The best defence?

by Jenna Wilson

Originally published March 2007 by Thomson Carswell on IPSource. Used with permission.

In the patent calendar, this may well be the year -- if not the decade -- of the troll.

In the United States, drives for patent reform have been given new life in view of recent lawsuits involving big names, big payouts...  and patents of dubious validity asserted by so-called patent "trolls". The Canadian intellectual property community has, for the most part, only watched from the sidelines, since most of the action is south of the border. However, this does not mean that Canadian businesses can assume they are safe from the same litigious behaviour.

There is no correct definition of a patent "troll". There is no particular set of characteristics that is exclusive to the so-called troll, because in fact many patentees share the same traits as trolls, but do not earn the same disapprobation in the industry or popular press. Patent trolls have been variously described as entities who own patents, but do not commercialize the patented inventions; who do not innovate themselves or make substantial commitments to research and development, but simply acquire patents from others for the purpose of enforcement; whose patents are of suspect validity (in the opinions of those alleged to infringe them, at least); or who derive their income principally from licensing their patents or collecting damages for patent infringement. However, even typical patentees may engage in this type of behaviour, and not deserve to be called a "troll". After all, patents do not bestow a right or duty to practise an invention; what they provide is the right to exclude others from making, using, or selling the claimed invention. The Patent Act gives a patentee the right to seek damages for patent infringement, or to seek (subject to the court's discretion) an injunction to prevent further infringement.

Still, the patentees who may justify the pejorative "troll" label include at least those patentees who use the threat of litigation to extract licence fees and royalty payments from their "victims", not necessarily because those the victims are actually patent infringers, but because it is cheaper for the victims to buy their way out. In the classic troll trap, the patents in issue are either of questionable validity or, if valid, may not actually cover the products and services of the victim that are alleged to infringe-but the cost to the victim of proving invalidity or non-infringement, and the remote risk of being subject to a permanent injunction, is greater than the "nuisance value" of a patent lawsuit.

Why is it that this particular patent enforcement strategy has become so popular now? One suggestion is that the seemingly sudden onslaught of patent trolls is due to explosive growth in software and business method patents, which are inherently more difficult for patent examiners to examine for novelty and inventiveness, resulting in the issuance of more patents of lower quality (i.e., questionable validity).

Trollish behaviour, to date, has been focused on the U.S. market, not Canada. However, Canada may, in fact, provide a useful proving ground for Canadian patents with U.S. counterparts. While patent enforcement in the Canadian market is clearly less remunerative, the cost of litigation (for those patentees unable to secure financing or contingency arrangements) is lower, and to date Canadian patents generally stand a better chance of withstanding validity challenges. Forcing a company that does business in both Canada and the U.S. to defend an action for infringement in Canada will give the patentee a good idea of the type of prior art that the defendant would advance in a U.S. proceeding; in the best case scenario, the defendant will settle or lose and the patentee will receive funds to add to its litigation war chest, and in the worst case scenario, only the Canadian patent will be put at risk, provided the patentee has not made any assertions that would give the defendant grounds to bring a U.S. action for a declaratory judgment of non-infringement or invalidity. In addition, if patentees discover that trolling is a successful business model, there's no reason to assume that they wouldn't try the same thing in other countries.

What, if anything, can be done about the troll problem? Tools and proposals for dealing with the problem of trolls include preventive, defensive, and remedial measures. However, many of these measures are either limited in their effectiveness, or harmful not only to trolls, but "legitimate" enforcers of patent rights. The only real solution is an improvement in the quality of patents issued.

Preventive measures: springing the troll's trap

The best practice for anyone in the business of supplying products or services has always been to manage intellectual property infringement risk with due diligence. Ideally, as part of the product development phase, patent infringement searches should be conducted to determine what patents and published patent applications already exist. The results of such a search may be used to shape R&D plans, and, with the accompanying legal opinions, to determine how to avoid infringement when a product or service is finally brought to market. Armed with that knowledge, in theory a company can avoid falling into a troll's "territory", since patent applications more than eighteen months old and patents are public documents.

In the real world, however, this type of due diligence generally remains an ideal, rather than an actual practice; many companies cannot, or think they cannot, afford to obtain prior art searches and non-infringement opinions prior to introducing a product or service into the marketplace. In addition, in practice a non-infringement patent search may not provide a definitive picture of the patent landscape.

