You may join an online discussion of this article at http://technocrat.net/article.pl?sid=04/11/08/1640206
Patents, originally created to stimulate innovation, may now be having the opposite effect, at least in the software industry. Plagued by an exponential growth in software patents, many of which are not valid, software vendors and developers must navigate a potential minefield to avoid patent infringement and future lawsuits. Coupled with strategies to exploit this confusion over patents, especially in standards setting organizations, it appears that software advancement will become stifled unless legal action is taken to resolve the situation. This article examines the current situation facing software developers and users, the methods employed by standards setting organizations to address these problems, and recommends strategies for resolving the problem caused by software patents.
Today's computer industry standards increasingly include technology that may be covered by a software patent. The owner of that patent has the right to demand a royalty from all parties that implement the patented principle, or may discriminate regarding who will and will not be allowed to license the patent. It is often the case that there is no way to implement a standard without making use of a particular patented principle. This effectively gives the patent holder absolute control regarding who will implement a standard containing his patented principle.
Such patents arise in two ways: they are knowingly embedded in the standard as it is being created, or they are submarine patents, unknowingly part of the standard until they "surface" after the standard is already in wide use. A pernicious patent holder can engage in patent farming: influencing a standards organization to use a particular principle covered by a patent. In the worst and most deceptive form of patent farming, the patent holder encourages the standards organization to make use of a principle without revealing the existence of a patent covering that principle. Then, later on, the patent holder demands royalties from all implementers of the standard.
Several recent court cases have involved patent farming or submarine patents in standards.
The Eolas case is a good illustration of a submarine patent. A purported inventor held a patent on a widely used feature of HTML, the standard that describes the format of web pages. The patent holder did not make his patent known for years, and then sued Microsoft for use of his principle. The patent holder was awarded a half-billion-dollar judgment by the court. The Eolas patent is now being reviewed by the U.S. Patent Office, which has released two preliminary findings claiming that the patent is invalid.
The Rambus case is a good illustration of patent farming. A manufacturer allegedly influenced an organization attempting to standardize computer memory to use their technology. It is further alleged that the manufacturer did not disclose that they had a patent application in process on the same technology. Years later, the manufacturer started bringing patent infringement lawsuits against all implementers of the memory standard. Unfortunately, the patent-disclosure policy of the standards organization proved unenforceable, and the alleged patent-farmer may yet prevail in extorting royalties from its victims, literally the entire computer industry.
Patent royalties tend to create discrimination against small-to-medium-sized businesses developing any form of software, and especially against Open Source developers.
The largest businesses in an industry generally have patent cross-licenses with their peers, and thus they may ignore each other's patents while smaller businesses have no choice but to license those patents if they use them. As a result, there is a "tax" upon technology that small businesses must pay while the largest businesses are exempt.
Small-to-medium-sized businesses that make real products are generally unsuccessful in enforcing their patents against the largest businesses. When a smaller business brings a patent suit against a larger one, the larger one will often bring multiple patent suits—justified or not—against the smaller company. The smaller company is unable to defend itself. The American Intellectual Property Law Association claims that prosecution or defense of a software patent lawsuit costs about $3 Million dollars1 The smaller patent holder simply cannot sustain the expense of defending himself, even when justified, and is forced to settle and license his patents to the larger company.
The only small business that can win a patent litigation against a larger one is the “patent parasite”, which makes no real products and derives all of its income from patent litigation. Since they make no products, the parasites are themselves invulnerable to patent infringement lawsuits, and can pursue suits against very large companies without any fear that those companies will be able to make defensive use of their own patents.
Complex structures like the Internet depend on a tall stack of standards. About 30 standards underlie every click of a mouse in a web browser: from the HTML file format through the TCP/IP communications protocol, all the way down to a standard that specifies the thickness of the gold plating on an Ethernet cable. If just five of these standards each contain a single patent bearing a 5% royalty, that will cut out all of the profit potential for any small company engaging in Internet software development. The largest companies in the industry, however, will be able to ignore those patents and can make a profit.
The situation is worse for Open Source developers. Those developers generally do not charge a royalty for their software, and thus generally cannot afford to pass on royalties to patent holders. Open Source developers work for a number of different purposes: some have motivation similar to that of an artist; for some the goal is research; some produce their software as an enabling component for a non-software business. The act of creating Open Source software is essentially a nonprofit activity, although some companies may attempt to profit from distribution of that software or services connected to Open Source software. Companies like Red Hat are clearly for-profit, but generally are not the developers of the Open Source software and have no commercial connection to those developers.
