The Software Protection Debate

Tri Ngo
Richard Sinn
December 19, 2005
6.901 Final Paper
Professor Robert Rines
Abstract
Software patents have been a controversial topic for a very long time. The unique
characteristics that govern software distinguish it from any other product that has been
historically patented. Software has existed and flourished before they were largely
patented, and it should be continued to do so without patents. Patents are not needed to
encourage innovation when it comes to software; innovation is an inherent characteristic
of software. Patents can actually discourage the independent inventor from innovating
because of fear of expensive litigation resulting from unknowingly infringing on a patent.

Introduction

Although patents were originally designed to protect the inventor and promote

innovation, we believe that patents with respect to software do not. In fact, software

patents sometimes impede innovation, and can actually be harmful to the independent

inventor.

The United States Constitution declares that the purpose of patents and copyrights

is to “promote the progress of science and the useful arts by securing for limited times to

authors and inventors the exclusive right to their respective writings and discoveries.” In

other words, a patent protects an inventor from imitators, and gives the inventor incentive

and financial compensation for the cost of his or her pains. If patent protection did not

exist, a potential inventor, especially an independent inventor, might decide against

inventing because of the expenses and risks associated with inventing and putting the

invention into form.

However, the idea of software patents has been historically controversial. Many

prominent figures in the “software world” oppose the ideas of software patents. There

are also many international movements against the idea of software patents, and a few

foreign governments have taken steps to restrict software patents. Even in the United

States, software patents were difficult to obtain until only recently.

1
 People often confuse software patents with software copyrights, but there is much

difference between the two. Software programs have always been copyrightable, but

patenting software has not always been straightforward. Because software programs are

written, they automatically fall under copyright protection laws, but there has been much

debate as to whether or not the process and algorithm behind software programs can be

patented. Copyrighting software gives protection to the expression of the written

program, but not to the actual implementation or idea behind the code. Thus getting

patent protection on a software program would give the inventor stronger protective

rights.

Controversy

Patents are generally considered beneficial in most industries, but patents in the

software industry are surrounded by controversy. Software tycoon Bill Gates in a memo

wrote,

"If people had understood how patents would be granted when most of today's

ideas were invented and had taken out patents, the industry would be at a

complete standstill today. ... The solution is patenting as much as we can. A future

startup with no patents of its own will be forced to pay whatever price the giants

choose to impose. That price might be high. Established companies have an

interest in excluding future competitors." (Lessig 2002-07-24: Keynote to

OSCON)

2
On the other ideological extreme is open source software advocate Linus Torvalds who

humorously writes,

"The fact is, technical people are better off not looking at patents. If you don't

know what they cover and where they are, you won't be knowingly infringing on

them. If somebody sues you, you change the algorithm or you just hire a hit-man

to whack the stupid git." (http://lwn.net/Articles/7001/)

These statements from representatives of the software industry demonstrate the dubious

nature of patents in this industry.

In fact, some countries have attempted to move beyond the controversy by

attempting to eliminate software patents. Article 52 of the European Patent Convention

(EPC) excludes "schemes, rules and methods for performing mental acts, playing games

or doing business, and programs for computers" from patentability. However, diverse

interpretations of this and related statements in the EPC have allowed many software

patents to be granted despite this law. In response, certain lobbying groups throughout

Europe and the rest of the world have launched a concerted effort to ensure the

elimination of software patents.

The appearance of recent movements from within the software industry has also

cast a shadow of doubt on the appropriateness of software patents. The open source

software movement and the industry formed around it have felt threatened by the

existence of software patents. Open source developers and companies argue that software

3
patents impede their ability to create new software due to fear of litigation from the

unknowing infringement of an existing software patent.

Legally, software patents exist on slippery foundations. Unlike physical

inventions, abstract software code is difficult to categorize. Some see software as a

mathematical algorithm and software innovation as the discovery of a new algorithm.

Mathematical algorithms can been seen as a product of nature and as such are ineligible

for patent protection.

In response to this controversy, economists have attempted to provide real world

evidence of the effects of patents in the software industry. Ironically, these studies have

increased the controversy rather than quelling it. Due to the expansive nature of economic

studies, researchers have found it difficult to fully eliminate confounding variables.

