The Open Source Identity Crisis

As the diversity in tech movement gains traction, open source faces an identity crisis.

by Noah Slater on July 21st, 2014

The geek identity, which has been undergoing a gradual redefinition for years, is fiercely defended by those who believe they have a real and justified claim to it.

A photo of a series of wooden fences, each saying 'no entry'. A statue of a yeti is in the distant right corner beyond the fences.

CC-BY Oast House Archive

As Joseph Reagle writes:

“It’s not surprising that a subculture polices claims about its membership (i.e., the affirmation of identity claims) and its boundaries (i.e., the positing of categorical contours). In this case, what are the contours of this thing called geekdom and who is accepted within its bounds? Indeed, geekdom is actually constituted by these ongoing struggles. Some of those wrestling with geek identity are those who have found a home away from the alienation they experienced in the mainstream; it’s a subculture where enthusiasm, knowledge, and skill are appreciated and praised over things like good looks and attendant popularity.”

The alienation he describes is core to the geek identity: many geeks (myself included) were bullied as children or found it hard to fit in with their peers. The geek identity works by transforming these feelings of otherness into positive defining attributes. Geekdom also draws its identity from hyper-masculinity and hyper-whiteness — Mary Bucholtz notes that this “identity, the nerd, is racially marked precisely because individuals refuse to engage in cultural practices that originate across racialized lines and instead construct their identities by cleaving closely to the symbolic resources of an extreme whiteness, especially the resources of language.”

As foundational elements of the identity, these attributes bind geeks together in a collective feeling of being exceptional in some way, a cut above the rest: greater passion, more abundant enthusiasm, superior intellect, advanced technical ability, and so on. Programmers typically believe they are the most important part of a team and consequently deserve more pay, more power, and better working conditions. This identity and its attendant sense of entitlement has shaped much of open source culture today, and serves as the yardstick by which outsiders are judged, and permitted into or barred from open source spaces.

As the diversity in tech movement gains traction, open source faces an identity crisis. Like Reagle, I also hear the refrain: “Don’t unsettle us. Don’t make us think about the consequences of our misogyny, or our entitlement, or our privilege.” This movement is bringing unwelcome scrutiny to a culture of masculinity, whiteness, technical skill, and privilege. And in doing so, the progress we are making is seen (perhaps unconsciously) as an attempt to unseat the very definition of what it means to be an open source contributor.

Shared Stories

If we want to understand how this identity is constructed, we have to look at the stories we tell each other that help make sense of it. Social dominance theory has this concept of the hierarchy-enhancing legitimizing myth. These myths, or stories, support the status-quo by justifying the inequalities that keep the dominant class in a position of power.

The first story we must look at is: who exactly is an open source contributor? The answer to this seems obvious enough. Open source contributors are hackers. The movement as a whole traces its roots back to Richard Stallman and the hackers at MIT in the late 70s. The hacker identity is a specialisation of the geek identity, which can be seen as a continuation of the scientist and mathematician archetype that stretches back centuries.

But this picture is wrong. And what I’d like to suggest is that an open source contributor is someone who contributes to open source. Nothing more complex than that. But this coupling of the open source identity and the hacker identity is so strong that most people do not realise a separation is even possible!

This is a big problem for open source, because it divides contributors into two classes. You have the first class contributors (who fit the hacker identity and can code) and then you have everyone else: second class contributors with second class skills. An underclass whose legitimacy (or lack thereof) is judged in reference to the archetypal hacker.

We mythologise men like Linus Torvalds who then serve as prototypes from which individual hacker identities are cut. It is unfortunate, then, that Torvalds is known (content warning: verbal abuse and ableism) for his overt acts of hyper-masculinity. Indeed, empathy, patience, humility, etc. are all shunned in favour of an ignoble dedication to “logic” and “hard data”.

The top of an entryway reading 'Linus Torvalds Auditorium'.

CC-BY Debarshi Ray, filtered.

We’ve spent decades trying to keep the “wrong sorts” out of programming by making it seem more difficult than it is. Melissa Pierce, who is doing research for a film about Grace Hopper, shared some tentative observations on Twitter that suggest this may have had something to do with a panic about job security 60s and 70s.

First tweet in a series by user @melissapierce, reading '1. I research the history of computing and the people who designed degrees were worried about making programming look too easy....'

Tweet source

Second tweet in the series, reading '2. In 60/70s these guys worried they'd be replaced by machines, so they started this kind of math cult priesthood around programming...'

Tweet source

Third tweet in the series, reading '3. By the time programmers realized compilers wouldn't make them obsolete, they'd already poisoned the water with the 'so complicated' myth.'

