Original post available on the Localization Lab blog.

Since the military coup in Myanmar on February 1st, Localization Lab has shifted its priorities to focus on coordinating urgent localization efforts to support access to free and open internet and safe communication in Burmese and other local languages. This has meant working with contributors to localize tools requested by affected communities and local partners, which have experienced increased downloads and active users. These include tools like TunnelBear, a Virtual Private Network (VPN) that can help people access blocked content; Signal, a messenger that prioritizes privacy and security; and Briar, a messaging and publishing app designed to work without an internet connection; among others.

We are very grateful for the rapid localization support we have received from several organizations and projects, allowing us to coordinate dedicated contributors to remove language barriers, making these tools accessible to millions more who could benefit from them.

We also wish these urgent language needs had been addressed before such a crisis hit.

A preparatory approach — one that prioritizes localization, accessibility, distribution and (if necessary) training prior to a crisis, instead of in response to one — needs to become the preeminent model for internet freedom tools.

Shutdowns have become increasingly common, and circumvention tools commonly experience a spike in local interest as connectivity plummets. At Localization Lab, it is common to see a flood of new requests to make relevant apps available in local languages when a crisis breaks out, and for new contributors to join projects and work on those languages.

Both are happening today in Myanmar. Within two days of the coup, Bridgefy, an app widely used by activists in Hong Kong during the protests in 2019 and 2020 to communicate via Bluetooth when internet connectivity was unavailable, was downloaded there more than a million times. After Facebook was blocked on February 4th, 2021, VPN usage in the country increased to more than seven times what it had been only days earlier. The Tor Browser saw a jump in interest and downloads, too, though usage dropped after various forums reported that military personnel were looking for people with the app installed on their devices.

Such was the case, too, in Belarus in August 2020, when internet access shrank to 20% of typical levels amid protests following a controversial presidential election. Psiphon, another proxy tool that we are localizing into Burmese, saw its adoption in Belarus escalate suddenly and dramatically, from around 10,000 users to more than a million in a single day.

One reason it was able to accommodate such a sudden spike in interest is that it is the only VPN provider available in Belarusian. As one Psiphon co-founder pointed out, “No one else is doing that.”

Localized tools and guides need to be as diverse as the needs of the communities they are built to serve. Movements are dynamic and their needs and tactics shift according to which information is blocked and censored, and which circumvention tools are working. The more resources are available in local languages, the more tools they have available to choose from that fit their needs best–leading to stronger and more resilient movements and communities.

These preparatory models need not be imagined, only prioritized, expanded, and scaled. Internet shutdowns, to some degree, are predictable. For example, last August the Zaina Foundation hosted a localization sprint in Tanzania ahead of the upcoming elections there. Elections have become common trigger points for internet shutdowns, especially (but certainly not exclusively) in many parts of Africa and southeast Asia.

Anticipating a possible shutdown, Zaina Foundation and Localization Lab collaborated on a localization sprint where participants translated circumvention software, including Psiphon, so it would be available in Swahili, building upon the previous year’s work of localizing digital security guides. Just months later in another sprint organized with the Arusha Women School of Internet Governance (AruWSIG), contributors localized additional resources into Swahili, including a censorship circumvention guide.

As expected, internet connectivity was severely disrupted before and during the elections in Tanzania, and while the Tor network appeared to be blocked, Psiphon continued to work.

This need extends to written guides as well, which can be quite useful for preparation and self-training. And while there is a clear need for more localization (we are currently working with the Electronic Frontier Foundation to localize their widely-cited Surveillance Self Defense guides into Burmese), there is also a need for clear, concise, and specific guidance on how to prepare for an internet shutdown. An accessible, localizable, and widely available resource dedicated specifically to the common challenges and potential mitigations around internet shutdowns would serve many and support the preparatory model championed here.

But a shift like this will require more than a new guide, and will require all of our cooperation. It will take funders providing more resources to toolmakers and local organizations for long-term resiliency programs; developers writing their tools with multiple audiences and contexts in mind from the beginning; and training and support organizations building shared contextualized expertise with local communities around circumvention and resistance so plans to access information are established before the crisis hits.

Most importantly, we can only expect a preparatory model to be effective if local partners and users are centered at every step. They know what tools and resources people need, how they use them, and the gaps that need to be addressed. Local partners are best positioned to develop the content that will work for their communities, and that can then be adapted to support regions experiencing similar challenges. Involving those communities should not be a box to be checked, but a central guiding tenet that shapes projects and partnerships.

We recognize, too, that there will always be a need for rapid response during crises. But by adopting a more preparatory approach, we shrink the number of challenges that require a rapid response when a crisis hits, so that funding and labor can be applied to different needs, ones that are now more urgent.

A time of crisis is not the time to build resilience, but to wield it. We stand in solidarity with those who seek to secure their communications and access the open internet in Myanmar, and we invite our colleagues in the internet freedom space to join us today in undertaking the necessary work to support and prepare the communities who will face the next internet shutdown, and the next one, and the next one.

Original post available on the Localization Lab blog.

In this moment, when we are reimagining our systems during calls to end racial structural violence and during a devastating pandemic, it’s important that we start with reforming language, replacing terms that exclude and reinforce oppression in our systems. Translating technical terms from English to other languages includes running the risk of unintentionally and unconsciously imposing Western values and constructs of gender, race, and individualism through technology. Terms like “whitelist/blacklist” and the “master/slave” programming command carry the weight of systemic issues deeply rooted in society. When these terms are localized into other communities, the values associated with them could be transferred into other cultures with different ideas of race and gender¹. Though the intent may not be to reinforce existing stereotypes and inequality, the effect is proven in the lack of inclusive language for a spectrum of users.

Translating terms with sexual innuendos perpetuate ideas of sex and gender, often in a binary bias. Sexist terms “preserve the ways in which casual use of English serves to uphold maleness as the norm and femaleness as the exception.”² Male domination, especially in decision making roles within the tech industry is evident by the exclusionary language used. By reinforcing sexist terminology, “gendered language performs some of the work of exclusion by reinforcing the message that female programmers are exceptions.”³ In order to systematically change tech culture and reach true inclusivity, language must also shift and reflect equitable values.

As we work in partnerships with marginalized groups, it is crucial that we are transparent about the problematic and at times violent histories of these terms; that we do not unconsciously pass on and perpetuate harmful ideology. For example, during our Localization Sprint in Thailand, participants collaboratively pushed back on existing constructs, intentionally careful in their methodology of language adoption. When technical terms are localized and translated, ideas are consequently transferred over, preserving or even promoting ideals of race, class, and gender onto other cultures without a clear understanding of possible harm. Language has the power to manipulate and change rhetoric, and it is our responsibility to acknowledge these problematic terms, change the narrative, and combat digital colonialism.

The tech industry is notoriously known as a heteronormative, white, male-dominated space; a bubble of intergenerational wealth, patriarchy, and privilege. The language and technical terms used in the tech industry are a reflection of those who are represented, hold the power, and control discourse. This exclusive fortress of power can create a hostile and uninviting environment for those who do not subscribe to, or identify with the norm; denying entry for diversity and inclusion based on colonial lines.

This article examines the technical terms used in cybersecurity, programming, and platforms with a gendered and racial lens, critiquing outdated and biased words used in the industry.

