Incorporating the Digital Commons

Front matter

=

BENJAMIN J. BIRKINBINE

Corporate Involvement in Free and Open Source Software

Contents

1. Introduction: Open Source Software and the Digital Commons 1.1. The Argument and Plan for the Book

1.2. Situating Free (Libre) and Open Source Software

1.3. Open Source Business Models

Index

121 135

List of Figures

2.1 Coupling Between Production and Reproduction Circuits 2.2 The Commons Circuit of Value

3.1 Netscape Navigator Usage Data 1994-2006

5.1 Major StarOffice Derivatives

41 44 60 98

List of Tables

1.1 Top Companies Contributing to the Linux Kernel, Versions 4.8-4.13

1.2 Types of Open Source Business Strategies 1.3 Typology of Property 1.4 Social Democratic and Critical Theories of the Commons

6.1 Centralised Versus Decentralised Peer Production

Acknowledgments

This book represents the culmination of more than six years of studying and using free and open source software, which began during my doctoral studies in the School of Journalism and Communication at the University of Oregon. I was first introduced to open source software during the summer of 2011 when a colleague, Jeremy Swartz, took me ona guided tour of the Open Source Convention (OSCON) in Portland, Oregon. At the convention, I was offered three different jobs by companies who had exhibits there. I was told, ‘we are looking for people like you. I was incredibly surprised by this because I knew practically nothing about free and open source software at that time. I had heard of Linux, but that was about it. Nonetheless, all of the exhibitors - including Facebook, Google, the New York Times, and the United States Government - seemed keenly interested in attracting free and open source software programmers to their organisations. That experience piqued my interest in why companies were so interested in open source, and it became the focus of my studies.

Along the way, I have had the privilege of discussing the subject with incredibly smart and patient people whohave helped me. I thank my dissertation committee for their patience, comments, and critiques throughout the process. Janet Wasko was an incredibly gracious and accommodating adviser throughout the process. Thanks for the support during my time at Oregon, the one-on-one meetings and, of course, all the fun. Bish Sen provided critical feedback throughout the process and always pushed me to think about the broader implications of my work. Gabriela Martinez was equally supportive, xii Incorporating the Digital Commons provided great feedback, and was always available for conversations. I thank both of them for the assistance throughout the process. Finally, I thank Eric Priest, who agreed to supervise an independent study on open source technology. My understanding is that Eric became one of the few law professors to serve on a dissertation committee at the University of Oregon.

I also thank the members of my cohort - Toby Hopp, Erica Ciszek, Francesco Somaini, Brant Burkey, and Fatoumata Sow - who were sources of inspiration, support, and friendship throughout our time in the program. In addition, Tewodros Workneh, Brenna Wolf-Monteiro, Pietro Monteiro, Randall Livingstone, Jacob Dittmer, Lauren Bratslavsky, Glenn Morris, Karen Estlund, Andre Sirois, Jolene Fisher, and Geoff Ostrove all made my time in Eugene enjoyable. I would also like to thank Kat at the University of Oregon for helping me with my first Linux install. Without that initial help, I would have probably never embarked on this adventure, and I am now happy to say that I’ve passed the gift along to others.

I also thank Christian Fuchs, Denise Rose Hansen, and Andrew Lockett at the University of Westminster for their support along the way. I feel incredibly lucky to have received a Fellowship at the Westminster Institute for Advanced Studies where I was able to expand and refine some of the ideas presented in this book. During my stay at Westminster, Ergin Bulut, Pasko Bilic, Sebastian Sevignani, Arwid Lund, Mariano Zukerfeld, Pieter Verdegem, and Maria Michalis all provided feedback and support. Also during my stay in London, a group of us attended the book launch of Massimo De Angelis’s Omnia Sunt Communia, and the ideas outlined in that book provided inspiration for many parts of this project.

This project has also benefited from the generous support of the University of Nevada, Reno, which has been my home for the past five years. I thank all my colleagues at the Reynolds School of Journalism, and especially Dean Al Stavitsky and Associate Dean Donica Mensing for their support and willingness to accommodate research activities. In addition, this project was supported by a grant from the Vice President for Research and Innovation’s office at the University of Nevada, Reno, and Ana de Bettancourt-Dias was most helpful in this regard. I also thank Rudy Leon of Evoke: Words For Hire for her assistance with indexing this book.

