Software is a Key Component of the U.S. Economy

Software is a Key Component of the U.S. Economy

1. Software is Playing an Increasing Role in Modern Innovations

1. Software Innovations Contribute to a Significant Portion of the U.S. Economy

Private entities and consumers spent $257 billion on software in 2010, accounting for almost 25% of the total investment in private equipment and software, and up from $57 billion spent in 1992.⁶ In 2006, the software and information technology industries employed more than 2.7 million Americans, adding 400,000 jobs between 1997 and 2006. While the market growth of other industries was declining, the software industry’s market growth has increased since 2006 and is projected to continue to grow through at least 2016, with the software labor market among the fastest growing through at least 2016.^7,8 The current business climate, including the U.S. patent system, has clearly created a fertile environment for the software industry to expand its significant contribution to the U.S. GDP, a trend that is predicted to continue. The Majority of the Software Industry is Comprised of Small Businesses

Just over 50% of new jobs in the software industry were created by small businesses during the 2010-2011 year.⁹ Despite main stream media’s attention to the battle among giant companies – e.g., Apple, Google, and Samsung – small businesses make up most of the software industry. Census Bureau statistics report that of the 7,000 firms that were in the software industry as of 2011, nearly 5,000 were small businesses with less than 500 employees.¹⁰ In 2014, 42% of all venture capital investments were in software firms.¹¹ This outpaces the biotechnology industry by 400%.¹²Software Patents are Key to Continued Growth in the Software Industry and the U.S. Economy

Failure to provide patent protection for software would weaken competition in the software industry (for example, by inhibiting startups) and could cede control of one of the most valuable industries in the U.S. to other countries. The few large software companies that can afford to withstand a lack of patent protection by virtue of their market share have a distinct advantage over smaller competitors that rely more heavily on patent protection. In addition, more software companies could turn to trade secrets to protect their innovations, counter to the desired policy of disclosure promoted by patents. Further, software companies could shift out of the U.S. to countries that offer stronger protection for software.

2. Software Patents are Essential for Startups

Software-related patents remain a critical tool for startups, as they rely on the patent system to obtain investments and recoup research expenses. Patents are vital in attracting investors.¹⁵ Businesses also rely on patents to gain a competitive advantage and prevent piracy and copy-cats.¹⁶

In an industry such as software where copying is both easy and rampant, a strong patent system is essential.¹⁷ This is especially true for new and small businesses that may depend on patent protection for their success.¹⁸ According to some,¹⁹ weak patent protection for software will weaken the U.S. economy. Those that copy the work of others do not, by definition, create new things. “When innovators cannot get protection they cannot get funding, they do not form companies, jobs are not created and pretty soon there won’t be anything left for the copy-cats to copy.”²⁰

Early patent protection is extremely important in the development process of a new invention. Unlike large companies, who have the economic resources to finance the process itself, startups rely heavily on venture capital. For startups, the need for patent protection is absolutely essential in the process of obtaining venture capital investment. The lack of patent protection disincentivizes venture capital companies from investing in startups.²¹ Without this investment, the startup companies will be more likely to fail, and the public may never realize the benefits of the innovations from these failed startups.

3. The Public is Harmed by Undue Reliance on Secrecy

4. The U.S. is Harmed by Weakened Protection for Its Companies

The patentability of software inventions is one reason why software is dominating venture capital investment. In 2014, just over $4 billion was invested in software deals by venture capitalists during the first quarter, four times as much as biotechnology.²³ In the first quarter of 2014, 42% of all venture capital investments were in software firms.²⁴

The link between weak patent protection and loss of investment can be seen clearly by examining patent regimes that traditionally shun strong IP rights. For example, a recent Forbes article suggests that weak Indian pharmaceutical patent protection has had the effect of weaker investment. The article states that “it is unsurprising, if regrettable, those innovative drug-makers are reluctant to sell their medicines there. Investors are also unlikely to put their capital at risk in Indian drug companies that seek to discover new medicines.”²⁵