First, most patent infringement searches will yield not only issued patents, but also published patent applications. In Canada, the claims of a pending patent application may be amended just about any time during the pendency of the application. Providing a legal opinion regarding potential infringement of any patent claims that may issue from a pending application involves a great deal of speculation, as the applicant may amend the claims to cover any subject matter described in the application as originally filed, provided the subject matter is sufficiently novel and inventive. In fact, it is possible that an applicant may amend patent claims to read more closely on a rival's or potential licensee's products launched after the filing date of the application, provided the application as originally filed will support such an amendment.

Secondly, patent applications are kept confidential by the patent office for the first eighteen months after the application's filing date or priority date, unless the applicant requests early publication of the patent application. Thus, a search of patent office records today can only provide a snapshot of the state of patent filings a year and a half ago. In a field with rapid innovation, there could very well be secret patent applications which will not be published until after a company's product or service is established in the marketplace.

At that point, upon discovery of the pending application, the company is left with the choice of continuing as originally planned and to accept the risk of future patent infringement, or work around the potential claims that will issue from the patent application now. In Canada, however, this problem is exacerbated by the lengthy pendency of patent applications - not strictly because of an examination backlog in the patent office, but also because an applicant may defer the request for examination for five years after the application filing date. While this provides the advantage of allowing the applicant to complete (in most cases) prosecution of a corresponding U.S. patent application, and to assess the value of proceeding with the Canadian application, it also lengthens the period of uncertainty for companies attempting to avoid future patent infringement. While one might assume that a patent applicant will seek allowance of the same claims in Canada as those granted in the U.S., there is no guarantee that the applicant will do so.

A company facing the prospect of possible infringement of a patent that may issue from a pending patent application also has the option of hastening examination of another party's published Canadian patent application, by making the request for examination if it has not yet been made by the applicant and/or requesting expedited examination of the application. Both of these options are available to the public and would have the effect of moving the application into, or further up, the examination queue at the patent office. Third parties can also file protests and prior art against a pending patent application, with the hope that the patent examiner will take the submission into consideration when assessing the patentability of the application. Furthermore, a company facing the prospect of possible infringement of an issued Canadian patent has the option of either requesting re-examination of the patent by the patent office, or bringing impeachment proceedings in the Federal Court.

But practically speaking, protests and re-examination are usually recommended by patent counsel with a certain amount of caution or reluctance. Once a protest or request for re-examination is filed, the conduct of the examination or re-examination, is out of the challenger's hands, and the patent applicant has the opportunity to amend the claims in order to avoid the cited prior art. Thus, the result of a protest or re-examination may be a stronger patent that is more likely to withstand an invalidity challenge, although possibly with narrower claims. An impeachment action, while it does not afford the patentee the opportunity to amend the claims, is likely to be prohibitively expensive. In addition, patents are not invalidated for obviousness at the same rate as they are in the U.S.

In short, due diligence only provides a limited amount of comfort in a field where any company may evolve into a troll, and where continual innovation quickly makes patent search results obsolete. Furthermore, even if a company can obtain a legal opinion that its products or services would not infringe another party's patents, it does not guarantee that this company will not fall victim to a troll attempting to extort nuisance value payments out of an entire competitive space.

Deterrent measures: the arms race

One strategy, which has been employed for years, is for a company that may be a potential target of patent infringement lawsuits to develop a patent portfolio with an eye to cross-licensing opportunities. These patents may arise from the company's own inventions, or may be acquired from third parties.

In the competitive context, possession of a sizable patent portfolio may serve as a deterrent to current competitors with patents of their own. Although a company may be tempted to try to shut down a rival's business by alleging patent infringement, the possibility that the defendant may turn around and assert its own patents against the plaintiff in retaliation would make reconsideration advisable. The patent portfolio also offers a resource for settlement purposes; a plaintiff may be willing to take an assignment of a subset of the defendant's patents in lieu of a cash payment, or may agree to a cross-licence instead.

Cross-licensing and retaliation are only useful tools when the plaintiff is vulnerable to the defendant's patents, or is in a position to exploit the licence. A plaintiff-patentee, such as a troll, who does not provide any product or service will be immune to a suit for infringement, unless the patentee can be held liable for inducing and procuring one of its licensees to infringe the defendant's patent.

But in some circumstances, a troll that is eager to build its own arsenal of patents may be willing to take on a portion of the defendant's patent portfolio instead of, or in addition to, money: in other words, the defendant escapes by feeding the troll. Thus, the strategy of building a patent portfolio may prove to be a deterrent to competitors, but not to a patent troll.