Last, but certainly not least, I would like to thank my family. Each of you, in your own way, provides the source of inspiration for this work. I thank my mother and father for their years of support. Without you, none of this would be possible, and I am forever indebted. My sister, Ann, inspires a wealth of adjectives: congenial, affable, good-humoured, gregarious, etc., and Pm thankful for the years of fun we've shared. My son, Caden, was only two years old when this journey began, and he is now well on his way to becoming a young man. I’ve enjoyed watching you grow over these years, and I look forward to many more. You are a blessing. Finally, my love and inspiration, Roberta, meu coracdo. You have truly been wonderful throughout all these years. You are a gift.

For the community, the commoners, Caden and meu coração.

CHAPTER |

In March of 2012, The Linux Foundation released a report entitled, ‘Linux Kernel Development: How Fast it is Going, Who is Doing It, What They are Doing, and Who is Sponsoring It. The kernel is an essential part of an operating system that facilitates communication between computer hardware and software, and the Linux kernel development project is considered ‘one of the largest cooperative software projects ever attempted’ (The Linux Foundation, 2012: 1). Aside from a technical overview of how kernel development has changed over time, the authors included a curious note in the report’s highlights: Microsoft was one of the top 20 contributors to the kernel. This marks the first time that Microsoft appeared as a top contributor, but it was not the only corporation in the top 20. Other corporate contributors included Intel, IBM, Google, Texas Instruments, Cisco, Hewlett-Packard, and Samsung, as well as others. The Linux operating system is a form of Free (Libre) and Open Source Software, or FLOSS, which allows users to freely study, use, copy, modify, adapt, or distribute the software. Why, then, would major corporations contribute directly to a FLOSS project, especially when that project seemingly does not directly contribute to corporate profits? This question becomes even more curious when one considers that many of the companies contributing to the kernel not only compete with one another in the market for information technology, but that companies like Microsoft and Google are direct competitors with Linux in the market for operating systems.

Indeed, Steve Ballmer, the Chief Operating Officer of Microsoft, once referred to Linux as ‘a cancer that attaches itself in an intellectual property sense to everything it touches’ (Greene, 2001). Ballmer was referencing the GNU General Public License, or GNU GPL, which is the most commonly used free software license. The GPL grants users of GPL-protected software the right to study, use, copy, modify, or adapt the software as they wish. In addition, users are granted the right to redistribute the software, as well as a modified version, and the user

Birkinbine B. J. 2020. Incorporating the Digital Commons: Corporate Involvement in Free and Open Source Software. Pp. 1-32. London: University of Westminster Press. DOI: https://doi.org/10.16997/book39.a. License: CC-BY-NC-ND 4.0 may even charge a fee for the modified version, provided that the distributor does not place greater restrictions on the rights granted by the GPL. The GPL does not preclude corporations from modifying free software or charging a fee for their modified versions, but the corporation must still grant free software rights to end users. Ballmer’s quote implies that free software is antithetical to commercial software companies. If this were the case, then Microsoft and other commercial software firms would have no incentive to contribute directly to one of the largest open source projects.

Furthermore, consider the fact that Ballmer made his denunciation of Linux on 1 June 2001. Merely 27 days later, on 28 June 2001, the United States Department of Justice found Microsoft guilty of monopolistic business practices in violation of the Sherman Antitrust Act primarily for bundling its Internet Explorer web browser with its Microsoft Windows operating system to rapidly increase its share of the market for web browsers. However, Microsoft has dramatically changed its position on Linux and open source since 2001, as signified by its inclusion in the top 20 contributors to the Linux kernel in 2012. That same year, Microsoft created Microsoft Open Technologies, Inc., a wholly owned subsidiary dedicated to facilitating interoperability between Microsoft and non-Microsoft technologies, while promoting open standards and open source. What changed during this 12-year period that Microsoft would so dramatically reposition itself in relation to FLOSS?

Microsoft is not alone. Indeed, corporate involvement in FLOSS has been increasing, especially since about 2007-2008. Table 1.1 provides an illustration of the companies that contributed to Linux kernel development for versions 4.8-4.13, which were released in 2017. The annual report for kernel development that year identified 225 companies that contributed to the project. While the Linux kernel is just one example of a FLOSS project to which corporations are contributing, other examples exist as well. This begs the question as to what motivates these companies to contribute to FLOSS projects. Furthermore, in what ways are they contributing to FLOSS projects? How do communities of FLOSS developers negotiate corporate involvement in their projects? Do communities of FLOSS developers have any recourse for unwanted corporate involvement or influence in their projects?