Racial & Ethnic Bias in Tech Terminology

Master-Slave command (Programming)

The term “Master-Slave” is a common computer programming phrase that refers to the idea that “components have total control over other components, or are controlled by a component, respectively,”⁴ a one-way direction of control and power. By continuing to directly connect its association with slavery, this term normalizes inhumane practices, perpetuating the 465 years of institutionalized systemic oppression of Black Americans.⁵ In addition, modern slavery in the form of the prison industrial complex, domestic servitude, and child soldiers still exist today.

By reducing “master” and “slave” to computer parts, the term desensitizes and softens the painful, historical dehumanization of enslaved peoples. This terminology, which has already been removed by Django, Twitter, Python, and Github contributes to the “power dynamic that’s the blood in the heart of racism.”⁶ The term “reflects the bigotry of those that dominated the field when much of the conventions for technology were formed.”⁷ The Master-Slave command is easily replaceable with more appropriate terminology like “database credentials” and “main” instead of “master”.

Blacklist, Whitelist, Black hat hacker, White hat hacker

In the most simplest terms, the racist ideology of “white is good, black is bad”⁸ is unfortunately echoed into the racist biases of society, which has then translated into the exclusionary binary of technical language. The dichotomy of the colors — white and black, which respectively equates to good and evil, desirable and undesirable, not only reflect racist and colorist sentiments, but mirrors the problematic racist stereotypes of people in society.

The terms “whitelist” and “blacklist” are used in digital security to differentiate approval and denial. Whitelisting has a “trust-centric approach, allowing access for approved entities”⁹. While “blacklisting” is “threat-centric and involves blocking access to suspicious or malicious entities.¹⁰ The implications of good and bad associated with these terms have a historically racist past, accompanied by the hierarchical direct relationship between race, class, and power. Additionally, the term “white/black hat hackers” have similar connotations of good and evil, oftentimes compared to ethical and criminal hacking, respectively. Subconsciously, the implications of the colors white/black bleed into the problematic biases for people of color. Mallory Knodel, the former Chief Technology Officer at Article 19 addresses that “this trope has significant impact on how people are seen and treated. As we’ve seen with metaphors, its use is pervasive and, though not necessarily conscious, perceptions do get promulgated through culture and repetition.”¹¹ There is no technical need to use “white/black” in describing these mechanisms. In addition to reinforcing structural racism, the choice to continue this type of language reinforces overt prejudice and unconscious biases against a community that is already underrepresented in the field of technology.

Grandfathering

Though not exclusively used in the tech industry, phrases like “that product was grandfathered in” are directly related to the grandfather clause used to discriminate against African Americans in the nineteenth century. This phrase is usually used to describe “individuals or companies who get to keep operating under an existing set of expectations when new rules are put in place.”¹² Racialized in its origins, the grandfather clause was created to discriminate and institutionally disenfranchise black voters and “their descendants from voting, while allowing poor and illiterate whites to vote.”¹³

Though the usage of grandfathering is now part of mainstream vernacular, its origins are rooted in oppressive policies enacted to institutionally disenfranchise African Americans. The term’s historical connections to slavery and casual reference to White supremacy does not reflect the ideals of an inclusive, safe, and equitable environment.

Mechanical Turk

Amazon Mechanical Turk (MTurk) is a crowdsourcing marketplace that allows for businesses to outsource their jobs to a distributed virtual workforce.¹⁵ The original Mechanical Turk is a fake chess-playing device, created in the 18th Century by a Hungarian engineer.¹⁶ The device was an automated machine that acted as a mechanical illusion, while a real person hidden inside the device played chess.

The name and use of the Turkish culture embodied in these two automated mechanisms is rooted in the idea of oriental, mysticism from the East; a stereotype pinned on those originating from Asia. Ayhan Aytes, a researcher on cultural history of Artificial Intelligence and Cognitive Labor, connects the way the West views the East as “performers of technological and cultural alterity, satiating the anxieties caused by the unfamiliar notion of mechanized cognition, by projecting them onto all-familiar ethnic and religious differences.”¹⁷ The parallels of Amazon’s platform and the Mechanical Turk encompasses the idea of a mystic, hidden labor force; fulfilled by others, unbeknownst to the mainstream audience. Heavily criticized for its low-pay and lack of labor regulations, Amazon’s platform exploits those who are in desperate need of employment, yet have adept technical skills.

Through Amazon’s platform, the word “turk” has additional dimensions. Employees are called “turkers”, and the verb form is known as “turking”. The term “turk” has a derogatory history, a blanketed label used to describe anyone with ethnic origins from the Middle East. Racialized in nature, Amazon’s use of the name “Mechanical Turk” highlights how the West is “overlooking examples from Islamicate cultures as an extension of colonialist desires to exclude non-western cultures from the Western-centric histories of science and technology.”¹⁸ Not only is it problematic as a platform, but the name reinforces the idea that “Orientalist undercurrents were exploited by Enlightenment discourse in order to configure the docile subject on the image of the Turk.”¹⁹ Usually performed by marginalized communities, the hidden labor of automated work is portrayed in both of these machines.

Sexist, Classist, and Cishet-normative Terminology

Fingering, Penetration and Pentest (Digital Security)

Digital security has several problematic terms rooted in sexism, including but not limited to fingering, penetration, and pentesting.

The term “fingering” has had various meanings in tech throughout the years. The original and most common use comes from a computer science term; an old Unix command called “finger.”²⁰ Created in 1971, it is used to find information on a particular user in a computer or a network.²¹ Though the origins of the term was not intended to be sexual, the meaning of the word has evolved to become so. “Fingering” as a computer command may be uncomfortable for a group of people, specifically women, who represent a minority in the tech industry.

The word penetration has many different meanings and has taken a definition of its own in the tech world. Phrases such as “internet penetration” and “penetration of products” insinuates a forceful entry and hints at a sexual innuendo. Though this term has entered mainstream vocabulary, it is important to address how it could conjure up past traumas or unpleasant imagery; especially for survivors of sexual assault.²²

In addition, penetration testing or pen test is an “attempt to evaluate the security of an IT infrastructure by safely trying to exploit vulnerabilities.”²³ Also known as ethical hacking, pen testing is a common method used in assessing security. While acknowledging this word can be used in nonsexual contexts, the mere use of penetration as technical terminology implies a sexual act and a masculine endeavor, a reflection of dominance in a majority male industry.

Male/Female Connectors

Female and male connectors are terms used in electrical and mechanical trades. A male connector is usually a “plug and has a solid pin for a center conductor.”²⁴ A female connector is “a jack and has a center conductor with a hole in it to accept the male pin.”²⁵ This mechanism is a direct analogy to sexual intercourse and the assumed binary nature of gender. By strictly promoting this cis normativity, we indirectly push this notion that gender and sex are binary, excluding and delegitmalizing those who identify as gender non-conforming or trans. In lieu of assigning a gendered relationship, using +/- could adequately represent this binary relationship.

John the Ripper

John the Ripper is a free, open source, “fast password cracker” used to detect weak Unix passwords.²⁶ The name is a play on the infamous “Jack the Ripper”, a serial killer who preyed on low-income women and mutilated the bodies of five sex workers in 19th century London.²⁷ The commemoration of a sadistic killer whose “crimes seemed to portray an abhorrence for the entire female gender”²⁸ feeds into the solicitous headlines that disregard the real fear of sexual assault. If diversity in the tech community increased, tools like John the Ripper might instead have titles that describe their actual function, without conjuring violent imagery.

