A Look Back on the Development of Software IP Law

A Look Back on the Development of Software IP Law

Aviv Gaon on Mr. David L. Hayes’ opening session presentation at the 20th Annual BCLT/BTLJ Symposium, Berkeley, California

Setting the Stage for the Future

Last April, I had the pleasure of participating in the 20th annual Berkeley Center for Law & Technology and Berkeley Technology Law Journal Symposium. The symposium was focused on the past, present, and future of software intellectual property (“IP”) from a theoretical as well as a practical perspective in the aftermath of the US Supreme Court decision in Alice Corp. (2014) and the denial of Google’s petition for certiorari in the Oracle case.[1] The symposium also generated remarkable discussions that called attention to contemporary IP technology industry dilemmas. In a way, the future development of IP law (as the Cheshire Cat states in Alice in Wonderland) ‘depends a good deal on where [we] want to get to.’

Technological Advancements: A Look-Back

In his presentation, Mr. David Hayes, a partner at Fenwick & West LLP and leading IP expert, described changes in IP law due to technological advancements over the past several decades.

In the first part of his presentation, Mr. Hayes laid out the relevant technological developments between the 1960s and the end of the millennium’s first decade. The first innovative landmark goes back to the mainframe era, with the release of IBM’s first transistor in 1959. In the 1970s, Digital Equipment Corporation launched its PDP 11 minicomputers. The 1980s was the mini PC decade, and in 1993, the Mosaic web browser initiated the Internet revolution. The millennium saw the rise of cloud computing and open-source software. However, the most recent milestone in our current decade can be attributed to Google’s Android and Apple’s iOS operating systems, which have expedited the development of cellphone and tablet applications.

Divide and Conquer: Patents, Copyright, and Trade Secrets

While the above-noted advancements have significantly impacted how society uses technology, the law has not responded to every change in kind. Indeed, changes in law often lag behind changes in society and technological development. In the following paragraphs, I will provide a bird's-eye view of the developments of software protection in three different dimensions: Patents, Copyright, and Trade-Secrets.


Mr. Hayes described the rise and fall of patents as a protection mechanism, tracing it from the Freeman-Walter-Abele Test – synthesized from the three US Supreme Court cases Gottschalk v. Benson (1972), Parker v. Flook (1978), and Diamond v. Diehr (1981) – to the 1994 the Federal Circuit decision In re Alappat that developed the Useful, Concrete, and Tangible Test. In Bilski (2008), the Court rejected both the Freeman-Walter-Abele Test and the Useful, Concrete, and Tangible Test, adopting instead the Machine or Transformation Test. However, just two years later in Bilski (2010), the US Supreme Court rejected the Machine or Transformation Test as the sole test of process patent eligibility, holding the test to be ‘a useful and important clue, an investigative tool, for determining whether some claimed inventions are processes under §101.’ The Court held the claims unpatentable because they were directed toward the abstract idea of hedging risk. The latest milestone in the patent saga has been the well-known case of Alice Corp. (2014). In Alice, the US Supreme Court established The Two-Step Test. Consequently, as Mr. Hayes explained, there has been a near complete collapse of business method patents and a substantial decrease in other patents directed at computer-implemented innovations.[2]


Computer programs and copyright protection have never been considered a good match. In the first stages of computer software development in the 1960s and 1970s, IP law did not offer protection for software. In 1974, the US Congress decided to establish the Commission on New Technological Uses of Copyrighted Works (CONTU). After years of deliberation, CONTU concluded that computer programs are indeed forms of expression and should be accorded the same rights as other works of expression, subsequently leading to the Computer Software Act of 1980. Following this, the Third Circuit Court decided in the case of Apple Computer, Inc. (1983) to extend copyright protection to operating systems and application programs, even if fixed in source code or embodied in ROM only. According to Williams Elecs., Inc. (1982), both source code and object would henceforth be protected. In these decisions, the courts acknowledged that a computer program is not only the code itself but also the product of the codes.

The difficulties of distinguishing between the idea and the expression in computer software led to uncertainty in the outcome of copyright infringement cases. Thus in Altai (1992), the Second Circuit Court stepped in and implemented a more reliable analytical framework, the Abstraction-Filtration-Comparison Test. In Oracle Am Inc., the US Federal Court reaffirmed this test.