1.1. The Argument and Plan for the Book

The overall purpose of this book is to investigate the seemingly contradictory relationship between FLOSS communities and for-profit corporations. Working from a critical political economic perspective, I investigate the power dynamics that exist between communities of FLOSS developers and the corporations that sponsor FLOSS projects or appropriate the software production of FLOSS labourers. After all, FLOSS products and the productive process that make those products possible have been widely lauded as revolutionary changes that

Table 1.1: Top Companies Contributing to the Linux Kernel, Versions 4.8-4.13 (Corbet & Kroah-Hartman, 2017: 14).

Company Changes Percent Intel 10,833 13.1% none 6,819 8.2% Red Hat 5,965 7.2% Linaro 4,636 5.6% unknown 3,408 4.1% IBM 3,359 4.1% consultants 2,743 3.3% Samsung 2,633 3.2% SUSE 2,481 3.0% Google 2,477 3.0% AMD 2,215 2.7% Renesas Electronics 1,680 2.0% Mellanox 1,649 2.0% Oracle 1,402 1.7% Huawei Technologies 1,275 1.5% enable greater degrees of freedom and autonomy on behalf of users and contributors (Benkler, 2006; Raymond, 2000; Stallman, 2002). This project intervenes in these debates by tempering these claims. I position technology as a site of social struggle, and I contextualise commons-based peer production within a broader social context to illustrate how such production intersects with capitalist production. I do this by demonstrating how the purportedly revolutionary changes brought about by FLOSS and commons-based peer production are now becoming incorporated into corporate strategies and corporate structures.

The central argument presented here is that free and open source software is dialectically situated between capital and the commons. On the one hand, communities of programmers are actively working to create software as digital commons that can be accessed, used, adapted by others. By developing software iteratively this way, the pace and scale of software production increases. This represents a virtuous cycle whereby an association of software programmers actively contribute to a community that claims collective ownership over FLOSS projects. As such, FLOSS programmers can be framed as commoners insofar as they remain committed to ensuring the reproduction and sustainability of commons-based software projects over time. On the other hand, capital attempts to capture the value being produced by FLOSS communities. This includes harnessing the processes (i.e. the collective labour, or commons-based peer production power) involved in FLOSS production as well as commodifying the products (i.e. specific FLOSS projects), which can provide a basis upon which to commercially exploit the collaborative production occurring in FLOSS communities.

This is not to say that the goals of the free software commoners and capitalist firms are always antagonistic. At times they are mutually beneficial, and researchers have demonstrated how commercial sponsorship of FLOSS projects tends to make those projects more likely to attract developers and, therefore, ensures the project’s longevity (Santos, Kuk, Kon and Pearson, 2013). However, we also have other examples of these relationships breaking down, particularly when it concerns the unwanted encroachment of capital upon commonly held resources like the digital commons. In these situations, the interests of the FLOSS community diverge from those of a commercial sponsor, and the relationship becomes antagonistic. The FLOSS community is faced not only with the challenge of ensuring that their digital commons remain viable, but also with ensuring that the project maintains the sense of community that enabled the project to grow in the first place. How, then, to negotiate the relationship between their digital commons and the unwanted intrusion by capital into their projects? There are a variety of factors to consider when attempting to negotiate this relationship, and the subsequent chapters provide empirical evidence for how these dynamics manifest.

The commons, generally, and the digital commons, more specifically, can be understood as an alternative system of value that is emerging from within capitalism. At times, circuits of commons value can intersect with capital accumulation circuits. Therefore, understanding the relationship between free software and capital dialectically is useful for accounting for the contradictions between these two forces that operate according to differing logics. Chapter 2 outlines these differences more specifically by drawing on theories of capitalism, digital labour, and the commons. The purpose is to develop a critical theory of the digital commons by incorporating a critique of capitalism within theories of the commons.