Evil Maid Attack

An evil maid attack is an “attack in which bad actors gain physical access to unattended computing devices for malicious activities.”²⁹ This term originated from the idea that devices are left unattended in hotel rooms, where a maid has unsupervised access.³⁰ Assigning this type of attack to a typically female, service profession can be seen as playing into the demonization and distrust of those in the lower income bracket and could perpetuate harmful and false stereotypes of stealing and dishonesty. From a translation perspective, the analogy supporting “evil maid attack” is also complex and the relationship is difficult to explain, making it a particular challenge to translate in addition to the term being problematic in nature.

SuCKIT

A rootkit is a “stealthy type of malicious software designed to hide the existence of certain processes or programs from normal methods of detection.”³¹ A type of rootkit created to detect this compromise is named “sucKIT.”³² The name is a play on the vulgar expression “suck it,” which is oftentimes used as an insult, insinuating phallic action. Frequently used colloquially, it is a condescending phrase often indicating gendered power and dominance.

Man-in-the-Middle Attack (MitM)

A man-in-the-middle attack is when “an attacker intercepts communications between two parties either to secretly eavesdrop or modify traffic traveling between the two.”³³ Known as one of the oldest cyber attacks, this term does not translate well into other cultures nor does it accurately represent its description. Instead, the term “on-path attackers” is more accurate, as “attackers can then collect information as well as impersonate either of the two agents.”³⁴ Assigning this position of power to a man again highlights the male dominance in this industry. Because this type of attack was assigned a binary gender, it further reinforces that cybersecurity, hacking, and tech are for men, excluding other identities and representation.

Moving Towards Inclusive Terminology

Inequitable terms can easily be replaced to create change without compromising the value of the technical work itself. Instead, a plethora of alternative, appropriate, and inclusive terms should be used and mainstreamed, creating a new norm.

Language and culture are intrinsically connected, reflecting societal norms and its values. ^35In a Western-centric industry, it is important to address and reconfigure the racist and sexist technical terms, even if unintentionally so, that are still used today. Knodel notes, “subtle configurations of sexist, racist, or ethnocentric language use[d] in technical documents can derail or interfere with readers’ ability and desire to comprehend and follow important information.”³⁶ When language serves to reinforce a dominant group’s values and power, marginalized groups are forced to work within these imbalanced power dynamics, stifling abilities of consent and inclusion.

Merely hiring people of color is not enough, and changing the problematic language used within the industry is one step towards shifting the culture. Although language sounds innocent and a relatively minute detail, its influence becomes exponentially elevated when translated and localized to adapt to other cultures. For example, languages that use a masculine/feminine form risk mainstreaming the male form and erasing female presence.³⁷ Practitioners must lead by refusing to use terminology that reinforces oppressive, exclusionary structures. These terms, which are embedded in the technical infrastructure of society, must represent an inclusive environment, as we work to reimagine and reconstruct societal foundations.

These expressions used in technology are a reflection of the historical biases within society. As we aim to create a diverse and inclusive industry, terminology matters. As we work to dismantle white supremacy and cultural bias, we must explore how Western binary definitions of race, gender, sexuality, class, and power affect those who rely on technologies developed in the West. When technology is localized to other cultures, especially adopted for minority groups, it is important that individuals feel safe and comfortable with the tools and language they are using. We need to replace terms that support casual racism, sexism, and cisheteronormativity with inclusive language, and ensure that dangerous ideologies are not perpetuated in other cultures, preventing harm in communities that rely on this technology. The tech industry has the power to change and shift its narrative, and it starts with inclusive language.

At Localization Lab, we hold localization sprints to discuss tech terminology and how to responsibly localize to different communities. This practice could be adopted for an English language sprint, bringing together diverse contributors in the tech space to tackle the complex and nuanced challenge of creating inclusive language and building a guide towards best practices.

Citations

¹ Taha, L., & McConnell, E. (2020, April 14). On Absenting Women from Arabic Public Discourse. Retrieved July 29, 2020, from https://www.localizationlab.org/blog
² Chevalier, T. (2014, February 24). Gendered Language: Feature or Bug in Software Documentation? Retrieved from https://modelviewculture.com/pieces/gendered-language-feature-or-bug-in-software-documentation
³ Ibid
⁴ Oberhaus, D. (2018, September 13). ‘Master/Slave’ Terminology Was Removed from Python Programming Language. Retrieved from https://www.vice.com/en_us/article/8x7akv/masterslave-terminology-was-removed-from-python-programming-language
⁵ Worland, J. (2020, June 11). America’s Long Overdue Awakening on Systemic Racism. Retrieved from https://time.com/5851855/systemic-racism-america/ , Solomon, D., Hanks, A., & Weller, C. E. (2018, February 21). Systematic Inequality. Retrieved from https://www.americanprogress.org/issues/race/reports/2018/02/21/447051/systematic-inequality/
⁶ S. (2018, September 9). Even in Tech, Words Matter. Retrieved from https://deninet.com/blog/2018/09/09/even-tech-words-matter
⁷ Ibid
⁸ The New York Times, T. (2017, April 02). Readers Respond: Which Racial Terms Make You Cringe? Retrieved from https://www.nytimes.com/2017/04/02/us/racial-terms-that-make-you-cringe.html
⁹ Blacklisting vs. Whitelisting. (2019, August 12). Retrieved from https://consoltech.com/blog/blacklisting-vs-whitelisting/
¹⁰ Ibid
¹¹ Knodel, M. (2019, September 12). Terminology, Power, and Offensive Language [Scholarly project]. In Network Working Group. Retrieved 2020, from [https://tools.ietf.org/html/draft-knodel-terminology-01][0]
¹² Greenblatt, A. (2013, October 22). The Racial History Of The ‘Grandfather Clause’. Retrieved from https://www.npr.org/sections/codeswitch/2013/10/21/239081586/the-racial-history-of-the-grandfather-clause
¹³ Riley, N. (2019, May 31). Words Matter: Why We Should Put an End to “Grandfathering”. Retrieved from https://medium.com/@nriley/words-matter-why-we-should-put-an-end-to-grandfathering-8b19efe08b6a
¹⁴ Mechanical Turk. Digital Image. Untold History of AI: When Charles Babbage Played Chess With the Original Mechanical Turk. 18, March 2019. https://spectrum.ieee.org/tech-talk/tech-history/dawn-of-electronics/untold-history-of-ai-charles-babbage-and-the-turk
¹⁵ Amazon Mechanical Turk. (n.d.). Retrieved July 02, 2020, from https://www.mturk.com/
¹⁶ What is Mechanical Turk? (2016, July 11). Retrieved from https://www.pewresearch.org/internet/2016/07/11/what-is-mechanical-turk/
¹⁷ Aytes, A. (n.d.). Media Archaeology, Cultures of Automata, and Mechanized Cognition. Lecture presented at Cornell University. Retrieved 202, from http://complit.cornell.edu/sites/complit/files/Aytes_Talk.pdf
¹⁸ Ibid
¹⁹ Ibid
²⁰ Finger protocol. (2020, May 16). Retrieved July 24, 2020, from https://en.wikipedia.org/wiki/Finger_protocol
²¹ Colbath, Sean. “Origins of the Finger Command .” Google Groups, Google, 1990, groups.google.com/forum/#!msg/alt.folklore.computers/IdFAN6HPw3k/Ci5BfN8i26AJ.
²² New, J. (2015, May 27). Asking Too Much, or Not Enough? Retrieved August 17, 2020, from https://www.insidehighered.com/news/2015/05/27/language-sexual-assault-surveys-criticized-students-triggering
²³ Penetration Testing. (n.d.). Retrieved July 02, 2020, from https://www.coresecurity.com/penetration-testing
²⁴ What is the difference between male and female connectors? (n.d.). Retrieved July 02, 2020, from https://www.l-com.com/frequently-asked-questions/what-is-the-difference-between-male-and-female-connectors
²⁵ Ibid
²⁶ John the Ripper Password Cracker. (n.d.). Retrieved July 02, 2020, from https://www.openwall.com/john/
²⁷ Dickson, E. (2019, March 18). Jack the Ripper May Finally Have Been Identified, Says New Study. Retrieved from https://www.rollingstone.com/culture/culture-news/jack-the-ripper-identity-study-aaron-kominski-809808/
²⁸ History.com Editors. (2010, November 08). Jack the Ripper. Retrieved July 02, 2020, from https://www.history.com/topics/british-history/jack-the-ripper
²⁹ Yedakula, K. (2019, April 14). What is an Evil Maid attack and how is it different from Evil Twin attack?: Cyware Hacker News. Retrieved from https://cyware.com/news/what-is-an-evil-maid-attack-and-how-is-it-different-from-evil-twin-attack-8a73a96
³⁰ Evil maid attack. (2020, June 07). Retrieved from https://en.wikipedia.org/wiki/Evil_maid_attack
³¹ Checking for Known Rootkits. (n.d.). Retrieved July 02, 2020, from https://security.web.cern.ch/recommendations/en/rootkits.shtml
³² Ibid
³³ Swinhoe, D. (2019, February 13). What is a man-in-the-middle attack? How MitM attacks work and how to prevent them. Retrieved from https://www.csoonline.com/article/3340117/what-is-a-man-in-the-middle-attack-how-mitm-attacks-work-and-how-to-prevent-them.html
³⁴ “What Is an on-Path Attacker? .” Cloudflare, www.cloudflare.com/learning/security/threats/on-path-attack/.
³⁵ Birner, B. (n.d.). Does the language I speak influence the way I think? Retrieved July 02, 2020, from https://www.linguisticsociety.org/content/does-language-i-speak-influence-way-i-think
³⁶ Knodel, M. (2019, September 12). Terminology, Power, and Offensive Language [Scholarly project]. In Network Working Group. Retrieved 2020, from [https://tools.ietf.org/html/draft-knodel-terminology-01][0]
³⁷ McConnell, E., & Taha, L. (2020, April 14). On Absenting Women from Arabic Public Discourse. Retrieved July 29, 2020, from https://www.localizationlab.org/blog