Trade Secret Protection

According to Mr. Hayes, trade secret protection was important during the mainframe era (the 1960s), since at the time software was rarely distributed in source code form. However, in the PC era (the 1990s), the importance of trade secret protection was diminished. Mr. Hayes provides two reasons for the trade secret decline: the rise of copyright protection and the Internet revolution, which made it relatively easy to view the functional coding behind webpages.[3] The decrease of trade secret protection began to level out with the rise of Saleforce.com cloud computing company in 1999. The advent of cloud computing at the beginning of 1999 also heralded multiplayer games like World of Warcraft (2004) and programs like Microsoft Office365 (2011) that contributed to the rise of trade secret protection.[4] Until the 2014 decision in Alice Corp., the use of trade secret protection in the software industry remained fairly constant. Mr. Hayes opined that the US Supreme Court decision in Alice might enhance the importance of trade secret protection in the future, mainly due to the diminished possibility of patent protection.

Where Do We Want to Go Today?

Mr. Hayes ended his presentation by synthesizing seven key characteristics of software that will affect the type of IP protection appropriate in the foreseeable future:

  1. Software is inherently functional. Though it requires creativity, the goal is a functional outcome. Creativity is considered a disadvantage for computer programs, and in most cases, programmers combine efforts to create the simplest, easy, and straightforward program.
  2. Software embodies multiple types of creativity. It may contain traditional copyrightable expression (as in the case of a game), line-by-line code, non-literal internal elements (such data abstractions, schemas, ), and external interfaces (such as APIs, input formats, GUI); all of which can be embodied in one program.
  3. Software evolution is often incremental. There is a limited record of prior art that often exists only in other source code, making it difficult to determine where patent protection is merited. Protection of object-oriented incremental changes in the code may not be imperative because of their small size and frequent reuse of code; in this case, innovation may be avoided. An efficient design is the simplest one.
  4. Software is increasingly short-lived. Mainframe software generally lasted 10 to 20 years; PC software could last six months to two years, and today’s mobile applications have an effective lifetime of a few weeks.
  5. Software development methodology has evolved. In the mainframe era, small development teams, usually employed by a hardware vendor like IBM, worked over an extended period of time to create complicated programs. Now, large enterprise development teams can develop incremental improvements in software multiple times per day.
  6. Software exists in many different markets. There are small, discrete markets for mainframe software, medium-sized markets for enterprise software, mass markets for consumer software, and essentially infinite markets for mobile applications. Each of these markets has unique characteristics driving different types of IP needs over different durations.
  7. Software has many distribution and use architectures. These strongly influence the type of IP protection that may be appropriate for a particular business model. Software can be on-site, client/server, cloud computing (private/public), SaaS/ASP/on-demand, grid computing, or involve other forms of distributed computing technologies such as peer-to-peer connections (with or without authorization) and mobile computing – all with their own intended distribution and life-cycle characteristics.

The conference’s main theme pondered the question: ‘What comes next?’ What is going to be the next test or idea that the courts will adopt? While the industry craves simple answers and reasonable legal method, the complexity of such tests, along with never-ending technological developments, leave the courts playing catch-up and implementing difficult and sometimes contradictory tests, and in the midst of this legal maze is the legal community trying to foresee the future in an unstable environment.

I posit that we are heading towards a more flexible era. In the future, the technology itself may set the rules of the legal ‘game’ and fix the boundaries. The legal role in this future may well remain in the outskirts of the technology realm with the courts restricting their judicial review to the reasonableness standard and deferring to the expert opinions of engineers and software developers.[5]


Aviv Gaon is a Ph.D. candidate at Osgoode Hall Law School. His research explores whether AI creations can have a claim to copyright protection. The author is grateful to Paul Blizzard, JD candidate at Osgoode Hall Law School and a Professional Engineer, for helpful comments and suggestions for this article.



[1] It should be noted that on May 26, only one month after the conference concluded, a federal jury decided for Google against Oracle claiming that Google Android system did not infringe Oracle copyright Java API’s since Google’s use is considered a “fair-use”.

[2] The main difficulty in Alice relates to the requirement that the software will add something extra, embodying the concept of incentive. As explained further here, software must be simple in order to be useful. Alice’s innovative arguments make it much harder to produce applicable candidates for patent protection that are both useful software and also encompass ‘something extra’.

[3] While some software languages such as JavaScript are interpreted by the browser and thus visible, much of the code running on the server itself is still under trade secret protection, and thus not viewable by the user.

[4] It’s easier to protect cloud-based programs since large portions of the code (the trade secret candidate) exist on secure severs rather than on end users’ computers (as in traditional software). Hence, the common method of reverse engineering (the main loopholes of trade secret protection) is much harder to apply.

[5] In recent decades, courts have taken a very intrusive approach; they examine the technology thoroughly, scrutinizing each component to decide whether the technology meets creativity requirements. This is very different from the legal approach in administrative procedures, for example, where courts only seek to decide whether the decision or claim at hand is reasonable. Courts do not interfere with reasonable decisions, leaving the matter for government agencies. Another example comes from torts claim, where courts rely on the opinion of experts (though in a different way, and to a different extent, than in administrative appeals).