Tweet source

Sometimes this gatekeeping is subtle, like the spurious notion that understanding programming is dependent on understanding mathematics. (Music theory is mathematics too, but that fact is not used to keep non-mathematicians out of music.) And sometimes it is obvious, like magical metaphors that imply programmers are essentially a different stock of human.

The Purity of Motivation

The archetypal hacker is a white heterosexual cisgendered man who dedicates all of his free time to coding. We sublimate this archetype into a collection of intersecting axes of purity by which people and their behaviours can be judged.

There is an expectation of single-minded devotion to the craft. Eric S. Raymond’s Jargon File (which I regrettably devoured during a glutinous phase of cultural assimilation) even describes a “larval stage” that people must pass through in order to be considered a true hacker. A sort of “rite of passage” if you will, it is described thus (emphasis mine):

“…a period of monomaniacal concentration on coding apparently passed through by all fledgling hackers. Common symptoms include the perpetration of more than one 36-hour hacking run in a given week; neglect of all other activities including usual basics like food, sleep, and personal hygiene; and a chronic case of advanced bleary-eye. Can last from 6 months to 2 years, the apparent median being around 18 months. A few so afflicted never resume a more ‘normal’ life, but the ordeal seems to be necessary to produce really wizardly (as opposed to merely competent) programmers[...]“”

Echoing this, David Heinemeier-Hansson argues that by “definition, involvement in open source requries [sic] at least some passion. Otherwise why would you forgo laying on the beach, on the couch, or in your bed all those hours spent crouched [sic?] in front of your screen?”

An image of hands typing on a Mac Book Air, code displayed on the screen.

CC-BY hackNY.org, filtered.

That he believes open source contribution is done in lieu of relaxing on the beach reveals a lot about the sort of person he expects to be contributing.

Indeed, we can think about this as purity along the axes of motivation. “True” open source contributors do it because they are “passionate” about it, and if you need (for example) financial support, then you are tainted. Heinemeier-Hansson even goes so far as to describe it as “selling out”, suggesting that money threatens to “corrupt” open source. This is textbook boundary policing language.

Unfortunately, these legitimizing myths about what constitutes “pure” motivation shut out anyone who isn’t privileged enough to have abundant free time, which is a sort of largely unacknowledged economic wealth. People with demanding jobs (perhaps more than one), a family, or other difficult life situations are stigmatised in the process. It also marginalises people who may be contributing for any number of other reasons, including skill acquisition, career building, or practicality (i.e. something is broken and needs fixing).

The Purity of Skillset

Perhaps the most obvious way in which the hacker identity has a hold over the open source identity is this notion that you have to code to contribute to open source. Much like technical talent is centered in the tech industry, code is seen as the one true way to contribute. This can be such a powerful idea that documentation, design, marketing, and so on are often seen as largely irrelevant. And even when this isn’t the case, they are seen as second class skills. For many hackers, open source is an escape from professional environments where collaboration with these “lesser”, more “mainstream” activities is mandatory.

What this means is that even if non-technical contributions are welcomed by a project, it can be hard to earn any official recognition through them. This is important, because on most projects, recognition also means admission into the power structure. Failure to recognise this sort of contribution shuts out an entire class of people from open source. It is tantamount to saying that their contributions are nice, but not enough to earn a seat at the table. Not surprisingly then, projects that do this can expect to receive fewer contributions of this type.

This focus on code as the primary method of contribution can be seen as purity along the axis of skillset. This doubles as a manifestation of the hard/soft spectrum, a centuries old idea we inherited from the sciences. Originally meant to be a hierarchy from theoretical to applied, it quickly became a hierarchy of legitimacy and gender.

Hard skills are seen as masculine, and soft skills are seen as feminine. Software was originally a soft skill and predominantly a woman’s field, being seen as secretarial work. As men started to enter the space (from hardware, naturally) it became seen as a hard skill, somewhat at odds with its name. This practice of masculinising a field and pushing out the women is something we see repeated over and over again; the rise of the “growth hacker” being one recent example. Doing so allows men to safely perform their masculinity, whilst simultaneously forming a hierarchy-enhancing legitimizing narrative that keeps women out.

Meritocracy as a Metastory

The meritocracy is a sort of metastory. It’s a story in so much as it’s something we repeat to each other. But it’s a metastory in so much as we use it to construct other stories. A story factory, if you will. When someone is included or excluded from our project, the meritocracy can be used to tell a smaller, specific story about why that is the case.