Remedial measures: equity and counter-attacks

Most attention has been focused on "remedial" measures for addressing the troll problem by limiting the legal remedies available to a patent troll. These remedial measures have focused principally on the availability of permanent injunctions, which provide the troll with the greatest leverage: a defendant may have sufficient revenues to pay a damage award, but a troll that obtains an injunction against an adjudged infringer is in a position to demand an exorbitant settlement. Proposed legislative amendments in the U.S. have included requiring proof of irreparable harm before granting a permanent injunction against an infringer, and restricting the availability of wilful damages for patent infringement so that mere knowledge of a patent would not trigger an assessment of treble damages.

The availability of a permanent injunction to a successful plaintiff-patentee was the subject of intense debate in the U.S. in 2006 in eBay v. MercExchange. eBay, the successful online auction service provider, had been found to infringe MercExchange's patent relating to electronic markets with a central authority to establish trust between individuals. The district court had denied MercExchange an injunction based on a number of considerations, including trollish symptoms such as MercExchange's willingness to license its technology, its failure to practise its own invention, and the business method nature of the invention. On appeal, the Court of Appeals for the Federal Circuit enunciated a "general rule" that the patentee is entitled to a permanent injunction in the absence of exceptional circumstances. The U.S. Supreme Court rejected both the CAFC's "general rule" and the district court's finding that MercExchange's business of licensing rather than practising the invention did not amount to irreparable harm, and remanded the case to the district court, and reiterated the well-established equitable principle that a permanent injunction requires demonstration of irreparable harm, the inadequacy of a damage award, that the balance of hardships favours granting the injunction, and that the public interest would not be disserved by a permanent injunction.

Prior to the U.S. Supreme Court's decision in eBay, RIM was faced with the task of arguing why it should not be enjoined from providing its BlackBerry messaging service after a verdict of infringement was handed down in favour of NTP in U.S. District Court. RIM relied heavily on evidence that the continued operation of the BlackBerry service was in the public interest, citing numerous examples of the reliance placed on RIM's communication solution by essential services, including U.S. Homeland Security, the Department of Defence, law enforcement, and hospitals. Because RIM and NTP settled before the court could issue a decision on the injunction, we will never know for certain if American public interest would have shifted the equitable balance in favour of RIM.

In the U.S., some defendants have struck back at a troll by requesting re-examination of the asserted patents in the USPTO after the commencement of the lawsuit. However, there is no guarantee in either Canada or the U.S. that this strategy will be of any use once a lawsuit has been commenced. In order to obtain a re-examination, the patent office must first determine that the party requesting re-examination has raised a substantial new question of patentability in the prior art documents submitted with the re-examination request. Re-examination is not automatic, nor is it immediate; a court may not stay a patent infringement action based on the defendant's hope that a re-examination proceeding will take place.

In Canada, punitive damages for patent infringement are extremely rare, to the extent that they are not a significant consideration for most defendants in assessing the liability risk. As in the U.S., in the vast majority of patent infringement cases in which infringement is proven, a permanent injunction is ordered against the losing defendant. However, there is precedent in Canada for withholding a permanent injunction on equitable principles in the right circumstances.

While potential troll victims may wish that courts would consider factors such as those considered by the U.S. District Court in eBay before ordering permanent injunctions, the decision to withhold or grant this equitable relief based on the value of the patent to the patentee is a dangerous precedent, because it would effectively give certain patents greater weight in court simply because of the way in which the patentee chose to exercise its legal rights: a patentee who chose to license its patents would not have the same remedies available to it as a patentee who practised an invention. The suggestion that certain types of patents, such as business method patents, might be entitled to less protection than others, is also disturbing; after all, there is no statutory distinction made between different types of inventions in respect of patent enforcement. Formulating any set of predetermined criteria calculated to catch the trolls may have the unintended effect of denying "legitimate", non-trolling patentees the remedies they are due.

The source of the problem: patent quality?

Some commentators have identified a drop in patent quality as the cause of the troll problem. With the surge in software and business method patent application filings in the U.S., not only has patent application latency (the time before the application is first examined) increased, but some say that the quality of the examination has decreased as well, as examiners are faced with increasing demands and not enough time to perform adequate prior art searches. As a result, more patents of questionable validity are issued, creating more potential revenue sources for trolls.

To address patent quality - and possibly trolls into the bargain - proponents of patent reform in the U.S. have proposed a number of substantial changes to the patent system, including the creation of a post-grant opposition process; changing the U.S. patent system from a first-to-invent to a first-to-file system; publication of all patent applications after eighteen months; and expanding the definition of prior art. Other suggestions have included a "use it or lose it" regime, giving the patentee a deadline for having the invention commercialized, failing which the patent would lapse, and tighter constraints on the filing of continuation (in Canadian terms, divisional) applications to prevent the situation where a patentee obtains a continuation patent with claims that read on subject matter that was actually developed by a competitor or potential licensor after the original patent application's filing date.