By Layla Taha

Translated from Arabic by Erin McConnell & Layla Taha

As an Arabic-English volunteer translator with Localization Lab, I often find myself addressing a “general” public; asking them to “download” this tool, or “open” that application. In doing so, the feminist in me and the woman whose mother tongue is Arabic are repeatedly interrogated.

In English, most technical communications, whether in an app’s interface or an organizational newsletter, address a given “you” in an attempt to sound personal. This creates a problem for English to Arabic translators because unlike English, Arabic has gender (masculine and feminine) incorporated into most aspects of its grammar — from nouns to adjectives to verbs. So, when using the imperative, such as “Open” or “Download” or “Install”, one must specify the recipient’s gender. Since it’s presumed you’re addressing a “person”, a masculine word in Arabic, most communications address a singular male.

To mitigate this, I started using the default masculine plural form to address “persons” of a mixed gender group, which is at least grammatically correct. But when I asked Rula Asad, the Executive Directorate of the Stichting “Syrian” Female Journalists Network, whether that’s a suitable solution, she casually responded: “Isn’t that a masculine plural?”

Rula was alluding here to another grammatical rule in the Arabic language which seemingly prioritizes male recipients of a message over their female counterparts. Much like French, there are masculine and feminine plural forms in Arabic. For example, in French you have the feminine singular madame (sayyidah in Arabic) and the feminine plural mesdames (or sayyidat in Arabic). The issue in Arabic however, is that when addressing a group of women and men, the masculine plural form, (al sadah) would suffice, grammatically-speaking. As for including the two plural forms, (similar to the French mesdames et messiuers or the Arabic al sayyidat wal sadah), that’s optional and left to the speaker’s discretion (or personal mannerisms).

In a long, virtual (and quite enjoyable) conversation with Rula, she said that the repeated use of the masculine form in public discourse in Arabic leaves women feeling excluded from the conversation. “There is an assumption that the language that is spoken and used is inclusive of women, despite the fact that it does not include women directly… In your mind’s eye, you still imagine a message intended for a group of men, rather than a group of women or a group of people with diverse gender identities.”

Rula believes that language is a tool and a medium that influences public opinion, shapes the discourse towards a given group or issue, and ultimately reflects itself in people’s daily lives, behaviors, and reactions.

When I asked her how the use of the masculine plural affects her personal reception of a given text, she replied: “I don’t feel involved in the subject and so I don’t get involved. I’m not exaggerating. It affects the extent of engagement or social responsibility that I feel. That’s why I am reticent to consume cultural production that doesn’t include me, or directly address me in its tools.”

It’s really not an exaggeration. In her book Invisible Women: Data Bias in a World Designed for Men, Caroline Criado-Pérez provides numerous examples that the data we use to inform economic, health, or other public policies does not take women into account. This exclusion often causes women to pay an extremely high cost in terms of time, money, and sometimes even with their lives.

Caroline explains that heart attacks, for example, are often misdiagnosed in women (one out of two times in Britain according to The Guardian) because the symptoms of the illness that appear in women are completely different than those of men — whose data is primarily used to inform medical research. In men, chest pain is one of the most common symptoms of a heart attack, whereby this is only the case for 1 in 8 women who suffer from a heart attack. For women, stroke symptoms are more likely to take the form of fatigue or digestion issues.

In her book Caroline argues that this phenomenon is also pervasive in the technology sector, from hardware architecture (eg. mobile phone size) to algorithms. Google’s speech recognition software for example is 70% more likely to understand men. Most “fitness tracking” applications underestimate the number of steps taken while doing housework at an error rate of up to 74%.

In an interview with Wired, Caroline explained that this is “both a cause and a consequence of a kind of unthinking that envisions humans as exclusively male.”

So we return to how we envision our humanity as we narrate, describe, and express it in our public discourse in Arabic, what could the consequences of continuing to “unthink” of women in nearly everything we write be?

“In its current form, Arabic is unfair, not only to women, but also to non-gender conforming individuals…”

“In its current form, Arabic is unfair, not only to women, but also to non-gender conforming individuals. If we take baby steps, our public discourse doesn’t even begin to reflect what is acceptable to society… We also have to take into account unspoken signifiers, or what is implied when unsaid. The use of language must be inclusive and fair. It must respect peoples’ choices and be free from judgment or stereotyping of specific social groups. This is the sort of public discourse we aspire to create,” says Rula.