In Chapters 3-5, I provide three detailed case studies that illustrate different aspects of the dynamics between FLOSS communities and corporations. I separate my discussion of corporate involvement in FLOSS into three thematic areas, with each case study providing an exemplary case of these themes. The three themes are processes, products, and politics. When considered together, these three case studies are indicative of more general tendencies of corporate involvement in FLOSS projects. Furthermore, each case study offers a nuanced understanding of the complex way these dynamics work, and they allow for a detailed unpacking of some of the contradictions inherent in the relationships.

To begin, Chapter 3 focuses on Microsofts contentious relationship with FLOSS. This relationship is indicative of the ways in which the processes involved. in FLOSS production effectively ushered in a new era of industrial software production. While other companies demonstrated a willingness to cooperate with FLOSS communities, Microsofts dominance of the software market for personal computing during the 1980s and 1990s makes it an instructive case for understanding how software production changed over time. The major historical event here is the antitrust ruling against Microsoft, which marked the end of an era in which software production was largely accomplished within a single firm that sought to exclude others from accessing its code. Indeed, one of the consent decrees in the Microsoft antitrust ruling was that Microsoft provide third parties access to its application programming interfaces (APIs). This was a radical departure from Microsoft's earlier practices, whereby the firm rose to power by using anticompetitive business practices.

Coinciding with Microsofts dominance of the software market and its eventual antitrust conviction in the 1990s were other software firms trying to find a way to transform FLOSS products into successful commercial products. My analysis of Red Hat, Inc. in Chapter 4 is indicative of how FLOSS products get incorporated into a commercial firms overall business strategy. Red Hat remains the largest and only publicly traded company providing software and services that are completely based on free software. As such, Red Hat cannot rely on traditional copyright protections to exclude others from using the underlying source code included in its software. Thus, my analysis of the firm explores how Red Hat has been able to create a profitable business based on free software.

Finally, the third case study in Chapter 5 focuses on how FLOSS communities cope with unwanted corporate influence in their projects. Sun Microsystems was an important corporate sponsor of FLOSS projects, but it was acquired by the Oracle Corporation, which had different plans for those projects. In that chapter, I focus on the diverse destinies of three such projects - the Open- Solaris operating system, the MySQL relational database management system, and the OpenOffice productivity software - and the ways that the communities involved in those projects resisted Oracle’s encroachment into their projects. In effect, the case study illustrates the politics involved in negotiating boundaries between FLOSS communities and corporations, while also demonstrating some of the strategies FLOSS communities can use to protect their projects.

In the remainder of this introduction, I provide more context for understanding the significance of FLOSS. This includes historically situating FLOSS within a broader discussion of the commons, as well as some of the key historical moments in the development of software, generally, and FLOSS, more specifically. In each of these sections, I also offer some notes on the terminology used throughout the book, which will hopefully assist in avoiding conceptual confusion. Following those sections, I discuss the cultural significance of FLOSS. I conclude the chapter with a note on the methodology used for the current study. Readers who are already familiar with the history of FLOSS and its defining characteristics may wish to skip directly to the next chapter or the note on methodology at the end of this chapter.

1.2. Situating Free (Libre) and Open Source Software

Although free software and open source communities are related and, in some cases, not mutually exclusive, each of them has distinct characteristics that can best be described by reference to the ethos underlying each movement. To contextualise the emergence of FLOSS within the evolution of the computing and software industries, a brief history of these industries is provided below. Following that discussion, I focus on situating two key figures associated with FLOSS within their historical context: Richard Stallman and Linus Torvalds. These two figures represent free software and open source, respectively.

1.2.1. Historicising Free and Open Source Software

The use of machines for processing information or calculating differences in numbers, human beings performed such work. But human calculations were, at times, prone to errors. To reduce this uncertainty, Charles Babbage, a philosopher and mathematician working at the University of Cambridge in 1822, proposed that it was ‘only by the mechanical fabrication of tables that such errors can be rendered impossible’ (Gleick, 2011: 95). Such was the proposition for Babbage’s Difference Engine, which performed routinised calculations mechanically, and was arguably the genesis for modern computers as we know them today. Later, Babbage expanded on his idea and planned a new type of machine that was capable of being controlled by instructions that could be encoded and stored to facilitate operation. The new iteration of the idea was called the Analytical Engine, but this still only provided the idea for the hardware or mechanisms necessary for such processes to occur. What was needed for this hardware was software.