Similarly, if software companies cannot receive appropriate protection in the U.S., foreign countries could strengthen their protection of software, attract businesses making significant research and development investment, and rise to become new leaders in software innovation. The software industry is one of the most important and profitable industries in the U.S. and one of the leading job creators.²⁶ Patent policy has been vitally important to promoting the innovation that has kept the U.S. at the forefront of software and hardware development. Bruce Lehman, former President of the International Intellectual Property Institute, referring to the lack of protection of pharmaceutical patents prior to the TRIPS Agreement of 1994, acknowledged that “[t]he lack of any means of patenting these inventions and the related lack of experience in licensing them to the private sector, suppresses the development of commercial enterprises focused on alleviating the disease burdens common to developing countries.”²⁷ The same conclusion applies to any effort to weaken or eliminate patent protection for software innovations. The inevitable effect will be less development in the software industry and the migration of the software industry to other countries with stronger software patent protection.

In the same sense, the BSA stated in its amicus curiae brief that “as early as 1992, congressional reports recognized that ‘patent protection is of importance to the U.S. software industry, both domestically and in the global market.’”²⁸ Without patent protection, software companies face the risk that competitors will copy without penalty the essential elements of their inventions.

As recognized by Congress and industry leaders, patent protection is vital to innovation in the software industry. The availability of software patent protection in the U.S. has contributed to, if not accounted for, the U.S. becoming the world leader in software innovation. Without patent protection for software innovation in the U.S., the U.S. could see reduced investment in software technologies whereby, over time, the U.S. would lose its position as a leader in software innovation.

2. Software Patents Promote, Rather Than Stifle, Innovation

3. Software Patents Do Not Empower So-Called “Patent Trolls”

Software patents are often criticized for high litigation rates, but this perception is driven by an over competitive business climate and involvement of ubiquitous popular products such as cell phones, not an inherent feature of software patents. Given that software is the medium for modern innovation, and central to new developments in manufacturing, healthcare, and nearly every commercial field, it is no surprise that battles over markets increasingly involve software innovations and thus software patents.

4. Software is Fundamentally No Different than Other, Already Protected, Forms of Invention

The efforts to limit the patentability of “software” (based on characterizations such as “it’s just math” or “it’s inherently abstract / intangible”) necessarily presume that software is a discrete, well-defined class of technology. Moreover, these arguments starkly contrast with the accepted, and indeed unremarkable, patentability of novel circuits or special purpose machines as a class of “apparatus.”

Disallowing software patents will harm the industry because (1) there are infinite ways to implement the same functionality using different means, as illustrated by the Church-Turing thesis, and (2) software is protected in electronic-component form. It would be impossible to identify all computer implemented inventions as an infinite series of possible hardware (electronic-component) implementations, and such an exercise contradicts the positive evolution of software towards more powerful, platform-independent functionality. Software patents are thus required to protect functionality, no matter the implementation. In other words, what matters is the invention, not whether some or all of the invention may be implemented in software.

1. The Fundamentals of the Church-Turing Thesis

A Turing machine is a simple concept, but one that has had profound impacts on the computing world. Developed by Alan Turing in 1936, a Turing machine is a model of a general purpose computer, though one with unlimited memory.⁵⁰ As Turing stated in the now-famous paper, “It is possible to invent a single machine which can be used to compute any computable sequence. If this machine U is supplied with a tape on the beginning of which is written the [state description] of some computing machine M, then U will compute the same sequence as M.”⁵¹ This concept gives us our first glimpse at what are now called stored computer programs.The Turing machine created a way to more-easily prove the calculability of a process, and in defining the machine, Turing showed the machine’s equivalence to Church’s lambda-calculus theory of calculability, which achieves the same end through arguably less-intuitive means.⁵² Together, the works of Church and Turing create a formal definition of “algorithm,” meaning that if it is possible to perform a process, then it is computable on a Turing machine, and if a process is computable on a Turing machine, then it is possible to perform it.⁵³ This connection/definition is the essence of the Church-Turing thesis.⁵⁴ Therefore, if a process is calculable by a Turing machine, then it is possible to perform the process in any manner of convenience. All working software is calculable by definition because it consists of processes/algorithms, and a Turing machine defines “algorithm.” If software were not calculable, it would not be possible to think through and write working software, which is a contradiction. Therefore, being calculable, the processes performed by software may be performed equally by mechanical or electronic means specifically configured to do so.