Open Source lives on collaboration, and its collaboration forms only when all parties have the same rights. The presence of a required royalty payment for implementation of a patented principle would probably abort any Open Source development in that area. Since today's software patents are written to be deliberately vague in order to have the widest possible scope of enforcement, it is a truism that any significant work of software infringes upon a patent.
The patent situation is bad enough when applied to the general field of software. When applied to standards, it specifically restricts the inter-operation of programs and communication between them, because those are the subjects of computer industry standards. Thus, software patents can be used to prohibit interoperability. It is no surprise that Microsoft has embarked on an aggressive software patenting program. A 2002 HP internal memo, later leaked to the press, announced that HP's cross-license with Microsoft explicitly excludes Open Source software2. An agreement negotiated between Sun Microsystems and Microsoft and disclosed by Sun in a 2004 SEC filing specifically excludes the OpenOffice Open-Source office suite from protection.3
The quality of the patents in question is low: Microsoft's patent on the "FAT" file-system used for floppy disks and USB sticks was invalidated by the patent office in 2004, at the request of the nonprofit Public Patent Foundation (pubpat.org). A patent that Microsoft filed in 2002 appears to cover the invention of the computer mouse, surely invented before that date. There are so many patents of this low quality that all of the world's nonprofit organizations would be hard pressed to invalidate every one. And unfortunately, getting to invalidate a bad patent is no sure thing. One must petition the director of the U.S. Patent Office to re-examine the patent, and currently that director has the option to refuse.
Many companies, and essentially all Open Source developers, cannot afford to defend themselves from a patent lawsuit no matter how poorly it is justified. They must settle, no matter how pernicious the terms.
Before Gutenberg, copyists, using pen and ink, duplicated written political dialogue laboriously. Only the wealthy and the church could afford to employ copyists, and during this period the paucity of communications limited the exercise of democracy to small groups. The advent of Gutenberg's press made the mass distribution of written political dialogue possible. People vote based on what they hear and read, and the improvement in communications brought by the press made egalitarian mass democracy possible. It is thus no surprise that the first amendment to the U.S. Constitution protects the freedom of the press.
Within the last century, electronic communications have increasingly become the vehicle of democratic discourse. Because radio and television broadcasting are expensive with limited frequencies available, the wealthy have dominated broadcasting. The Internet and World Wide Web place into the common man's hands the capability of global electronic broadcasting. Clearly, the Internet is the most important tool of democracy since Gutenberg developed movable type.
In order to protect democratic discourse in the future, the Internet must remain a fair and level playing field for the distribution of political speech. The full capability of the Internet must remain available to all, without restriction by religious, business, or political interests.
A number of "Internet radio" devices have become available today. Most of those devices only receive stations that have been enabled through the gateway site of the device's manufacturer. This means that the manufacturer of an Internet radio can control what stations the device provides access to, and thus what political viewpoints are available via the device. One day in the future, most of us will receive text, audio, and video programming via the Internet, either wired or wireless. Imagine the problem for democracy if, when that day dawns, the manufacturers of our access devices are a few companies that have attained a market lock on Internet broadcasting, thus determining what political viewpoints the electorate can receive.
Standards are essential to the Internet because there can be no connections unless the parties agree on the way we will connect, the protocols we will use to communicate, the file formats we will use to exchange information. Unfortunately, because patent holders have absolute control over who implements their patented principle and how they do so, industry standards that embed patents can be used to restrict the Internet to the benefit of a particular company or group of companies. That company or companies will have significant control over the conduct and content of the Internet. Because of the discrimination factors explained above, the largest companies will be the ones likely to gain control over the Internet, and thus the trend is toward plutocratic domination of the future Internet.
The most likely way for a software patent to find its way into an industry standard is innocent inclusion by people who are not aware of the existence of the patent. However, the same factors that make innocent inclusion possible work to the favor of the deliberate patent farmer. Those factors are:
U.S. patent law allows treble damages to be extracted from a party that knowingly infringes a patent, while one who unknowingly infringes must pay only simple damages. The quality of existing patents is so low, that it is generally not possible to determine if a patent found in a search is indeed valid. Because it is impossible to develop a significant work of software today without infringing on a granted—but perhaps invalid—patent, the safest course is not to search for applicable patents, and thus avoid treble damages in any future court proceeding.
