D effort.
On average a software and business method application requires more time, is more demanding in term of skills (knowledge of patent law, of computer science, and of the business to which the invention pertains) and experience, and has less chances to be granted than another domain application. When the legal system cannot implement a system that would only allow the richest to apply, today only large organizations actually can hire experienced patent drafters and deal with high percentage of rejections.
Defensive patenting is a practice of incumbent companies. The goal of defensive patenting is not to acquire a rent but to secure market shares. The adjective “defensive” can be misleading: a patentee enforcing a defensive patenting policy can sue other companies. Defensive patenting just means that the patentee sues when it finds that the alleged infringer copied its patented system or method in a way harming its business. Therefore patentees enforcing a defensive patenting policy are more interested by the right to exclude than by license fees and damage awards. Their patents protect components and methods implemented in their products in order to prevent their competitors from providing the same functions in nearly the same way.
The primary motivation for implementing a defensive patenting policy is that, if competitors get patents and if the company does not get patents, then the company will be excluded from providing some products without having the means to exclude others from providing some other products. Defensive patenting is a business in which companies competing to serve the needs of a market watch each other and use patents against each other. A market determines a patent sub-domain in the software and business method patent domain to which it is possible to apply almost the same comments as for drug patents. An important difference relates to prior art.
As we have seen drug inventors can safely ignore healer prior art. People can use prior art pertaining to another market domain to invalidate patents. There are two reasons for that:
borders of market domains are blurred;
companies competing on a given market do not use specialized workforce. Most market domains use the same project management, methods, operation research techniques and programming languages.
Portfolio building policy is of the same nature as defensive patenting policy though more ambitious. A portfolio building policy consists in filing patent applications, not only for inventions practiced in commercial products as in defensive patenting, but also from inventions reduced to practice only in prototypes. Only software giants have the resources and the need for building patent portfolios. Patent portfolios serve three purposes:
Securing patent rights in areas in which the company does not plan to invest for the moment.
Securing a cross licensing capability. If a company B sues a portfolio owner A, said portfolio owner will check whether company B infringe any of its portfolio patents. If B infringes one of said portfolio patent A can exclude B for practicing an invention and B can exclude A from practicing an invention. Then A and B negotiate a settlement in which they agree to allow each other to practice their inventions.
Hiding the company’s plans and assets. Patents and patent applications are public. If the company was filing applications only for systems and methods practiced in their products analysts could deduce the plans and product designs of the company from its applications.
Both defensive patenting policy and portfolio building policy are designed to be used as weapons in the struggle with competitors. Companies enforcing a defensive patenting policy and portfolio owners are focused on their business. They are not seeking rents from much smaller companies because identifying such companies would be costly and because suing a much smaller company could be bad publicity.
Software and business method patents are as legitimate as other patents. They differ from other patents in the following aspects:
wider domain;
larger number of potential prior inventors and infringers;
software and business method inventions do not follow from a planned R&D effort. There are made throughout product development;
software and business method patents are obfuscated by the applicants’ concern to product what they have done and not their innovations.
Four types of inventions may coexist in this domain:
mass market inventions;
forerunner inventions;
defensive inventions;
portfolio inventions.
Defensive and portfolio inventions are protected and used like inventions of other industries. Incumbent companies file patent applications for these inventions to complement their arsenal in the long term struggle with competitors. Any other use of these patents would be interpreted as a lack of focus on their core business.
Forerunners file patent applications to protect their business from organizations that have more resources to exploit the invention than them and enhance the value of their companies.
Mass market develops in royalty-free directions. A patent protecting a mass market invention deters the public from using said invention and hence made said patent and invention useless. Because the patent system is thus unable to provide an adequate incentive for mass market inventions, said inventions have to be made by organizations such as universities using public funding.
The software industry did not widely adopt the principle of protecting innovation through software and business method patenting. Wider adoption depends on the capability of the legislative power, courts and patent offices to provide guidelines and clear rules and to consistently support said guidelines and rules a long enough time.