A different approach advanced by Professor Beth Noveck of New York Law School is the "peer-to-patent" system. Based on the premise that the problem of patent quality is rooted in inefficient access to inadequate information on the part of patent examiner, the peer-to-patent system would have patent applications submitted for review by members of the public, who may cite prior art for the examiner's consideration. The USPTO agreed to implement a pilot project with a target start date in spring 2007. This community-based approach is reminiscent of various open source communities' efforts to collect prior art to invalidate a patent - which calls into question the objectivity of the individuals who participate in such a peer review.

In Canada, complaints about patent quality are not as frequent. This may be due only in part to the significant delays in obtaining a Canadian patent. While the Canadian Patent Office is working towards an eighteen month latency from the date the request for examination is made, currently an applicant can expect to wait upwards of three years before receiving a first office action for a computer-implemented invention. If an applicant defers examination for as long as possible, it may be eight years or more before the Canadian application is examined. In the meantime, the USPTO may have issued a corresponding patent that should not have been allowed; and if the patentee decides to turn to trolling in Canada, the fact that the U.S. granted a patent and the low Canadian threshold for inventiveness may frighten a Canadian victim into a settlement, since, as noted earlier, the ability to assess a future risk of infringement is hampered by the long pendency of patent applications.

Thus, ensuring that only high-quality patents are issued is a significant step towards correcting the troll problem; however, it is only part of the solution. Potential infringers or troll victims still need to be able to perform an adequate assessment of the risk of infringement when it matters - not only when the patentee alleges infringement, but when the potential infringer's due diligence reveals the existence of a pending patent application.

One change to the Canadian patent system that may help a potential infringer avoid infringement with certainty is a shorter pendency for pending applications by removing the option to defer the request for examination, and by requiring examiners to issue final actions earlier in the process. Even if a company does its due diligence and attempts to get non-infringement opinions on pending applications, the reliability of those opinions is compromised because those patents are not yet issued. Another possible change is to limit divisional filings in Canada to those applications filed in response to examiners' requisitions to restrict the number of pending claims in the parent application, rather than allow applicants to file divisional applications on previously unclaimed subject matter. Because some trolls are known to keep pending divisional or continuation applications while they assert the parent patent against an infringer just in case the patent is invalidated, an even more radical change to divisional practice would be to amend the Patent Act such that once a patent is asserted in an infringement action, any parent or divisional application in the same family that is still pending automatically lapses.

This is not to say that additional steps could be taken to assist examiners in during prosecution of the patent application to help improve the quality of examination and of the patents ultimately issued. For example, the Patent Rules currently require Canadian applicants to identify relevant prior art if it was cited against a corresponding patent application in another country - but only if requisitioned to do so by the examiner. This particular rule of patent practice has lost its importance in recent years, as the USPTO and EPO have made their file histories available online, making it easier for the examiners to locate that prior art themselves. An alternative rule requiring applicants to advise the examiner of all relevant prior art of which the applicant is aware, in the manner of the information disclosure obligation in place in the U.S., would be of more use to examiners in identifying relevant art or determining the direction of a prior art search. A more extreme version of the disclosure obligation is to impose a mandatory prior art search obligation on the patent applicant; for example, requiring the applicant to commission a search of one or more patent databases, or designated databases of non-patent literature, to assist in the location of prior art.

From even the brief suggestions above, it can be seen that almost any suggestion to improve patent quality or increase certainty for the parties affected by patents will be found to be objectionable by one stakeholder or another - in short, there will not be any "fix" to patent quality or the troll problem that will please everyone.

Determining which solutions should be put in place requires, first, that we decide who should pay for problems that now need fixing: do we place the burden on patent applicants by imposing more stringent requirements on patent applications and possibly driving up patent filing fees, forcing applicants to perform prior art searches, or restricting their opportunity to file divisionals, or do we punish the businesses who supply goods and services by subjecting them to weak patents and uncertainty? The answer, we suggest, is to place the burden on the patent applicants. Forcing troll victims to bear the cost of weeding out bad patents through impeachment litigation will not work, because the high cost of litigation creates a favourable climate for trolls - the U.S. is in that situation already. While raising the barriers to entry to the patent market may seem unfair to small businesses and independent inventors, it should be remembered that patents are assets, and assets cost money; furthermore, in Canada, the majority of businesses are small businesses.