Towards Finding Local Solutions

On the sidelines of the Internet Freedom Festival of 2019, Erin McConnell of The Localization Lab organized a session on “Exploring Approaches to Gender Inclusive Language in Arabic”. Farah Barqawi, co-founder of the Wiki Gender project, a participatory platform that produces open feminist knowledge on gender issues and women in Arabic, moderated the session.

The session brought together a number of people interested in this localization dilemma including Ahmed Gharbia, the former artistic director of the Arab Digital Expression Foundation (ADEF).

When I asked Ahmad why he participated in the session, he said that he is specifically interested in this issue and in contemporary Arabic linguistic phenomena in general. “This issue has many dimensions: linguistic, morphological, as well as semantic and rhetorical. It’s important for me professionally to take part in creating an environment where women and men can cooperate in different areas of their lives — one that is more reflective of womens’ visions of the world.”

In the session, Farah presented a background of local attempts to deal with the dilemma of inclusion in Arabic.

A Summary of Current Approaches to Gender Inclusion:

The Backslash
Using the feminine suffix after a backslash (‘/’) may be one of the first attempts to address gender inclusion by some Arab media organizations. Farah advises against this approach as much as possible because the plethora of backslashes “break the flow of the text visually, linguistically and otherwise.”

The Masculine Plural
This approach, simply using the masculine plural instead of the masculine singular, is usually accompanied by using nouns as opposed to verbs when possible, in an attempt to avoid verb conjugation altogether.

In response to the use of the masculine plural, Farah says: “We settle for the plural because it addresses a group, however it is still the masculine plural. If there is a group of 9 women and a lone male suddenly joins, the feminine plural no longer applies. Even in our cultural heritage the masculine is given high status. It is not blatant discrimination, but that does not mean it is not descriminatory. In these situations, let’s just use both plurals (masculine and feminine).”

As mentioned at the beginning of this post, I myself often use the masculine plural in my writing. I see this as an approach that addresses women and men (Admittedly I wrote “men and women” and then corrected myself!). I take this approach because it does not draw much attention to itself and away from the content (like replitive conjugations or backslashes do). It is also commonly used in colloquial language. When speaking to a group of women in colloquial Arabic you say: “Shall we (m.pl.) walk (m.pl.)? Do you want to drink (m. pl.) something?”, as opposed to using the feminine plural which would require more thoughtful deliberation.

To this Ahmad adds that colloquially, “Arabs use masculine and feminine forms without paying attention to the actual significance of gender, and they use other determinants when there is an actual desire to determine gender. These determinants are considered excessive or erroneous if we strictly apply the rules of morphology and grammar. For example, ؛Today I saw two women engineers (f. pl.).؛ This phenomenon exists in various forms in the Arab world. For example, most Egyptians have united around use of the masculine plural for demonstrative pronouns (this, that, these, those) and most plurals. However, Levant dialects have standardized use of the feminine plural for demonstrative pronouns even when the aforementioned contain masculine nouns. In the dialects of the Maghreb we find variety in the use of some verbal tenses. This is all to say that maybe the solution is to ignore masculine and feminine markers all together and not count any customization of gender… In my opinion, the most important thing is the discourse, the ideas and the concepts, and how they are presented, treated and analyzed.”

Using both Genders
Using both the masculine and feminine plurals as it appears in the Quran:

“المُسْلِمِينَ وَالْمُسْلِمَاتِ وَالْمُؤْمِنِينَ وَالْمُؤْمِنَاتِ”

“Muslims (m. pl.) and Muslims (f. pl.), believers (m. pl.) and believers (f. pl.).” As for Wiki Gender, its editorial policy prefers giving precedence to the feminine plural over the masculine plural, or only using the feminine plural.

Alternation of Gender Within the Text
Farah, the co-founder of Wiki Gender and the moderator of the IFF session, stated that this problem of discourse originates from imagining that we are addressing a “person”, a masculine word. “Person” is a word that claims neutrality. It refers to an unknown person (sex, gender, etc.), which can be pluralized, such as “people” (m.pl.) and “persons” (m.pl.), but may not be feminized in the plural form. Hence the idea of diversification (which requires imagination) or switching gender signifiers throughout the same text and in the plural forms of speech (eg: journalists (f.pl.), researchers (m.pl.), students (f.pl.) and interested parties (m.pl.))

An Experimental Portmanteau
Farah says that some time ago, the members of Wiki Gender began experimenting with solutions to the problem of the backslash, which often separates the masculine and feminine suffixes when addressing a group of people (much like “s/he” in English). They played with removing the backslash and combining the two suffixes into one word. The result was a completely new word that signified both genders, rolled off the tongue, and was still declensional.

Ahmad says he discovered similar experimentation in the Maghreb, but he doesn’t know how it originated or who is behind it. .

For her part, Rula believes that this method is truly creative and adds to the language from within. She advocates for linguistic experimentation and sees that finding a solution will require patience: “For so long, people tolerated gender stereotyping and exclusion in most public discourse and knowledge production. So why can’t we be patient with these texts that are in an experimental stage? So long as there are no solutions, people can eventually get used to it.” She adds, “Alternative media outlets, who are advocating for gender equality or female empowerment, should understand the social and emancipatory aspects of such experiments, rather than simply treat media discourse as something to be sold and consumed.”

On Not Falling into the Translation Trap

Following the session, I followed up with Farah on what motivates her. In a Jitsi call she said, “Arabic has a rich vocabulary and a legacy of a popular imagination that was sexually abundant and had no problem in naming sexual acts or acknowledging diversity of gender and sexual orientations. But with the establishment of organized religion, states, and wars, and the spread of capitalism that engulfed the region as well as the rest of the world, entered a holy reverence for heterosexuality and stereotyped gender roles that consist of the productive family of children, the working father and the mother doing unpaid domestic work. So, to treat the subject of inclusion as a novelty, because it’s surfacing in a North American context is to fall into the trap of translation first. It’s also a trap of borrowing things that may not be essential in our contexts.”

She adds: “A common manifestation of this is the obsession with finding one word in Arabic, when Arabic is not a language of single words. How many words do we have in Arabic for “love” compared to in English? Why don’t we celebrate this diversity? I’m not saying we shouldn’t address the issue, I’m saying we must be careful to ask the questions that arise from these dilemmas, in our organizations, in writing, in discrimination against female writers, etc. Who are the editors and the publishers and how do they produce what is written or published in Arabic?”

This is the second in a two-part post about integrating localized training, user feedback, and developing outreach strategy into localization sprints — events that bring individuals together to localize (adapt and translate) technologies and resources for their communities. Read Part I of this post.

By Erin McConnell

Diving into Local Needs: User Stories and Persona Building

The Thai Localization Sprint challenged sprint attendees to imagine different Thai user profiles and their digital security needs and motivations through user story creation and persona building (for example, imagining a middle-aged farmer or a teenage student and their needs.) This helped the participants better understand Tor, its diverse use cases and the large breadth of individuals who could benefit from using it. Particularly with tools like Tor that come with a lot of misconceptions, it is important for users to imagine themselves and others in their networks using the application. It’s imperative that individuals be able to create narratives that resonate with their communities so that they better understand the tools and adopt them when appropriate.