Unlike the physical sciences where hypotheses can often be tested directly through a

carefully designed experiment, the creation of an economic experiment to test the effects

of patents in the vast software industry is difficult. Researcher Robert W. Hahn concludes

in his article, "An Overview of the Economics of Intellectual Property Protection," that

"...despite its long history, the literature on intellectual property rights has found few hard

conclusions." (37)

Landmark Cases

The United States Patent and Trademark Office (USPTO) website says that under

the patent law statute, only “process, machine, manufacture, or composition of matter”

can be patented, while mathematical algorithms or scientific truths cannot be patented.

4
Originally, the USPTO considered software to be a mathematical algorithm, and thus not

patentable. However, through the years, the ability to patent software has grown

significantly easier.

In 1981, the United States Supreme Court resided over the landmark case

Diamond v. Diehr (www.law.uconn.edu). The Supreme Court decided that a machine

controlled by a computer program was patentable. Although it did not fully allow a

computer program to be patentable by itself, “it set the stage for later precedent-setting

decisions that extended protection to software.” (Hahn 2). The ruling meant that even if

most of the invention under review was composed of a computer program or a

mathematical formula, it could be patentable as long as the invention as a whole met the

requirements of a patent. Through this ruling, “creative patent attorneys were now able

to wrap software innovations into patents for tangible processes or products.” (Hahn 2).

The 1998 Supreme Court case, State Street Bank & Trust Company v. Signature

Financial Group, Inc. further eased software to be patentable. In 1999, only one year

after the “State Street Decision,” John W. Rees wrote that the decision immediately

triggered a “boom in [software] patent application filings.” Although the Supreme Court

decision was originally directed toward business method-related software, it eased the

restrictions on the patentability of software in general. The State Street decision basically

stated that anything could be patented as long as it could “produce a useful, concrete and

tangible result” (Arnold 2001). This “boom” in software patents still exists, and experts

estimate that there are now thousands of software patents, with another 10,000 patent

pending (Syrowik 2003).

5
 Ironically, as patent protection for software increased throughout the years,

copyright protection has weakened. Early on, court rulings tended to favor software

copyright holders, but copyright protection was substantially weakened in the case of

Lotus Development Corp. v. Borland International, Inc. (Hahn 3). In 1996, the Supreme

Court upheld without comment the decision made by the First Circuit. This was the case

that essentially established the idea that copyright protection for a software program

could only be extended to the expression of the program, and that anybody could

duplicate the code “provided one does not copy the literal code” (Lundberg 55).

Characteristics of Software Innovation and the Software Industry

The controversy that surrounds software patents may have its source in the unique

characteristics of the software industry and software itself. Software is fundamentally

different from any other invention or innovation the patent system ever had to deal with.

Software is not limited to any physical constraints, and as a result, it has created “a

different kind of industry with its own particular economic structure” (Irlam).

Gordon Irlam and Dr. Ross Williams argue that because software is free from

physical constraints, it’s much more complex than any other industry. Software’s

complexity has grown so much that some computer programs cannot be understood by a

single person. A typical industry can have a product that is comprised of twenty parts,

and a more sophisticated industry such as consumer electronics can have products with a

thousand parts. However, a software program can comprise of millions of lines of codes,

and an uncountable number of parts.

6
 Another aspect of software that makes it more complex is the abstraction

techniques of computer programming. Software programs are abstracted into

components to be used in larger software programs, and in turn abstracted again to be

used in even larger programs. Therefore, “software’s abstraction makes it difficult to

partition these technologies” (Irlam). Irlam and Dr. Williams continue to argue that most

industries have products that are covered by only a few patents, but in the software

industry a “product can contain thousands of inventions, any of which might be

patented.” Because of the ability for one software program to contain many potential

patents, and because of the difficulty to partition the many innovations within one

software program, it is difficult to analyze each software program.

To add to the complexity of software, the software industry is very dynamic. It

develops much faster than any other industry, and new products in software come out

every few months, while other more typical industries “typically produce a new

generation of products every ten to twenty years” (Irlam). Therefore, during the twenty

year term of a patent, many generations of software programs can come and go. A very

well known example of this fast evolution in software can be seen with the Windows

operating system by Microsoft. Microsoft released Windows 3.1 in 1992, and only nine

years later Microsoft released Windows XP, the most used operating system in the world

today. In between those nine years, Microsoft went through many significant changes

and innovations to their product. To compare the speed of innovation in the software

industry with some other conventional industry let’s take a look at the media storage

industry with respect to the VHS and DVD. The VHS format was released in 1976, and a

7
new widely accepted standard in video and media storage did not appear until the DVD

came out in 1996, almost 30 years later.