The meritocracy is a boundary policing tool, in the very literal sense of determining who is permitted access to open source power structures. We think we’re judging people objectively, but that is demonstrably not the case. (The homogeneity of open source being the most obvious indicator that something is wrong with this picture.) Studies show the meritocracy actually exacerbates in-group bias, ensuring that the only people we let into the community are those similar to us. And as Nóirín Plunkett points out, the meritocracy also disproportionately rewards self-confidence and assertiveness, thereby institutionalising a gender gap.

“The meritocracy” is a sort of newspeak, then, for the act of judging people (with no accountability) against what we, privately, think contributors ought to look like. In this sense, it is perhaps the most insidious hierarchy-enhancing legitimising myth in open source. Not only for its power to convince us that the structural inequality is just, but in its near universal application to any sort of perceived impurity on an individual-by-individual basis.

This ought to be a concern for anybody who cares about open source. The monoculture is holding us back. To pick one example, good design and UX are practically non-existent. This is a big problem, as I’m sure anyone who’s ever (tried) to use open source software will agree. And by excluding marginalised groups, we’ve built a culture that starves us of the very skills and perspectives that open source needs the most.

Liminality

There is some good news, however. The diversity conversation is getting harder and harder to avoid. Core pillars of the establishment, like the meritocracy, are being challenged by an increasing number of people.

And, perhaps not unrelated, gender diversity in open source seems to be on the up. In 2006, the FLOSSPOLS survey concluded that 1.5% of open source participants were women. Many people still quote this number when talking about average figures. But a look at recent surveys seem to indicate that this figure may be as high as 11% now.

Participation of women in open source over time:

A chart graphing the percentage of women participants in open source from various surveys between 2002 and the present. Data points graphed as follows: From the FLOSS Project 2002, 1.1% women participants to 1.5% as measured by the FLOSSPOLS survey 2006. From the Ubuntu Community Census 2006, 2.4% women participants to 4.4% of Ubuntu Members in 2009. The Perl Survey 2007 is marked at 3%. The Drupal Community is designated from 7% in 2008 to 17% in 2014. GUADEC (GNOME's) annual conference attendees are marked at 5% women in 2009, 17% in 2012 and 18% in 2013. Google Summer of Code participation is marked at 7.1% women in 2011, 8.3% in 2012, 9.5% in 2013, and 9.8% in 2014. From the FLOSS Survey 2013, women are noted as 11% of FLOSS contributors. And finally for GNOME Foundation members, 10.3% were women as of January 2014.

Data taken from the Geek Feminism Wiki.

As Reagle writes, “women’s interest in a historically geeky topic can seemingly devalue it. Busse wrote of a character in the comic Foxtrot who lamented that ‘Orlando Bloom has ruined everything’ because of his sister’s new interest in the Lord of the Rings[...]” What’s actually going on here is that women’s participation de-masculinizes the activity. It is “ruined” because it is no longer possible to perform maleness through it. Similarly, as women and LGBTQ folk enter open source, the identity loses some of its heteronormative cisgendered maleness.

As communities start to bridge the gaps in culture, race, and class, communication styles and social norms are being challenged. This is especially true of projects that attract contributions from countries like China and India where the Power Distance Index (and other cultural dimensions) can differ considerably from countries in the Anglosphere.

And it doesn’t stop there. It’s becoming increasingly common for people to seek funding, grants, and other forms of financial support for doing open source work. Either through crowdfunding tools, through institutions, or through the projects themselves. While this allows projects to enjoy more diverse contribution, it also threatens to expose unexamined privileges.

Add to that the increasing amount of pressure to start recognising the legitimacy of non-technical contributions, and it’s clear that what it means to be an open source contributor today is undergoing redefinition. For many people on the inside, this is a fight that is going to feel intensely personal for them. Identity is a source of self-esteem and pride. Having the legitimacy of that questioned is going to be an unwelcome experience.

This allows us to understand the boundary policing we see in a new light. This isn’t about whether gendered pronouns are acceptable, or whether racist terminology should be taken out of the documentation. This is about who these people believe open source is for, and by extension, their own self-identity. Indeed, when people are challenged to explain themselves, you get nonsense and cognitive dissonance. People are being openly hypocritical, with no apparent awareness. The closer they get to confronting the truth, the more likely they are to break down in anger and confusion.

These sorts of situations are not new. Liminal periods are often messy and disorienting, but bring with them the potential for radical change. Where the dominant class supplies hierarchy-enhancing legitimizing myths to normalise structural inequalities, the onus is on us to supply hierarchy-attenuating ideologies, like diversity and intersectionalism, that counter hegemonic modes of thinking.

This will be hard, laborious work. But we need to keep forcing these conversations. Now is a time for iconoclasm. The opportunity is ours to redefine the identity and the culture of open source.