Patents are currently allowable on such mechanical or electronic means, but software remains isolated and scrutinized as “abstract,” and potentially unpatentable under Section 101. This is inconsistent with trends in modern computing and technology.

2. How the Technological Functions Implemented in Software Are Protected in Other Forms

• 
  A specific combination of discrete components forming a circuit that performs the algorithm;
  
• 
  An application-specific integrated circuit (ASIC) or system on a chip (SoC) that has been fabricated to perform the algorithm;
  
• 
  The hardware description language (HDL) schematic for the above ASIC;
  
• 
  A field programmable gate array (FPGA) that has been physically configured to perform the algorithm;
  
• 
  A device that can configure an FPGA to physically perform the algorithm;
  
• 
  A processor coupled with a nonvolatile memory device storing machine-level instructions that, when executed, cause the processor to perform the algorithm;
  
• 
  The nonvolatile memory device itself, loaded with such machine-level instructions, which is usable with such a processor (a “computer-readable medium”);
  
• 
  A volatile memory device storing the same set of machine-level instructions;
  
• 
  The combination of machine instructions loaded into a nonvolatile and/or volatile memory;
  
• 
  Low-level source code that can be compiled to express such instructions;
  
• 
  A high-level script that can be interpreted by an interpreter into machine language that (etc.);
  
• 
  A pseudocode expression of the script; and
  
• 
  An electromagnetic “signal” expressing such instructions.
  

To make matters worse, the diversity of computing technologies is expansive and growing. Computing can be conducted by specifically engineered proteins, DNA sequences, and bacteria; by quantum computers; by purely mechanical machines powered by a hand crank or hydraulics, etc.

If borders are drawn around favored classes of computing technology, a particular logical invention can be deemed to merit patent protection when described in one form of computing technology, and yet to be deemed “abstract” when implemented in a different form of computing technology – even though the distinction is technically irrelevant, and is largely a design choice.

For example, consider a claim that is submitted as a “software system” or a “signal” or a “general-purpose processor programmed to perform certain steps. This invention could be rejected as ineligible for patent. But if the preamble recites “a codec integrated circuit” for performing certain steps, this claim is deemed patent-eligible. There is no legal basis for ascribing to that phrase the “magic words” effect of conferring patent-eligibility.

5. Patent Protection is Needed for Software Because Other Forms of Protection Are Inadequate

1. Copyright

There have been both judicial and legislative extensions to copyright law that provide some measure of protection that might overlap with that provided by patents, but closer examination reveals how limited copyright is for software. As an example, the growth in “look and feel” copyright protection in the 1980s (e.g., Whelan v. Jaslow) was followed by significant narrowing in the 1990s (e.g., Computer Associates v. Altai). Some protection for software available under the Digital Millennium Copyright Act (DMCA), particularly for anti-circumvention measures, undoubtedly helps strengthen the protections that software vendors can enforce via license agreements. At the same time, however, express limitations on copyright protection for computer software, whether in the U.S. under 17 USC §117 or in other countries makes copyright protection very much a “lowest common denominator” analysis. Other limitations, such as fair use (17 U.S.C. § 107) likewise are often employed to prevent a software copyright owner from enforcing rights in anything close to a manner equivalent to what is available under patent law. Finally, while copyrights can protect certain software features such as graphics-intensive user interfaces fairly well, they are not nearly as well suited for protection of software functions—in fact copyright can only protect such functions incidentally to other features. See, e.g., Storage Technology Corp. v. Custom Hardware Eng’g & Consulting, Inc., 421 F.3d 1307, 1312 (Fed. Cir. 2005), reh’g denied, 431 F.3d 1374 (Fed, Cir. 2005) (even where only a portion of code is required to be loaded into random access memory for equipment to operate, the entirety of the corresponding program was subject to Section 117’s provisions, protecting defendants’ right to copy that program).