None of the alternatives to that process is appealing: go through extensive court or administrative proceedings to determine the validity of every patent found in a search, license every patent found in a search whether valid or not—probably removing all profit from the product, or give up the software project.
Because of all of the above, most software engineers and standards working group members are instructed by management and legal counsel not to look for existing patents covering their work. Thus, the patents may become evident only later, when their owners bring infringement proceedings.
Since software patents are deliberately vague in order to have the widest possible range of enforcement, no patent search can ever be said to be completed. It may simply not be evident to the searcher that a particular patent could apply to the principle in question. A patent holder may always assert a patent that has not come up in a patent search. Only an expensive court proceeding will determine whether that patent is applicable.
The principle being patented must be kept undisclosed until the patent is published, or the applicant will be denied the right to patent it. In some cases, one of the entities working on the standard has a patent application in process—either still being written during the 364 days permissible to work on an application after an invention occurs, or submitted to the U.S. Patent Office, but not yet published. If the patent applicant were to disclose that application, it would mean the loss of the right to be granted a patent, so they can't tell the standards organization what is going on.
This is an extension of the "Can't Tell" problem.
In the recent IETF MARID proceeding to standardize anti-spam technology, Microsoft disclosed that it had a patent in process on some of the technology included in the standard, but was unable to disclose the substance of the patent application due to fear of losing the right to patent. In addition, Microsoft declined to license the unknown patent under terms that were acceptable to all of the participants in the standards working group. This resulted in a deadlock in the working group, which disbanded without achieving ratification of its standard.
Some standards organizations are so afraid that their work will be aborted by the vague assertion of a patent, as happened with IETF MARID, that they decide to largely ignore patents in the interest of completing their work, and leave the issue of patent licensing for the implementers to solve. They thus create a technological standard which may not allow any possibility of implementation except by the largest companies, due to the embedded infringement issues.
In the U.S., it is illegal for anyone but an attorney to provide legal advice. Thus, standards organizations are loath to appear to provide legal advice to their implementors regarding any patents that may be infringed by an implementation of the standard. Some standards organizations resolve this problem by simply not saying anything about patents. Other more reasonable ones provide information that will be a clue to the attorneys of implementors, with a disclaimer. That disclaimer generally states something of the form: "This is not legal advice. Only an attorney who has contracted to advise you can provide legal advice."
This is a form of the "We're Not Your Lawyer" problem. Many standards organizations are afraid to provide advice, even with careful disclaimers that they aren't your lawyer. They fear that they could be liable if the advice is incorrect or insufficient. Implementors can be sued or suffer other damage to their business in connection with infringement of a patent exercised by a standard. They may then attempt to pass that liability on by suing the standards organization or even the working group members and their companies.
Like the "We're Not Your Lawyer" problem, some standards organizations deal with this by refusing to say anything about patents. The more reasonable ones add yet more disclaimers and give as much advice as possible.
Patent farming is deliberate insertion of a patent into a context in which that patent will later bring a royalty to its owner. The patent farmer offers its technology to a standards organization. The farmer may not disclose that a patent covers the technology. Without disclosure, the standards organization proceeds blindly, unknowing that the patent will be prosecuted against implementors of the standard later on. This sort of patent farming should be considered a form of fraud.
Some standards organizations embrace patent farming. Some just look the other way. In such organizations, when the patent holder discloses that they have a patent that applies to a principle used in a standard, the standards organization still accepts it into the standard. This is especially true in standards organizations that have a RAND (“reasonable and non-discriminatory”) patent licensing policy.
Unfortunately, a RAND policy is not necessarily either "reasonable" or "non-discriminatory". None of the standards organizations that claim to have a RAND policy define what is “reasonable”. The presence of a monetary royalty would generally operate to discriminate against Open Source developers, who collect no royalties from the distribution of their own software and thus have none to pass on to patent holders.