The software industry is also very different economically from many of the other

industries out there. The aircraft industry has medium research costs, high development

costs, and high production costs. Conversely, Irlam and Dr. Williams state that the

software industry has low research costs, high development costs, and low production

costs. Irlam and Dr. Williams argue that software has such low research costs because

“development has not been able to keep up with research” and that software has a high

development cost because “it takes a lot of human effort to write production-quality

software.” Production costs for software is low because it does not cost much money to

copy code.

Originally, patents were intended to encourage inventors to invent by incurring

the costs of innovation by promising exclusive rights on the new invention for a limited

amount of time. However, in an industry where the cost to research and produce may be

minimal anyway, it is unclear if software patents serve this purpose. Software patents

may actually be devastating to small companies because litigation over patent

infringements can be very expensive to both parties, and it could destroy small

companies.

Because of these unique characteristics in software innovation, the software

industry innovates in a sequential and complementary way. Bessen writes "The

sequential and complementary nature of innovation is widely recognized, especially in

high-tech industries" (3). The innovation in the software industry is sequential because

every new invention builds directly upon a previous one. Additionally, since innovations

8
in the industry often occur through the complementary efforts of many inventors,

software innovation is labeled as being complementary. In light of these characteristics,

patents can actually impede innovation. Patents can interrupt the sequential innovation

mechanism by preventing the use of a prior invention as a basis for further innovation.

For example, patents can restrict competitors from contributing ideas that may actually

help achieve new innovations. In an industry that exhibits sequential and complementary

innovation, a firm that patents its product "can prevent its competitors from using that

product to develop further innovation" (Bessen 3).

The Open Source Movement

The recent open source software movement is opposed to software patents, and

refuses to take out patents on any of their new innovations, and yet it has been

surprisingly successful.

The open source movement can be interpreted as a response to the tension in the

software industry regarding intellectual property rights. Unsatisfied with the proprietary

development model, open source software wishes to make explicit the implicit

characteristics of the software industry. The movement aims to guarantee the right of the

software innovator to use prior technology as a basis for innovation. The movement also

encourages extensive collaboration on the development of software. These two aims run

parallel with the unique industry characteristics of sequential and complementary

innovation.

9
 Many reliable and praiseworthy software programs have been created as result of

the open source movement. Many academics and software scholars are avid supporters

of the open source movement. Tim Berners-Lee, Richard Stallman, and Linus Torvalds

are a few names that are commonly associated with open source software. They and

many others believe that because software was borne out of academic spirit, information

should be free and available to everybody.

Commonly Suggested Solutions

Although there is much debate as to whether patents benefit or harm the software

industry, it is the consensus that tensions and problems exists concerning software

patents. In order to alleviate these problems scholars have suggested some improvements

to the current treatment of software patents. These suggestions all attempt to place patent

policy in agreement with the sequential and complementary innovation present in the

software industry. To improve the current system, some have suggested that the

obviousness barrier to obtaining a software patent be heightened, making it more difficult

for the USPTO to grant a software patent. Other suggestions include that the scope of the

granted patents should be narrowed, which would make the number of patents actually

granted much lower. The biggest problem with all these suggestions is that it is difficult

to quantify these regulations, which is why it was a problem in the first place. It is very

difficult, if not impossible to sometimes quantify what “obviousness” means, or what

“novelty” means.

Another idea that is commonly suggested is that software patents should be

banned all together. Although this may sound like an extreme proposition, the basis for

10
this argument is well thought out and convincing. There are a lot people who support this

idea, and many of them are accredited computer scientists and software programmers.

Our Proposed Solution and Reasoning

In order for the software industry to develop efficiently and software to innovate

to its potential, it would be best to ban software patents. Software patents not only hinder

the progress of software innovation, but it has grown to the point where software patents

are almost useless.

Patents do not make sense for the software industry in many ways. First of all,

the software world is so dynamic that taking a twenty year monopoly out on a particular

software invention does not make much sense. During those twenty years, several

different generations of a product can be innovated and produced, and a software

invention that was non-obvious twenty years ago may seem very obvious only two years

later.