The idea for software arguably originates with Augusta Ada Byron King, the Countess of Lovelace, otherwise known simply as Ada Lovelace. In 1843, she developed the idea that Babbage’s Analytical Engine could perform a series of operations beyond the mere calculation of numbers. By abstracting from the differences between two things, Lovelace posited that the Analytical Engine could be programmed to perform operations that relied on symbols and meanings, which, in turn, could be communicated to the machine. Although Lovelace’s idea was never realised in her lifetime, she is credited with developing the idea for software and is known as the first programmer.

While Babbage and Lovelace are credited as pioneers in developing the ideas for modern computers and software, the construction of such machines did not begin until World War II. Developments in the field of computer science and information theory - like Kurt Gédel’s incompleteness theorem, Alan Turing’s idea for a Universal Turing Machine, Claude Shannon’s mathematical theory of communication, and Norbert Wiener’s cybernetics - provided the intellectual inspiration for the development of such machines. Before, during, and after

World War II, many of the developments leading to modern computers were used for military purposes. Most notable, perhaps, were the German Enigma machine that was used to encrypt secret messages and the electromechanical bombes used by the United Kingdom to decipher those messages (Smith, 2011). However, in 1941, Konrad Zuse, a German electrical engineer, built the Z3, which is regarded as the first electro-mechanical, programmable, fully automatic digital computer (Zuse, 1993). The first comparable computer in the U.S. was developed by John Atanasoff at Iowa State University in 1942 (Copeland, 2006). Only one year later, the first fully functioning electronic digital computer was put to use by the cryptanalysts working at Bletchley Park in the U.K. as part of the Government Code and Cypher School. The Colossus, as the new machine was known, was programmed to decipher German communications during the war. By the end of the war, Bletchley Park had 10 Colossi working to decode German communications (Copeland, 2006).

Following these initial landmarks, the development of modern computers accelerated as many of the early pioneers began working for academic institutions and private companies after the war. In the United States, Grace Hopper, who served in the United States Navy Reserves as a member of the Women Accepted for Voluntary Emergency Service (WAVES) during World War II, was assigned to the Bureau of Ships Computation Project at Harvard University. While there, she worked on the Mark I computer project, which was built by IBM in 1944. Later, after she began working for private companies, Hopper popularised the idea of machine-independent programming languages. This led to the development of the Common Business-Oriented Language (COBOL) in 1959. Hopper is also credited with popularising ‘debugging’ as a term for removing defective material or code from a program. While Hopper may not have invented the term, she popularised it by literally removing a moth from a Mark II computer at Harvard University after it had caused the machine to short circuit (Deleris, 2006).'

During the 1960s, the creation of microprocessors drastically reduced the cost of computing. As a result, communities of hobbyist programmers and computer enthusiasts began to experiment with the technology in the following years. One notable example was the Homebrew Computer Club, started by Gordon French and Fred Moore in 1975 at the Community Computer Center in Menlo Park, California. The club provided an open forum for hobbyists to trade parts and advice about the construction of personal computers. The goal was to make computers more accessible to others. More will be said about this specific hobbyist community in Chapter 3, as it played an important role in the rise of Microsoft. Aside from these hobbyist communities, the majority of computer development occurred within the military, academic institutions, and private companies.

Most notable were the initial developments within the Defense Advanced Research Projects (DARPA), which was created in 1958, as well as the Artificial Intelligence Lab at the Massachusetts Institute of Technology (MIT), which was founded in 1970.” Programmers working at the time were using a proprietary programming language called Unix, the intellectual property rights for which were owned by AT&T. One of the programmers working at MIT was Richard Stallman, who began working in the lab in 1971. Stallman found that when he wanted to work with the Unix programming language outside of officially sanctioned spheres, he was denied access to the code by AT&T. In protest, he posted messages to computer-based bulletin boards in 1983 announcing that he was developing a Unix-based language that would be available for free so that others could use the language however they saw fit. In 1985, Stallman published “The GNU Manifesto, which outlined the goals of his new project, his reasons for developing the project, and what the project was aimed at fighting back against.* The programming language was called ‘GNU, a recursive acronym standing for ‘Gnu’s Not Unix. Along with the programming language, Stallman developed the GNU Public License (GPL), which stipulated that anyone could access the source code for free, and that anyone using the GPL agreed to make their contributions available under the same conditions. This would ensure that computer programmers could freely share their work with one another, thereby creating a common form of property that developed in opposition to its proprietary and closed counterparts.