Particularly with tools like Tor that come with a lot of misconceptions, it is important for users to imagine themselves and others in their networks using the application.

The persona building and user story activities — done in small groups — resulted in very different outputs, drawing from experiences in the human rights sphere and outside of it. The personas and stories highlighted how Tor and other digital security tools are not limited to those in activist or human rights work. Digital security issues are relevant to all and tools like Tor can address the needs of a social media influencer, a land rights activist, or someone who wants to privately used a shared computer or mobile device.

Imagining Thai User Personas

The Okthanks Persona Builder gamifies the creation of user personas to help organizations and developers explore the security concerns of individuals from different backgrounds and working in different contexts. The activity was used to introduce and explore Thai personas, using USABLE personas as a guide, so that participants could integrate them into their future work.

Following an adapted version of the Persona Builder game, each group determined the name, age, education, location and job title or role of their persona and then picked four more user persona components at random from a shuffled stack of options. Examples of available persona components included: the individual’s unique priorities, the devices and applications they use, their technical strengths and frustrations, information about their connectivity, political dynamics in their region etc.

“We work on digital security in urban areas, and we often forget the limitations and difficulties that those in rural areas are facing. This was a good exercise to look back on these issues.”

— Thai Localization Sprint Participant

The end result was two drastically different personas. The first user persona, Nu Nui, reflected the experience of participants who worked with land rights issues in Thailand. Nu Nui was a farmer in the Upper South of Thailand who was involved in community organizing around land rights. She had little technical literacy and faced physical threats from corporate developers trying to acquire land. Through the activity, the group identified several digital security threats and needs for Nu Nui and her community, including: secure communications, a panic button application in case of physical threat, secure information storage for legal documents, secure backup, and the ability to delete sensitive content quickly on a mobile device.

The second group imagined a very different persona than Nu Nui. They profiled Gibza, an social media influencer and owner of an online business. Gibza, whose priorities revolve around her physical appearance and social status, would only use Apple devices, which participants noted is a status symbol in Thailand, limiting her to applications for iOS and macOS. One of her key concerns was keeping banking and financial information secure from government and family members to avoid paying import taxes on resale items and to avoid family requesting financial aid. While Gibza had security concerns, any solutions would have to fit seamlessly into her carefully curated image.

Who is the Thai Tor Browser User?

While many individuals are familiar with the Tor Project, the project’s image remains fraught with myths and misconceptions. Particularly in certain regions, the browser is associated with illegal activity or dismissed as a tool only for those with “something to hide”. To challenge these perceptions and explore the myriad of uses for Tor the group was tasked with creating user stories based on an overview of Tor earlier in the day.

Participants were split into two groups and each received copies of the recently released Tor outreach fliers, which contain user stories focusing on key uses for Tor (anti-censorship, anonymity, privacy). Each group wrote a Tor user story in Thai for a local context which identified an individual, the goal they wanted to achieve and how Tor helped them reach that goal.

One user story centers on a woman who wants to purchase a sex toy, and uses the Tor Browser to navigate barriers stemming from policies and sociocultural norms.

User Story #1:
แอ๋ม เป็นสาวพนักงานธนาคารแห่งหนึ่ง แอ๋มสนใจอยากซื้อ Sex Toy แต่จะเซิร์สดูก็กลัวว่าเพื่อนร่วมงานจะรู้ว่าแอ๋มเป็นคนเปิดดู เพราะในออฟฟิศมีฝ่าน ITคอยตรวจสอบการใช้งานอินเทอร์เน็ต บวกกับที่บ้านใช้คอมพิวเตอร์ร่วมกันจึงมองหาเครื่องมือที่ปลอดภัย โดยไม่สามารถระบุ IP Address หรือมีประวัติการใช้งาน แอ๋มบังเอิญได้ดูคลิปของเจ๊ตุ๊ก ซึ่งเป็นนักกิจกรรมผู้หญิงบน Facebook แนะนำให้ใช้ Tor Browser จึงหันมาใช้ Tor Browser ตอนนี้แอ๋มรู้สึกสบายใจค้นหาความรู้ด้าน sex toy อย่างสบายใจและมีความสุข

English Translation:
“Am” is a bank worker who wants to purchase a sex toy for herself online. Since selling sex toys in country A is illegal, there is no sex shop where Am can purchase the toy. However, she’s worried that if she searches or purchases the sex toy with her work’s laptop, the IT staff could trace the traffic and expose her activity. She also cannot use her personal laptop at home since there is only one shared laptop for the whole family. Am wants a tool that would allow her to be anonymous and keep her IP address and search history from being traced. Am saw a clip on Facebook of a woman activist mentioning Tor Browser. She therefore used Tor to search for and purchase a sex toy.

The user story was inspired by the recent doxing of a Thai woman who purchased a sex toy online. According to the group members, when the toy was being delivered, the box was damaged allowing the delivery person to see inside. The delivery person took photos of the contents and the woman’s full name and address and shared them on social media. Participants shared that sexuality is still very taboo in Thailand, and particularly for women. Being exposed purchasing a sex toy can have serious social and professional ramifications.

In contrast, the second user story revolves around an activist who wants to share an online campaign about election transparency and fraud. The individual reads that someone has been arrested for sharing a similar campaign after being tracked down by authorities through their IP address. A fellow activist friend recommends using Tor Browser so that the individual’s activity cannot be traced back to their IP address.

This scenario was also born out of a real-life situation in which an activist was arrested after online political activity was traced back to him through his IP address. When sharing these user stories, it was important to remind participants that logging into accounts or making purchases using identifying information — even via Tor — affects security and anonymity significantly.

User Story 2:
น้องยีราฟเป็นนักศึกษาและนักกิจกรรมที่อยากจะเริ่มรณรงค์เรื่องความโปร่งใสและเป็นธรรมของการเลือกตั้งบนโซเซียลมีเดีย แต่ก่อนหน้านี้ ยีราฟอ่านข่าวพบว่ามีคนเคยโดนแจ้งข้อหาจากการวิพากวิจารณ์ กกต. บนโลกออนไลน์ โดยใช้ IP Address เป็นหลักฐาน น้องยีราฟจึงมองหาเครื่องมือที่ปลอดภัยที่สามารถรณรงค์ได้โดยไม่ถูกตามจับจาก IP Address เพื่อนนักกิจกรรมหลายคนจึงแนะนำให้ยีราฟใช้ Tor Browser ยีราฟจึงหันมาใช้ Tor แทนเบราว์เซอร์ทั่วไป ยีราฟรู้สึกแสดงความคิดเห็นเสรีและปลอดภัยมากขึ้น

English Translation:
“Giraffe” is a student activist who wants to create an online campaign on transparency of country B’s national election. However, Giraffe read a news article about a case in which a person shared an online petition to dissolve country B’s Election Commission because the election was a fraud, and that person was then arrested. The police arrested that person by tracing their IP address. Giraffe is worried that she could be arrested if she does online campaigning. Therefore, she’s looking for a tool that would help her do an online campaign without sharing her IP address. Her activist friends recommended she use Tor Browser, instead of normal browser. Giraffe used Tor and felt more comfortable to express her opinion online and mobiliZe her campaign.