Moreover, each software program can potentially contain hundreds if not

thousands of different innovations and inventions that can qualify as a patent under

current law. This makes developing new programs very difficult because programmers

and developers have to make sure that their new software programs, which can also

contain hundreds or thousands of new innovations, do not violate any existing patent. All

this could be very taxing on small companies and independent inventors because they

might not have the resources to search through thousands of patents to make sure that

their innovations in their software program do not violate any existing patents. This

11
would not only slow down innovation, but also discourage small companies or

independent innovators from innovating for fear of litigation.

In regards to software, the industry will be able to thrive even if patents did not

exist. The cost to produce software is so low, that a form of compensation from patents

is not needed. Also, because of the unique characteristics of the software industry,

innovation is a necessary quality a company or an inventor must have in order to be

successful in the software world. Thus, the software world does not need patents to

encourage people to innovate; it is already inherent in the industry.

Software patents have grown to the point where major software companies are in

a patents “arms race.” Big software companies try to take out as many software patents

as possible to protect themselves from other software companies that also have many

software patents. The reason for this behavior is because if one big company decides to

sue another big company over patent infringements, the other company will be able to

counter-sue with their own claims of patent infringement. The idea is that because there

are so many patents out there, the chances of someone infringing on a software patent is

very high, and that each company must be infringing on a patent somewhere. Therefore,

the more patents a company owns, the more likely it can defend itself if someone decides

to sue them. Additionally, if a big software company wants to eliminate a small software

company as a competitor, it can sue them for infringing on one of their patents. Fighting

litigation in court is very costly, and it can destroy a small software company or an

independent inventor whether they win or lose.

Patents in software do not serve the purpose of what a patent should be. It creates

economic inefficiency because the industry behaves differently from other more

12
traditional industries. Patents tend to be more beneficial to the bigger companies, and

they are not necessary to stimulate innovation in the software industry. In the early

histories of software, software was developed under a patent-less environment that

followed from the academic spirit of information sharing. Software was able to thrive in

the beginning without any patents, and it should continue to thrive today without patents.

Conclusion

The idea of patents in the United States existed since the drafting of the

Constitution, and the United States patent system has existed for two hundred years.

Then why is it starting to break down now for software? Software is so fundamentally

different from anything we have seen before, that the patent system just does not work for

it. The idea that patents are supposed to protect the inventor and promote innovation

does not apply to software; software can continue to prosper without patents. At best, the

patent system is ineffective for software, and at worst it can be injurious to the entire

software industry.

13
Works Cited

Arnold, Beth and David Lane. “Patent Strategies for Protecting Bioinformatic Inventions:
It May be Worth Venturing Out of Group 1600.” Foley Hoag, LLP. 1 February
2005. 10 December 2005
<http://www.fhe.com/publications.asp?pubID=000323292105#_ftnref9>

Bessen, James and Eric Maskin. “Sequential Innovation, Patents, and Imitation.”
researchinnovation.org January 2000. 10 December 2005
<http://www.researchoninnovation.org/patent.pdf>

Hahn, Robert W. Intellectual Property Rights in Frontier Industries. Washington, D.C.:

AEI-Brookings Joint Center for Regulatory Studies, 2005.

Irlam, Gordon and Ross Williams. “Software Patents: an Industry at Risk.” 25 January
1994. 10 December 2005 <http://lpf.ai.mit.edu/Patents/industry-at-risk.html>

Lundberg, Steven W. and Stephen C. Durant, ed. Electronic and Software Patents.
Washington, D.C.: The Bureau of National Affairs, Inc., 2000.

Rees, John W. “‘State Street’ Decision Causes ‘Boom’ in Software Patent Filings.”
FindLaw.com. 1 March 1998. 10 December 2005
<http://library.findlaw.com/1999/Mar/1/128488.html>

Syrowik, David R. and Roland J. Cole. “The Challenge of Software-Related Patents: A

Primer on Software-Related Patents and the Software Patent Institute.” SPI.org. 9
December 2003. 10 December 2005 <http://www.spi.org/primintr.htm>

The USPTO – General Information Concerning Patents. 2005. USPTO. 10 December

2005 <http://www.uspto.gov/main/patents.htm>

Diamond, Commissioner of Patents and Trademarks v. Diehr. University of Connecticut.

10 December 2005
<http://www.law.uconn.edu/homes/swilf/ip/cases/diamond_di.htm>

14