The counter to the "We're Not Your Lawyer" problem and the "Liability" problem is the legal load upon the implementors. If every implementor has to perform extensive legal research before implementing any standard, there will be no opportunity to make a profit. In addition, if every implementor has to do the same legal due diligence, then there is tremendous duplication of effort. It makes more sense to pool all of the resources of potential implementors to get legal advice that applies to all of them, and the standards organization is the logical entity to coordinate that activity. However, this brings up questions: should that advice be released to the general public, in the pursuit of keeping an open standard open, or should it be given only to the folks who pay for it?
Standards have always come with governance. In the earliest days, communities had a monarch who, among other duties, would set standards for his realm. In modern standards governance, we replace the king with a democratic consensus process. Developing consensus across an industry requires an element of fairness and equitability—or agreement will always be less than total. The best practice in this regard is referred to as an open standard. Open standards admit all parties that would use them, without discriminating against one party at the expense of another. Their governance is similarly supposed to be free of discrimination, so that one party cannot unfairly influence what the standard says at the expense of another party. However, few standards organizations have adequately addressed the issue of software patents and their effect on the open and fair character that the standards organization is supposed to maintain.
The rise of intellectual property licensing in software, and its inevitable extension to software standards, has led to a more rigorous definition of what governance and intellectual property terms are called for in something that is to be considered an open standard. The problem is that discriminatory intellectual property policy and unfair governance can actually encumber a standards document thought to be “open” simply because of its availability. This fact may not be immediately visible to the implementors or even the standards working group. This has recently has prompted several organizations to create formal definitions of open standards.
WTO, the World Trade Organization that promulgates trade treaties to which the U.S. is signatory, provided these principles for international standards organizations:
Transparency.
All essential information regarding work planning, work under progress and final results should be made easily accessible to all interested parties and throughout all stages of standards development. Appropriate time and opportunities for written comments should be provided.
Openness.
Any interested national member should be provided with meaningful opportunities to participate in standards development.
Impartiality and consensus.
The standards development process should be conducted in a non-discriminatory way and should not give privilege to, or favor the interests of, a particular supplier or country. Conflicting arguments must be reconciled to satisfy all national members.4
The Danish government statement on open standards includes the succinct statement: “An Open Standard is accessible and free of charge to all. It remains accessible and free of charge.”
The above national and international policies argue for careful consideration of the matter of patents within standards, because of the potential for discrimination inherent in the rights that are granted to the patent holder. However, this issue is insufficiently covered in existing treaties, national legislation and standards organization policy.
Standards organizations have enacted several different policies with regard to software patents in standards.
A disclosure-based policy requires disclosure of information regarding patents that might apply to the technology being specified by the standards working group. In general, the aim of a disclosure policy is to bring to the attention of the standards working group the possibility that a patent could encumber a particular element of the standard. Armed with that knowledge, the working group can then decide:
Disclosure policies assume that companies who are most likely to hold patents on technology used in the standard will be the ones most interested in a particular standard. A disclosure policy depends on the cooperation of those companies. If a patent holder is not a member of the standards working group or the standards organization, the disclosure policy won't apply to them. Thus, disclosure policies are not able to cause the discovery of all patents that might apply to a standard.
There are many different levels of disclosure possible. A standards organization could require:
Submitter disclosure policies require that direct submitters of technology to a standards working group disclose any patents that they are aware of that apply to the technology. This policy is intended to thwart deliberate patent farming, but depends upon all submitters behaving honestly. A patent farmer might make casual suggestion to the standards organization rather than formal submissions, and might thus influence the standard without becoming responsible to disclose his patent.
Working group member disclosure policies require that members of a standards working group disclose information about patents that they are aware might apply to the standard, even if they have not submitted the particular technology in question. This sort of policy goes beyond the need to deter deliberate patent farming, and seeks to inform the standards working group if the patents of its own members might encumber implementations of the standard.
This requires that all members of the standards organization disclose patents that they believe might apply to the standard. This sort of policy generally applies to a last-call before ratification of a standard, so that companies that are not on the working group of a standard are not required to keep close track of the entire development of that standard.
The three types of disclosure policies mentioned above only require participants to disclose information about the patents if they are aware of them. A standards organization can in addition require that its members (or the working group members, or the submitters of technology) search their own patent portfolios to discover additional patents that might apply to the standard.
The standards organization may in addition require that its members (or the working group members, or the submitters of technology) promise to license, with prearranged terms, any patents they hold that are not disclosed or found in a search before the standard is ratified, but are discovered later on.