Localization as a Tool for Education, Feedback and Outreach

Localization Sprints are not only a way to efficiently localize technologies. They are an opportunity to introduce new technologies and technical concepts, provide culturally & regionally specific feedback for developers, create stronger regional networks, and devise outreach and localized marketing strategies for sharing localized tools with the community. Together these elements can lead to more informed and sustained adoption of technologies as sprint participants better understand the tools they are using, have them available in their preferred language, identify clearer lines of communication with developers, and have more ownership over the tools as contributors to them themselves.

“I have more understanding of the concepts behind the tools. I’ve also learned that these tools are important in daily life and we can use them in real life.”

— Thai Localization Sprint Participant

If you are interested in learning more about Localization Sprints or would like to coordinate one, you can connect with Localization Lab via localizationlab.org or Twitter.

This is the first in a two-part post about integrating localized training, user feedback, and developing outreach strategy into localization sprints — events that bring individuals together to localize (adapt and translate) technologies and resources for their communities. Read Part II of this post.

“I wish to see [human rights defenders who] are at risk, especially in rural areas, using these tools as efficiently as urban-based activists. Thanks for making these apps more accessible.”
 — Thai Localization Sprint Participant

By Erin McConnell

In June 2019, Localization Lab facilitated a digital security workshop in Thailand with Internews and a number of amazing, local civil society groups. The main goal was to collaboratively localize and provide user feedback for four digital security tools that the Thai participants had identified: Mailvelope, Tor Browser, TunnelBear and KeePassXC.

With only 5 days, the localization sprint had lofty goals:

• Support freedom of speech online;
• Collect qualitative feedback on the tools from the community;
• Give Thai users ownership over digital security tools to ensure sustained use and engagement; and
• Strengthen the existing network of Thai localizers and users of digital security tools.

Participants joined from a diverse set of backgrounds and several had never heard of these tools before. A localization sprint is inherently collaborative. People with more technical experience helped those with less, and everyone put forth their best effort. It takes considerable work for a small group to translate and review four different applications in five days. And yet the feeling in the room was upbeat — people played their favorite music and felt empowered to use and contribute to these technologies. The process also challenged everyone in the room to think about how normal Thai people could use digital security tools to protect themselves online.

“I really liked this workshop, especially the tool training parts which helped me understand the tools more clearly and understand how — and in which situations — I should use these tools, which I never knew before. These tools are very useful for Thai people and we should help spread their use.”

— Thai Localization Sprint Participant

Localization and User Feedback: A match made in Pattaya

Incorporating user feedback into the localization process is the perfect opportunity to gain insight into the needs and preferences of users from a diversity of backgrounds.

The sprint participants came with varied technical backgrounds. Some attendees had trained others on several of the tools that were being localized, and others had never used them before. This provided a valuable chance to better understand the user experiences of absolute beginners and more seasoned users of each tool.

To capture some of this feedback, attendees were asked to follow and fill out USABLE Tool Task Ranking forms, customized for both Mailvelope and KeePassXC. Participants rated the difficulty and commented on tasks as they walked through steps like creating a keypair and sending encrypted email with Mailvelope, or creating an encrypted password database in KeePassXC.

At the end of each day, after participants had gone through training and localization of the tools, they were given a One Liners form developed by Okthanks to capture concise feedback about the user experience. The One Liners forms glean meaningful information from users about their first impressions of an app and their understanding of its function.

Integrating both of these feedback activities into the localization process — in addition to observations from throughout the day — the sprint documented valuable information about Thai user needs and preferences for all of the projects localized.

Localization as a Tool for Education, Feedback and Outreach

Localization sprints are not only a way to efficiently localize technologies. They are an opportunity to introduce new technologies and technical concepts, provide culturally & regionally specific feedback for developers, create stronger regional networks, and devise outreach and localized marketing strategies for sharing localized tools with the community. Together these elements can lead to more informed and sustained adoption of technologies as sprint participants better understand the tools they are using, have them available in their preferred language, identify clearer lines of communication with developers, and have more ownership over the tools as contributors to them themselves.

“I have more understanding of the concepts behind the tools. I’ve also learned that these tools are important in daily life and we can use them in real life.”

— Thai Localization Sprint Participant

If you are interested in learning more about localization sprints or would like to coordinate one, you can connect with Localization Lab via localizationlab.org or Twitter.

By Erin McConnell

When discussing access to the internet, information, and technologies the role that language plays in encouraging or inhibiting that access cannot be ignored. Recognizing this and the fact that the majority of online content and tools are available in English and not in local languages, organizers have hosted a full day Localization Sprint prior to each Arusha Women’s School of Internet Governance (AruWSIG). This year, the focus of the Localization Sprint was adapting digital safety educational materials for Swahili speakers.

The 2nd edition of the AruWSIG brought about 30 participants from across Tanzania to the feet of Mt. Meru. Over the course of 4 days, they learned about local and international internet governance and the challenges facing Tanzanians as they navigate the digital sphere and try to gain access to it.

Adapting Digital Safety Guides for Swahili Speakers

The localization sprint focused on 3 digital safety guides that address real and pressing needs for Tanzanian and Swahili-speaking netizens.

The Safe Sisters digital safety guide was developed through The East African Women’s Digital Safety Fellowship, for and by East African women. The guide emphasizes digital safety practices addressing some of the most common issues faced by women online: using social networks securely, creating strong passwords, securely sharing devices, and being aware of how information shared online can be monitored and collected by others.

The two Surveillance Self-Defense guides, How to Use WhatsApp on Android and Protecting Yourself on Social Media Networks, supplement the Safe Sisters guide by taking deeper dives into how to understand and use social media networks and WhatsApp more securely. Both guides were fully translated by sprint participants and once the guides have been fully reviewed, they will be formatted and shared on the Localization Lab website and distributed by partners and sprint participants.

KsGEN on Twitter

AruWSIG19 We'd like to share some of the moments from the workshop so far....

Why Localize for Swahili Speakers?

Despite having such a large number of speakers and an official presence in East Africa, the amount of localized technology and original content in Swahili online is relatively small and efforts to further develop and standardize technical terminology — like that of the Open Swahili Localization Project — lack resources.

This lack of information in Swahili (and other regional languages) can contribute to a litany of challenges faced by East African users as they enter into and interact in the digital sphere. Maybe of most concern is how it can affect already marginalized populations. The continued gender gap affecting womens’ access to technology and the internet was one issue highlighted by participants in an AruWSIG session on the feminist principles of the internet, hosted by Rebecca Ryakitimbo, Founder and Programs Manager of KSGen. Participants shared that the lack of content in Swahili and other local languages disproportionately kept women from obtaining technical literacy, access to the internet, and to information.

“Language is very key with the issue of connectivity,’’ said Ms. Ryakitimbo. She then asked the room how much information on family planning and medical issues was available in Swahili and local languages.

The issue of accessibility was echoed by other sprint participants, and was highlighted by Zaituni Njovu as she introduced the Localization Sprint:

“Some people cannot access content online because it is in English or in other languages. We are empowering Swahili speaking people to use the internet.”