The failure of copyright to be a proxy for patent protection is seen in practice by the large numbers of software vendors, large and small, who have embraced patent protection even where both patent and copyright protection is available. In fact, numerous software vendors have even opted to embrace design patent protection rather than copyright protection in clearly overlapping areas. Given that patents cost tens of thousands of dollars to procure and maintain, while copyrights have at best a nominal cost, this alone shows that copyright is not thought to provide protection anywhere close to that of patents.

One good example of copyright’s limitations to protect software-implemented innovations is found in Lexmark Int’l v. Static Control Components, 387 F.3d 522 (6^th Cir. 2004), reh’g denied, 2004 U.S. Appl. LEXIS 27422 (Dec. 29, 2004), reh’g en banc denied, 2005 U.S. App. LEXIS 33330 (6^th Cir. Feb. 15, 2005). Lexmark marketed toner cartridges with embedded program code that allowed, for example, an indication of how much toner was remaining. The printer authenticated cartridges upon installation. Defendant SCC made microchips that mimicked the authentication sequence and admittedly included, verbatim, a small program Lexmark had on its microchips. Lexmark sued for both copyright infringement due to the copying and for violation of the anti-circumvention provisions of the DMCA. The Sixth Circuit ruled that compatibility considerations brought into play both the merger and scenes a faire doctrines of copyright law, dooming the copyright infringement claim. The Sixth Circuit also ruled that anti-circumvention protection is only available to restrict access to the literal code of the chip, which did not exist in this case. Comments by judges in concurring opinions cast doubt on whether the DMCA could or should be used to “create monopolies for replacement parts” and that fair use might be a defense if any such attempt were made. Here, the focus of copyright law on original expressions rather than products or innovations per se is shown to make it very difficult to obtain anything close to patent-like protection.

2. Trade Secrets

First, trade secrecy is only available for a subset of software-related innovations: those that are not evident to a user immediately upon use. Thus, while it may be perfectly fine to protect a sorting algorithm hidden deep inside proprietary source code as a trade secret, it may be impossible to protect a highly innovative user interface as a trade secret because the user interface is immediately evident to anyone using the product. Traditionally, trade secrecy has been favored for the “internals” of software while patents have been the mainstay of protection for the “externals,” with the possible exceptions of software “apps” that may have limited novelty.

Third, trade secrecy remains a poor cousin of other forms of intellectual property because there is no federal law under which a party can bring a civil action for trade secret misappropriation, and there is little treaty-based international protection. There is currently a strong move in Congress to enact such federal legislation, but until that time, the differences among state laws and the inability to readily reach foreign defendants or reach across state borders for enforcement create weaknesses for trade secret protection compared to patent protection.

3. vis-à-vis patents.Trademark

6. Applicants Need Clarity and Stability in the Law of Subject Matter Eligibility to Effectively Seek Patent Protection

   1. Alice and Bilski Have Caused Confusion Regarding Section 101

Recently, each time the Federal Circuit or the Supreme Court issued a significant decision invalidating a “software” patent on Section 101 grounds, the USPTO had to adjust its guidelines for protecting computer-implemented inventions (e.g., MPEP 2106), and likewise, patent practitioners had to adjust their approach to drafting software patents. Given that the courts’ interpretations of Section 101 have recently changed numerous times, it has been all but impossible for practitioners to draft patent applications with a degree of comfort that the resulting patent will survive a future challenge under Section 101. This is an untenable situation. The Section 101 patent eligibility test for software inventions cannot remain a moving target for patent practitioners.