A standards organization may require that its members (or the working group members, or the submitters of technology) license patents that apply to the standard with pre-arranged terms. These terms may include:
The standards organization may require that licensing terms are non-discriminatory. In general, this is a requirement that the terms be the same for everyone. However, any royalty requirement whatsoever may act in a discriminatory fashion against small-and-medium-sized enterprises and Open Source developers.
The standards organization may require that certain entities be offered different terms than others. For example, they may require that Open Source developers be offered royalty-free terms while a royalty may still be demanded of proprietary software developers. Some proprietary software developers consider this sort of carve-out to be a discriminatory policy.
The standards organization may require that the patent holder never require a royalty from any party that uses the patented principle to implement the standard. This can be used to protect small-and-medium-sized enterprises and Open Source developers.
The standards organization may permit patent royalties, but may require that they be "reasonable." Some standards organizations ask for reasonable royalties without ever defining limits to what is reasonable. Because a single application may rely on a stack of 30 different standards, even a very low cap on royalties per patent may prove onerous because of the potential for there to be many patents that apply to that large stack of standards.
The concatenation of a reasonable royalty requirement and non-discriminatory terms is sometimes referred to by the acronym "RAND," for "reasonable and non-discriminatory."
The standards organizations may require its members who hold patents that apply to a standard to join a "patent pool" operated by the standards organization. The pool collects licensing payments from implementors of the standard, and apportions them to the (potentially many) individual patent holders in the pool.
The standards organization may identify aspects of the standard that are essential to its implementation, and require that the patent holder license the technology necessary to implement essential components of the standard, rather than all essential or optional components of the standard.
The standards organization may require that the patent holder license the use of the patent only to implement the standard, rather than license all possible uses of the patent within a program that implements the standard.
The standards organization may require the patent holder to license the use of the specific claims within the patent that are necessary to implement the standard, rather than all claims of the patent.
The standards organization may define how a working group will react to questions about patents. Some standards organizations define a patent assessment process that is intended to answer, incompletely, whether the patent applies to a particular technology, whether the standard can be modified to avoid the patent, and whether appropriate licensing is available.
The standards organization may define in advance the process to be followed if a patent is revealed after the standard has been ratified. This process may include withdrawal of the standard if patent issues cannot be resolved.
The potential for enforcement of patent policy varies widely between different standards organizations and between different participants in a particular standard. The range runs from completely unenforceable to enforceable-by-lawsuit.
The most important factor is that standards organization policies are not legislation. Thus, they can only be held to apply to entities that have willingly contracted to follow them. If a patent holder isn't part of the standards organization, they won't have to follow its rules.
The broadest possible means of applying them would be to attach terms to the use of the standard itself as copyrighted property. That would require submission to a patent policy as a term of use of the standard. This would also require a "tear-open" license upon the standard. It might not admit to enforcement in every case due to legal issues regarding the nature of consent to a tear-open contract, and whether certain uses of the copyrighted standard constitute fair use under copyright law and would not require a license.
Another strategy is to require that members of the standards organization contract to follow a patent policy. Violation of the policy could be cause for a lawsuit from the standards organization as well as its individual members.
A popular but often ineffective strategy is the announcement of a patent policy without explicit consent by all parties. Some standards organizations read their patent policy aloud at the start of a meeting, in the hope that it can be held to apply to those who hear it and participate in the standards development.
An additional problem is that many standards organizations allow participation in standards discussions by parties who are not members of their organization and who have not contracted to comply with the policies of the standards organization.
The most carefully stated patent policy in a standards organization to date is that of the World Wide Web Consortium. Led by Tim Berners-Lee, who is widely hailed as the inventor of the Web, this organization counts among its members most of the leading companies concerned with the Internet, as well as a number of educational and nonprofit organizations. After several years of argument, W3C determined that all patents incorporated into W3C standards must have royalty-free licenses. Then they went on to address, in writing, the many nuances necessary to successfully operate a royalty-free policy without requiring that W3C members give up all of their patent rights. Because of the number of companies involved and the importance of W3C to all of them, W3C's policy has literally been reviewed word-by-word by a dozen attorneys. It is an excellent starting point for any other organization that wishes to establish a workable policy to handle the collision of software patents and industry standards.
