Open Source and Diversity

Open source still has a long way to go before software, and software development alike, becomes more equitable. Open source has proved useful in building a community of developers and creating high-quality products, but it is also reveals some of the worst parts of humanity. Discrimination is embedded in small parts of open source practices and licenses, and it is important to shed light on these issues in order to make the open source community more diverse and inclusive.

Github’s mission is to “support a community where 27 million people learn, share, and work together to build software.” However, it is also a place where code contains extremely “racist, sexist, and homophobic language.” The developer community is intended to be self-policing, but how can we ensure that open source maintains its freedom while, at the same time, placing restrictions on abusers of open source platforms? Analogous to political leaders, open source leaders must set good examples and promote inclusivity in open source projects. Former president of the Open Source Initiative (OSI), Russ Nelson, did just the opposite. Nelson was removed from his position a mere month after taking office for racist posts. Though he claimed that the mission of the OSI was extremely important to him and that he hoped “the community could continue its focus on working together to advance the integration of open source software into the wider society,” his personal blog included one post that read, “blacks are lazy.” While some immediately brought attention to his racist remarks, others in the open source community claimed that this was simply an “unfortunate circumstance” for Nelson to be in.

Open source community leaders must value diversity in order to maintain the freedom and fairness reiterated in the open source mission. Frannie Zlotnick, a Github data scientist who leads the Open Source Survey, believes that open source project managers can increase diversity by making sure “that all of their employees have a chance to contribute to open source on the job.” This would allow employees from all walks of life to become familiar with open source and have their voices heard within the community. Patricia Torvalds, daughter of Linus Torvalds, describes how many open source leaders argue that the tech world is a meritocracy and that “if someone is good enough at their job, their gender or race or sexual orientation doesn’t matter.” A meritocracy “assumes a level playing field, in which everyone has access to the same resources, free time, and common life experiences to draw upon. It fails to take into account the barriers marginalized people face in contributing to open source projects. Torvalds points out that the meritocracy argument is a cop-out and that “the lack of diversity is a mistake, and that we should be taking responsibility for it and actively try to make it better.”

One of the ways open source projects attempt to honor diversity is by including a Code of Conduct. A Code of Conduct “establishes expectations for behavior for your project’s participants” and is intended to breed positivity and “facilitate healthy, constructive community behavior.” However, simply including a Code of Conduct doesn’t immediately protect people from harassment or discrimination. It is the duty of a project maintainer to ensure that the guidelines in the code of conduct are followed. In order to place responsibility into the owners of the repository, open source project leaders should include a Contributor Covenant. The Contributor Covenant details a pledge for contributors and maintainers to “make participation in our project and our community a harassment-free experience for everyone” regardless of gender, sexual orientation, race, and so on.

Another way to make open source more inclusive is to include better documentation in a variety of languages. Research shows that only 21% of open source developers speak English as a first language. Breaking this language barrier involves communicating clearly and avoiding use of technical jargon that might exclude contributors. Research shows that overuse of technical jargon might also be related to gender issues, as men report higher levels of confidence in their coding capabilities and tend to use complicated language to demonstrate their skills. Statistics also reveal how gendered the technical world is. Of 5,500 open source developers surveyed, 95% of respondents were male, and according to the U.S. Census Bureau, “black, Asian, and Latino programmers account for a total of about 34 percent of programmers in the US.”

Proponents of free software and open source raise a fair point about the security of individuals being at risk when groups of people are able to control users by limiting software for their own personal interests. The strictness of free software licensing in preventing future derivatives of software from switching licenses actually protects individual rights. Open source allows for more flexibility in licensing, but this can be a problem when it comes to making software more equitable. Licenses such as the MIT License and the Apache License allow developers to suddenly trade the benefits of Open Source for the interests of a single corporation. This raises issues of class and power, which oftentimes come hand in hand with racial discrimination, as the interests of marginalized groups of people are often cast away for the interests and monetary gain of individuals or corporations.

One example in which we witness the problem with incorporating proprietary software and open source software is seen in the case of Google and Android. Google’s Android licensing integrates aspects of proprietary software. For example, much of the core code, such as the code required for the Android Open Source Project (AOSP) to boot, is not open. The confusion about whether or not Android is actually an open source project might be related to the unclear distinction between Google Android and the Android Open Source Project. Google Android adopts proprietary software practices, some of which can be problematic when it comes to making software more equitable. Companies can convert open source software to proprietary software as soon as the open source status becomes a monetary disadvantage. In many cases, this reduces the ability for all groups of people to use or contribute to Android, particularly people who are not financially well off.