2. Patent applications are drafted according to the interpretation of patent eligibility in effect at the time of application preparation. Patent practitioners interpret the rules/guidelines from the USPTO, court decisions, and prior prosecution experience (e.g., rejections provided by examiners) in order to identify appropriate claim and specification language. Practitioners prepare the specification and claims prospectively and only later discover whether such language will be acceptable to the USPTO, in accordance with the USPTO’s interpretation of the then-current patent eligibility standard, or even later and only if an infringement action goes to trial or summary judgment, to the courts.

“Good” (i.e., valid) patents are not as often the subject of litigation, review, or appeals, so practitioners are typically left with the “borderline” or “bad” patents as the subject of much case law. Bad cases typically make bad case law with an emphasis on what not to do, with no clear guidance as to what is desirable or even acceptable. This lack of good examples in the case law also makes it difficult for the practitioner to advise his or her client on matters of patent eligibility and is the driving force behind preparing this Software White Paper.

The Alice court acknowledged that, at some level, all inventions embody an abstract idea. Patent applications drafted pre-Alice are constrained by existing language or constructs that were established or approved under earlier standards (e.g., by the courts or the USPTO) with little or no consideration as to whether an abstract idea may be present. Alice places a cloud over all such patents. While these standards may have been difficult to follow pre-Alice, Alice’s notion that all inventions may somehow implicate a statutory exception, coupled with the Court’s explicit refusal to further define what an “abstract idea” is, make it highly challenging to describe an invention—any invention, software or otherwise—that does not in some way represent an embodiment of an abstract idea. Of course, this statement is made in an environment where an “abstract idea” is undefined. As noted by District Judge Wu in Eclipse IP LLC v. McKinley Equipment Corporation, [cite] in the absence of a definition of an abstract idea, the abstract idea exception evokes like Justice Stewart’s most famous phrase: “I shall not today attempt further to define the kinds of material I understand to be embraced within that shorthand description; and perhaps I could never succeed in intelligibly doing so. But I know it when I see it….” See Jacobellis v. State of Ohio, 378 U.S. 184, 197 (1964); cf. Alice, 134 S. Ct. at 2357 (“In any event, we need not labor to delimit the precise contours of the ‘abstract ideas’ category in this case.”). The doctrine of abstractness has been around for a long time, but no one has ever been able to define or describe it in a useful and workable way. The courts must do better than this.

Inconsistent Examination by the USPTO as to Section 101 Requirements

Patent examiners have not had clear guidance on what constitutes patent eligibility under Section 101. Particularly when an invention covers a technical area that is completely new, language is changing, and technology is advancing rapidly. With no clear guidance, examiners have no option but to cite the Section 101 form rejection without any real analysis and try to encourage applicants to amend claims to recite “something more,” another undefined and poorly articulated standard from the Court. As of the time of this writing, the USPTO has not issued its final Alice guidelines, so the Section 101 rejections thus far have been issued in a rote fashion with no understanding on the part of the bar as to how to respond and even less of an idea by the examiner as to how to consider the response.

In addition, different patent examiners apply the USPTO guidance and rules differently. An examiner’s approach may be determined by his or her supervisor’s opinion. Worse yet, Section 101 rejections are often evaluated by a Section 101 “guru” within each technology center. The examiner cannot speak for the “guru” and the applicant cannot speak to the “guru,” so overcoming Section 101 rejections becomes needlessly convoluted. Consequently, practitioners get different decisions from different examiners on similar innovations. This leads to further confusion as to where the boundary line between eligible and ineligible inventions lies.

Call to Action: Let’s Work Together to Develop Clear Guidance as to What Makes a Good Software Patent

Though the tide is clearly running against software patents in the courts, no court decision (or legislation, for that matter) has categorically stated that software is patent ineligible. In fact, many decisions (including Alice) have taken pains to indicate that the intent is not to outlaw all inventions that embody software.