W3C members sign a contract with the organization, and thus the potential for successful enforcement of W3C policies is high. W3C restricts submission of technology for a standard to its own members and invited experts who are not members but are also required to sign a contract with W3C as a term of participation. At a number of points in the standards development W3C publishes the standards document and invites public comment from parties that have no contractual relationship with W3C. There may be some potential for patent farming through the public comment process.
W3C starts with a disclosure policy. This policy is different for granted patents than it is for applications in process, because of the potential that early disclosure of an application in process could cause dismissal of the application. W3C requires that applications be disclosed before the last call for discussion before ratification of the standard. This generally would give the applicant enough time to submit an application to the U.S. Patent Office and see it published.
W3C's requires royalty-free licensing of essential claims of a patent necessary for implementation of the standard, and only for implementation of the standard. Any other use of the patent claims within the program is not covered by W3C policy and the patent holding company has full latitude to license these uses on its own terms.
One of the best components of the W3C policy is a procedure for a Patent Assessment Group (PAG). The discovery of a patent that may apply to technology in a proposed or ratified standard triggers a requirement that a PAG meeting be convened, and that the PAG report back to the working group and the organization management. The existence of the PAG process allowed W3C to handle the Eolas case while other standards organizations would have been at a loss for procedure to follow.
W3C's policy is to withdraw a standard if a submarine patent affecting the standard is revealed and the patent holder is not amenable to royalty-free licensing. This sends a clear message to patent holders that they cannot hold W3C standards hostage—the organization will work actively to invalidate the patent or write it out of the standard. W3C's clear policies caused their membership, the major companies in their industry, to rally behind W3C leadership in fighting the Eolas patent.
IETF, the Internet Engineering Task Force, is at the opposite end of the spectrum from W3C's licensing policies. This organization was responsible for standardization of most of the innovations behind the Internet before the development of the web browser. However, it was founded before software patenting was legal in the U.S., and is not structured to protect its standards from patent farming. IETF is an unincorporated association rather than a corporation, keeps no membership roll, and insists that all of its members are participating as individuals rather than employees of their companies. IETF proceedings generally take place upon open email lists, and anyone can submit technology to IETF standards development without any prior arrangement. Since IETF is unincorporated, there is no entity that would be a credible plaintiff to prosecute violations of IETF rules. Since there is no membership roll and no contract with members, there is no indication of consent to IETF's rules.
IETF has a patent disclosure policy that is read at the start of meetings and published in textual form. Some attorneys have opined that the maximum possible enforcement of this policy would be to deny the violator entrance to subsequent meetings.
In the recent IETF MARID anti-spam standard development, Microsoft revealed that it had a patent application in process that covered some of the technology in the standard. Microsoft refused to license the technology in a manner that was compatible with Open Source software licensing. The MARID working group was not able to reach a consensus on the use of this patent and disbanded without ratifying the standard draft.
There are a number of laudable sentiments behind IETF policy. Individual IETF members are, to a great extent, opposed to software patenting. They believe, however, that it is a problem to be handled by legislation rather than IETF, and thus they have so far been loath to create policies that would effectively manage the problem. IETF's culture includes a belief that everyone should be able to participate equally in standards development. For this reason, they eschew red tape such as contracts that must be signed before someone can contribute content to a standard, and they keep no membership roll. They insist that all participants in IETF development act as individuals rather than as employees of their companies. Unfortunately, this tends to cause IETF members and management to be blind to the fact that many IETF participants are indeed promoting their corporation's agenda. People who act explicitly as corporate representatives at other standards organizations are often seen at IETF meetings, fulfilling the same roles, but under the guise of "individuals representing no corporate agenda."
Obviously, IETF cannot continue to be a viable host for future standards development without protection from patent farming. They seem to expect that protection to come in the form of legislation.
There are significant problems in the U.S. implementation of software patenting that are really the root cause of the problems of patents in standards. An examination of them would merit another document, so I will stick to specific concerns regarding patents in standards here.
When any entity encourages another to make use of technology, without disclosing knowledge that there are patent applications or patents covering that technology, this is a form of fraud. The eventual goal is to demand patent royalties from persons who implement the technology, when these persons would not have implemented the technology at all had they known that it was encumbered by a patent.