Why is diversity in open source so important? As artificial intelligence technology becomes increasingly powerful, it is important that we train systems to respect everyone equally. In 2016, Microsoft released Tay the Twitter bot, an AI chatbot that learned from tweets. Shortly after Tay was released, the chatbot was trained to tweet all sorts of “misogynistic, racist, and Donald Trumpist remarks.” This example shows how susceptible AI is to human biases, revealing the worst parts of humanity.

Another reason is that software is intended to meet the needs of many groups of people. One argument for increased diversity in open source is the idea of “vendor lock-in,” which is when a company is tied down to a single supplier. In the open source world, we can think of vendors as contributors. A narrow community of developers limits “the range of experiences from which we can draw.” It is easier to catch bugs and find quicker solutions in a team with many diverse members. While small, micro tasks such as writing single lines of code might not benefit from diversity, larger and more important tasks such as code design certainly will. If the goal of open source is to create better products, then the diversity of contributors will translate into the applicability of software for a wider audience. Lack of diversity impacts the quality of open source consumer applications because people with more diverse background can help figure out how to make software more user friendly. Data from the Open Source Survey backs up the fact that lack of diversity and negative experiences on open source projects actually impact the quality of the project. Approximately 21% of people have witnessed or experienced a negative interaction while contributing to an open source project, and of these people, 8% chose to stop contributing. Project leaders lose talented developers by failing to enforce a positive, inclusive working environment. To demonstrate how pervasive this bad reputation open source has, more than 50% of contributors have witnessed a negative interaction on an open source project. In order to create better products and address the needs of all groups of people, we must ensure that every voice is heard. Making the open source community more diverse is the necessary first step towards a more equitable future.

Citations

Levi, Ran. “The History of Open Source & Free Software.” Curious Minds(audio blog), June 3, 2016. Accessed December 18, 2017. https://www.cmpod.net/all-transcripts/history-open-source-free-software-text/.

“The Open Source Definition (Annotated).” Open Source Initiative. Accessed December 18, 2017. https://opensource.org/osd-annotated.

“Build Software Better, Together.” GitHub. Accessed January 18, 2018. https://github.com/about.

Horn, Leslie. “There Is Blatant Racist and Sexist Language Hiding in Open Source Code.” Gizmodo. February 1, 2013. Accessed January 18, 2018. https://gizmodo.com/5980842

Broersma, Matthew. “Open Source Head Sacked in Racism Row.” ITworld. March 8, 2005. Accessed January 18, 2018. https://www.itworld.com/article/2810378/open-source-tools/open-source-head-sacked-in-racism-row.html.

Open Source Survey. Accessed January 18, 2018. http://opensourcesurvey.org/2017/.

Finley, Klint. “Diversity in Open Source Is Even Worse Than in Tech Overall.” Wired. June 7, 2017. Accessed January 18, 2018. https://www.wired.com/2017/06/diversity-open-source-even-worse-tech-overall/.

“Contributor Covenant: A Code of Conduct for Open Source Projects.” Contributor Covenant. Accessed January 18, 2018. https://www.contributor-covenant.org/.

“Torvalds 2.0: Patricia Torvalds on Computing, College, Feminism, and Increasing Diversity in Tech.” Opensource. August 3, 2015. Accessed January 18, 2018. https://opensource.com/life/15/8/patricia-torvalds-interview.

“Your Code of Conduct.” Open Source Guides. Accessed January 18, 2018. https://opensource.guide/code-of-conduct/.

Irwin, Emma. “Diversity and Inclusion: Stop Talking and Do Your Homework.” Opensource. September 1, 2017. Accessed January 18, 2018. https://opensource.com/article/17/9/diversity-and-inclusion-innovation.

Stallman, Richard. “What is free software?” GNU Project. April 4, 2017. Accessed December 18, 2017. https://www.gnu.org/philosophy/free-sw.en.html.

Kerner, Sean Michael. “Is Android Really Open Source?” EWEEK. January 17, 2018. Accessed January 18, 2018. http://www.eweek.com/blogs/first-read/is-android-really-open-source.

Vincent, James. “Twitter taught Microsoft’s friendly AI chatbot to be a racist a**hole in less than a day.” The Verge. March 24, 2016. Accessed January 18, 2018. https://www.theverge.com/2016/3/24/11297050/tay-microsoft-chatbot-racist.

Cotton, Ben. “Why the Open Source Community Needs a Diverse Supply Chain.” Opensource. November 21, 2017. Accessed January 18, 2018. https://opensource.com/open-organization/17/11/inclusivity-supply-chain.

Gruman, Galen. “What’s Really Behind Silicon Valley’s Apparent Racism?” InfoWorld. August 2, 2016. Accessed January 18, 2018. https://www.infoworld.com/article/3098607.

Ronacher, Armin. “Diversity in Technology and Open Source.” Armin Ronacher’s Thoughts and Writings. Accessed January 18, 2018. http://lucumr.pocoo.org/2017/6/5/diversity-in-technology/.