However, the current “wait and see” approach to software patent eligibility is unworkable. Software innovators and their business partners will not invest in research and development at the same levels, and will not seek patent protection if there is a high level of uncertainty as to whether their inventions are patent-eligible. This will, in turn, hinder innovation, as discussed above.

Moreover, current owners of legitimate software patents that were drafted and granted by the USPTO under a different Section 101 standard now face an uphill battle when attempting to engage in technology transfer, licensing, and enforcement of their hard-earned intellectual property rights. Many of these patents drafted under a different legal standard are now likely to be called into question as a result of the new Alice test, regardless of the substantive merits of the innovations described by these patents. The “know it when we see it” approach to determining patent eligibility ex post (after the patent application has been prepared, filed, and issued) fosters an attitude of disrespect for patent rights that usurps the legitimately obtained patent rights of the original innovators.

Attitudes toward software patents have ebbed and flowed over the years, but software inventions have always remained—at some level—patent eligible. No valid reason has been advanced to inject the intolerable uncertainty that the recent 101 cases (particularly Alice) have created in the area of subject matter eligibility for software. Undermining patent protection for software inventions are patent-ineligible will have profound effects--not only in the software industry, which is full of innovation, but in and across all other industries in which software is used. discontinue patent protection for software. Categorically ruling that all software inventions are patent-ineligible would have profound effects—not only in the software industry, but also in and across all other industries in which software is used.

Clear guidance as to the requirements for software patent eligibility is needed. Patent owners, licensees, and investors in software technology, as well as patent practitioners, patent examiners and courts, all need to be able to consistently and reliably determine whether the criteria for a valid software patent have been met.

For at least these reasons, we urge the AIPLA to become actively involved in filing amicus briefs defending software patents in appropriate cases addressing Section 101, to lobby Congress for legislation overturning the most onerous aspects of the Bilski and Alice decisions, and to work with the USPTO to establish guidelines for drafting, claiming, and prosecuting patent applications directed to innovations implemented in software. The following sections offer modest proposals from members of the ECLC.

3. The AIPLA Should File Amicus Briefs in All Federal Circuit and Supreme Court Cases Addressing the Issue of Patentable Subject Matter

4. The AIPLA Should Lobby Congress to Modify Section 101 of the Patent Statute to Secure Appropriate Protections for Software-Implemented Inventions

\

Section 101 of the Patent Statute has not been amended since its adoption in 1952. We recognize that amending Section 101 will require mindful debate and thoughtful action by a reticent Congress. We also recognize that other stakeholders will have much to say about the contents of such an amendment. However, the AIPLA can lead the way.

1. Return to State Street: Useful, Concrete, and Tangible Result

2. Make it Clear that Section 101 is a Coarse Filter

Eligibility for a patent under this Section shall be determined by considering the claimed invention as a whole independent of any other Sections of this Title.

This proposed legislation would negate the most pernicious aspects of the Alice decision: parsing of claims to pull out an abstract concept and consideration of the claim breadth under Section 112 and novelty and nonobviousness of the claimed invention under Sections 102 and 103 as part of the patentable subject matter inquiry.

The AIPLA Should Work with the USPTO to Ensure that the USPTO’s Patentability Guidelines Are Workable to Ensure that Meritorious Software-Implemented Inventions Remain Patentable

The Bilski and Alice decisions and the USPTO’s response to these decisions has created unprecedented uncertainty in connection with the patentability of innovations implemented in software. The software industry is too important for the AIPLA to allow the erosion of patent rights for software-implemented inventions to continue. The AIPLA must redouble its efforts to persuade the courts, Congress, and the USPTO to take steps to remove the uncertainty around software patentability while preserving a degree of protection that will allow the software industry to continue to thrive. Practitioners within the ECLC recognize that this will not be an easy task and that we may have to move outside of our comfort zones and perhaps change the way we draft and prosecute patent applications implementing software. But something must be done. The ECLC is ready, willing and able to act. It is our hope that the leadership of the AIPLA will lead the way.