The role of patent law is to provide statutes to regulate business and provide incentives aiming to promote the progress of useful arts. Software and business method patents are not the correct tool to support mass market innovations but they are the cheapest and most effective way for supporting other innovations in the software field. So patent offices correctly grant patents to software and business method inventions.
The core issue is that the software industry does not properly handle innovation. The problem probably comes from two facts:
To the opposite of the industrial revolution the computer revolution was planned and funded. Governments and largest companies gave away huge amounts of intellectual property to develop this industry. People learnt through examples and copy and need now to not reproduce without proper judgment what others have done, to credit their sources, and to recognise innovation.
Software industry uses a skilled, ingenious and unspecialized manpower. This is a post-industrial division of work in which people and not organizations are given the task of finding solutions, and hence invent at unpredictable places and times and at an unmeasured cost.
The history teaches us that an adequate protection of innovation makes a big difference in term of growth and represents a big competitive advantage for those who implement first the correct protection. The seventeenth century patent law preceded the industrial revolution and helped this revolution to happen in UK. We do not believe that a patent law designed to address industrial needs can address post-industrial needs without changes but we believe that the only reasonable solution is to start with the patent law we know.
[1] The barrier to new entrants is high. Said barrier notably includes know-how and distribution channels. A new entrant usually needs more time to market its products than incumbent companies need to react to its offer. Entering such markets require large investments in specialized assets that cannot be easily sold or reused [high entry and exit costs]. Quite frequently conditions that enabled the creation of incumbent companies (low labor cost, no training needed because the know-how was yet to invent, no existing distribution channel to compete with) no longer exist on their home market.
[2] We could define the intellectual property as the means provided by the society to protect what is made public through a product or service:
Aspect | Means |
|---|---|
innovation | patent |
expression | copyright |
marketing | trademark |
[3] For instance at a given stage developers may find a huge sizing error or realize that a little forgotten detail breaks the planning. Such errors are powerful incentives to invent. Such inventions should not be overlooked on the ground that they were made to fix mistakes.
[4] The software industry is and will remain an industry without barrier to new entrants because:
it is inexpensive to create and sustain a software business,
programs like songs and unlike drugs can be made tangible as electronic files that can be inexpensively promoted and distributed through Internet.
[5] There are trials (Eolas v. Microsoft, MercExchange v. eBay...) and there are settlements. For instance on June 8, 2004 several sources reported that Yahoo was suing Google for patent infringement. Said patent, 6,269,361 “System and method for influencing a position on a search result list generated by a computer network search engine” was granted to a company, Overture Service (goto.com), bought by Yahoo. On August 8, 2004 Google and Yahoo settled out of court, Google giving shares to Yahoo, notably for a license to use 6,269,361 and other Yahoo patents. We can make three remarks about this settlement:
Stock markets do not like uncertainty. A public company and even more a company that plans to file an IPO have strong incentives to negotiate.
This is not easy to determine the cost of defendant shares as demonstrated by the current discussions about the cost of stock option programs.
Overture service’s patents were considered by Yahoo when they decided to acquire Overture service and actually paid back a part of the acquisition price. Buyers always check patents and patent applications of companies they may acquire and the question to determine is which weight is given to patents and patent application in the buying decision. And when a software company goes bankrupt there is almost always a company or firm to buy IP rights.
[6] Are the possible inventions solving a given problem, in a given environment, with given building block scarce, given that
an invention that differs from another invention in minor details is equivalent to said other invention;
an invention that would be found by any person of the art given the problem is obvious over prior art?
If possible inventions are scarce then software and business method patents addressing mass market needs are harmful.
Patents Presentation Search Issues Strategies Business Methods Patentability Analysis MercExchange eBay Trial Reexamination Business view Granted patents Examination USPTO EPO PanIP Eolas 1-click family 1-click analysis 1-click prior art Trademark Copyright
Contact:support@pagebox.net
©2001-2005 Alexis Grandemange
Last modified