There is a legal doctrine called Laches that covers a similar issue. Under the doctrine of Laches, a patent holder who becomes aware of an infringement cannot delay enforcement until the market is larger and would bring greater royalties. In general, courts hold that a six-year delay is necessary to activate Laches as a defense for those accused of infringement, although Laches cases have been won for shorter intervals. It is difficult, expensive, and no sure thing to win a case on Laches grounds.
Laches is a defense. Non-disclosure of a patent or patent application with intent to collect a royalty or license in a discriminatory fashion later on should be both a felony and a tort.
Opponents of my statement above might suggest that anyone can search a patent, and thus there should be no responsibility to disclose. However, modern software patents are written in so vague a fashion that searches don't catch them reliably. And, of course, one cannot search for patent applications.
The potential 364-day delay between the date of the purported invention and the date that the patent application is filed makes abuses possible by those who are willing to perjure themselves regarding the date of their invention. It also encourages purported inventors to hold back disclosure of their invention for as long as possible in order to increase the amount of time in which they can enforce their patent by up to one year. It would be advisable for the U.S. to award patents based on the date that the application is filed, rather than the date of the purported invention. This would motivate the rapid disclosure of inventions, and would thus close loopholes that currently encourage inventors not to disclose pending patents to standards organizations and other technology users.
Some national administrations allow the use of patented principles for purposes of interoperability without any need for the implementor to pay a royalty or enter into a license. This carve-out for interoperability would protect standards organizations, and would in general prevent manufacturers from using mandated incompatibility to drive competitors out of the market.
The problems presented by software patents are numerous and must be addressed on many levels. Standards setting organizations can partially resolve this problem by following the W3C’s model in which the intellectual property policy is clearly stated and members are required to adhere to that policy. New legislation providing protection from patent-farming and submarine patents is necessary. In addition, governments should recognize the importance of interoperability to any free market for computer software, and should legislate to allow the royalty-free use of patented principles for interoperability purposes. Developers can perform due diligence, but no strategy is available to a software developer that will be fully protective against the risk of patent prosecution. Thus, developers should investigate whether adequate insurance coverage is available, and should continue to agitate for protective legislation.
Worldwide government organizations can also impact this problem by scrutinizing the purpose and process of software patenting. Is software patenting really promoting innovation and contributing to the vigor of free markets? If governments are to award software patents at all, they must drastically increase the diligence carried out before those patents are awarded, and the opportunity to challenge those patents both pre and post-award.
The Open Source community faces a particular challenge in the area of software patents and standards. Since many Open Source developers are non-profit, they don’t have the ability to pass on royalty payments to consumers. Yet, the Open Source community is now the predominant provider of software for many applications. Standards setting organizations should help the Open Source community to at least partially avoid the patent minefield. To accomplish this, the two communities must work together to develop policies that meet both of their needs in a way that continues to fulfill the ultimate goal of standardization: interoperability.
Bruce Perens is a leader in the Free Software and Open Source community. He is the creator of the Open Source Definition, and founder or co-founder of the Open Source Initiative, the Linux Standard Base, Software in the Public Interest, and No-Code International.
Perens is a Senior Research Scientist for Open Source with George Washington University's Cyber Security Policy Research Institute. He operates his own strategy consulting firm, Perens LLC. He is a member of the executive boards of Open Source Risk Management, an insurance company covering risks of Open Source Software, and Software In The Public Interest, Inc., the non-profit organization behind Debian GNU/Linux and other Free Software products.
Perens is series editor of the Bruce Perens' Open Source Series line of books with Prentice Hall PTR publishers. The books sell well in retail stores, even though their texts are released under an Open Source license that permits anyone to copy the books and sell the copies. Previously, Perens spent two years with Hewlett-Packard as Senior Global Strategist for Linux and Open Source, and was president of Linux Capital Group, a venture fund that specialized in Linux.
Sheri Bolin edited this document, and it will appear in her The Standards Edge book series. The generous support of NTT Docomo Labs USA is making this and other articles possible. I developed a page or two of this material while consulting for Checkfree Corporation, and am grateful for their patronage. Thanks to The Cyber Security Policy Research Institute of George Washington University, and Tony Stanco there, for granting me an affiliation with their prestigious institution. Carl Cargill, director of Sun Microsystems Standards department, sponsors Bolin's series and invited me to write this article.
You may join an online discussion of this article at http://technocrat.net/article.pl?sid=04/11/08/1640206