Stallman became the figurehead of the movement against proprietary software. He viewed access to source code as a fundamental right, which he wanted others to believe in as well. He summed up this view in his famous dictum, ‘Free as in freedom, not as in free beer, thus positioning free software as a moral right (Stallman, 2002). The free software definition stipulates that ‘users have the freedom to run, copy, distribute, study, change and improve the software’ (Free Software Foundation, 2012). As the principles of free software grew beyond the borders of the U.S., others have tried to reduce the confusion over the English term ‘free’ by using the French term libre rather than gratis. Stallman established the Free Software Foundation (FSF) to promote his movement against proprietary software, and he represents an impassioned counter-cultural figure who continues to espouse his free software philosophy.

While Stallman is generally considered to be the figurehead of the free software movement, open source software is generally associated with Linus Torvalds. In many ways, Torvalds and Stallman have similar stories, but differ on philosophical terms. During the 1980s, free software projects were being developed but generally on a smaller scale. Free software had not yet found a way to coordinate efforts on a larger scale. Torvalds wanted to work on kernel development for an open-source operating system. Rather than relying on numerous programmers all working independently on such a task, Torvalds released the source code for his project, which he was calling ‘Linux, a portmanteau of his name, Linus, and the language he was working with, Minix (itself a simplified derivative of AT&T’s Unix). Torvalds suggested that anyone who was interested. in contributing to such a project was encouraged to do so, if they released their work back to the community so that others could progressively work toward completing the kernel. The project proved to be successful, and eventually led to the creation of the open source operating system, Linux. Coordinating such a large-scale programming project was accomplished by asking those working on the code to release their work, no matter how small the changes seemed. The rationale was that coordinated efforts reduce the amount of redundant work, which was summed up in the adage ‘with many eyes, all bugs are shallow, which Eric Raymond refers to as ‘Linus’s Law’ (Raymond, 2000).

Stallman and Torvalds differ with respect to how they view the relationship between free software and proprietary software. Whereas Stallman tends to be somewhat more confrontational in his opposition to proprietary software, Torvalds is less so. Williams (2002) describes a decisive moment at a conference in 1996 where Stallman and Torvalds appeared on a discussion panel together. Torvalds expressed admiration for the work that Microsoft was doing and suggested that free software advocates could work together with companies. Such a suggestion was generally seen as taboo since Stallman was perceived with esteem by the programming community, and the Free Software Foundation generally took a very adamant stance against proprietary software companies. Powell (2012) frames this distinction between free software and open source similarly:

open source software as an industrial process grew out of the culture of free software development, but departed from the latter’s political focus on the value of sharing and the maintenance of a knowledge commons, and instead focused on the efficiency of open source processes for software production (692).

This momentat the 1996 conference thus marked a watershed moment in which the fervour of the free software movement thawed a bit, as Torvalds came to represent a more liberal approach to free software. By ‘liberal’ here, I am referring to the literal definition rather than a specific political position; the term should be understood as something that indicates an openness to new perspectives or behaviours while willing to abandon traditional values. In this regard, Linus’s expression of support for the work that Microsoft was doing signalled an openness to working with Microsoft (or other commercial firms) simply to produce the best software rather than an adherence to the anti-corporate stance of Stallman and the Free Software Foundation.

In sum, then, we can understand the free software and open source movements with respect to these differing philosophical positions. Stallman and free software advocates tend to make moral claims against supporting proprietary software, while Torvalds and open source tend to be associated with a more liberal and inclusive stance. While Stallman and Torvalds have been used to illustrate the differences between free software communities and open source communities, they should not be viewed as mutually exclusive communities, nor should they be seen as representative of the entire free software and open source communities. One of the peculiarities of the free and open source software community is that, although the overall community is united in their belief that software ought to be free for users to study, modify, adapt, or customise, its members will often vehemently defend their preferred free software project while deriding others. In a sense, this signals to others where their loyalties lie and engenders stronger ties within niche communities that exist within the larger FLOSS community. The present project is less concerned with these intra-group fissures than the relationship of the community to the corporations that rely on their labour. To that end, the combined term ‘Free (Libre) and Open Source Software’ or ‘FLOSS is used to refer to the overall community.*