When asked why they were personally motivated to participate in the Localization Sprint, another participant stated, “Our whole life is online and our 1st language is Swahili.” While many Tanzanians do speak English, the majority speak Swahili and access to English language education is largely determined by an individual’s financial resources.

“We need [content in Swahili] for user protection,” shared an attendee, emphasizing the need to break down language barriers for users, particularly around terms of service and privacy settings. “We are part and parcel of these technologies, including social networks,” added Diana Damson, an ICT Law student at Iringa University. “If something goes wrong, we are affected.”

For sprint participants, making technology and more online content available in Swahili was about more than basic access. It was also clear that expanding the presence of Swahili in the digital sphere had cultural and sentimental value. Zaituni Njovu shared:

“It is our culture. It is one of our cultures…People feel more comfortable when they use [content] online in their native language”

Technical Terminology and a Case Against Borrowed Words

Across the board, technical terminology in Swahili was shared as both a challenge and takeaway.

The majority of localization sprint participants shared that their biggest takeaway was becoming familiar with new terminology in Swahili and being able to talk about digital safety without relying as heavily on borrowed words from English. One participant added:

“It’s important to localize into Swahili and learn new terms in Swahili. At times we were using English terms to explain English to someone who speaks Swahili.”

Not all localization sprint attendees were ideologically aligned when it came to use of technical terminology in Swahili however. A few participants shared that using Swahili translations could make it more complicated for users who already have familiarity with English terms. They reported that some technical translations in Swahili, though following Swahili word formation norms, can be confusing or even appear nonsensical.

The overwhelming consensus of facilitators and participants was that using Swahili, and avoiding borrowed terms was the best approach. Translation should follow the standards set by projects like the University of Dar es Salaam’s Open Swahili Localization Project as well as the University of Zanzibar, one of the premier institutions for Swahili expertize according to event facilitator Bonface Witaba of the Centre for Youth Empowerment.

Diana Damson broke down her dedication and the responsibility of Swahili speakers to localize, despite the challenges:

“Changes need time. We need to be patient. We can’t give up. Going back to English will not help us. We need to put in more effort to make our language keep up with technical development.”

Special thanks to Bonface Witaba and Rebecca Ryakitimbo for organizing the 2nd edition of AruWSIG and inviting the Localization Lab to co-facilitate the Localization Sprint with the inspiring Zaituni Njovu.

From WhatsApp to Instagram, the evolution of technology has brought about some pretty interesting innovations. However, the ever-changing nature of tech has also resulted in what has been termed “the digital divide”.

In low-income countries like mine, privatized mobile network operators and Internet Service Providers are the main drivers of internet access in the country. Because of the high cost of living and huge start-up costs when setting up infrastructure, internet access in Zimbabwe is generally pricier than the regional average — leading to major roadblocks in digital innovation. While efforts have been made to lobby the government and promote infrastructure sharing in order to reduce the cost of access in Zimbabwe, the process has proven to be long and drawn out. A feasible infrastructure sharing plan which would see the players who have more resources agreeing to bring down costs seems far away.

But really, even if access were cheaper, would it actually address the digital divide? Maybe not. Perhaps, part of the solution lies with volunteer translators and the work they do to localize circumvention tools in the tech space.

As someone who comes from a part of the world known in some quarters as “The Global South”, my mother tongue, Shona, is considered a minority language. For this reason, it usually isn’t very high on the “food chain” when it comes time to chose which languages are made available online. In Zimbabwe, economic challenges have meant family members have moved away to different continents altogether. Thankfully, the advent of technology has made communication with families easier, but it has also meant that most communication is now mostly through writing (think Instant Messengers, social media, email etc…). For people whose native language is something other than other than English (or another “majority” language), an extra effort has to be made to interact with applications — something that in itself presents a barrier to many. Countries, such as mine, have not developed new words for a while now and this means that localized terms for WiFi, encryption, and interface (among other tech terms) are virtually non-existent in our language. If technology and the internet are going to make any inroads into developing nations, it is important to understand that technology will only be adopted when the local culture and language are reflected in the interface of the tools we are expected to use.

The work of Localization Lab is an opportunity to create information in local languages for communities that need that information the most. Though Localization Lab’s core mission is the translation of apps for circumvention tech, the space is an opportunity for volunteer translators like me to make a difference by ensuring that in an ever-changing world that continues to be geared towards tech my language survives, continues to evolve and does not disappear completely. Annual and country-specific collaborative translation events — or Localization Sprints — are the building blocks that have helped shape the way I look at how information on the internet is accessed and consumed by citizens in my home country. The Sprints have also enabled me to see that the challenges with language, context, and attempts to bridge the digital divide are not in vain.

The work done here enables information security, digital literacy, and anyone working in the internet activism space to work towards equipping local communities with digital literacy skills in a language that is familiar and has relatable context. As an information security trainer in Zimbabwe, this aspect is especially important given that we sometimes find ourselves in rooms full of people who have little to no knowledge of very basic Internet concepts and personal digital security techniques. With each training that I have had to adapt, change and play around with to suit the group, I am left with a deeper impression that we as trainers need to have applications and information in our mother tongue. This can only happen if there are people working to translate the information and the applications.

Through Localization Lab, the much bigger conversations around language development are beginning to take shape and that, in my teapot shaped country, is the start of the journey of 1,000 miles.

So I leave you with this thought of mine:

“The translations are a starting point. A starting point to ensure my mother tongue stays relevant, a starting point to bridge the digital divide caused by language and context, and indeed the starting point to unlock the potential of my fellow countrymen and women. Indeed all this starts with volunteering to translate applications”.

The Localization Lab knows that the work we do would be impossible without the help of incredible contributors who make up the community. If you have used Psiphon in Arabic, browsed the Tor Support Portal in Indonesian, downloaded Signal in Khmer, or accessed OONI probe in Greek, it is all thanks to community members who are dedicated to making Internet freedom technology accessible in different languages.

The Contributor Spotlight is a way to highlight some of the community members who are unlocking access to digital security technology. For this post, we want to show our gratitude to the contributor AO. We spoke with AO about his work as Localization Lab’s French-language coordinator, the importance of quality localization, and what drives him to localize Internet freedom technology.

How did you first hear about and get involved with the Localization Lab?

After three decades of working for various employers, a couple of years ago I was ready to give to the community through volunteer work. I was using various kinds of open source software on Ubuntu and I thought the translations could be better. The proliferation of mobile apps came with, what I considered to be, the spread of bad translations. I knew nothing about crowd-sourced localization at the time, but soon discovered a global volunteer force translating apps, websites, operating systems and the like. I started with Ubuntu and with some open source applications that I was using on a regular basis. I liked the challenge, and I liked seeing the result of my translations going into production.

One of the apps I wanted to work on was translated on Transifex, and I discovered a host of other projects I could dedicate myself to. I was interested in digital security, so I quickly came across projects hosted by the Localization Lab. I fully dedicated myself to the task, first translating, then reviewing and coordinating language teams. I guess the amount of work I was putting into the Localization Lab’s managed projects drew the executive director’s attention. She contacted me and asked if I was interested in attending the Internet Freedom Festival in 2015. I accepted the invitation and the rest is history. I have been coordinating the French language translations of all Localization Lab’s projects ever since.

What has motivated you to take such an active role in the Localization Lab community?

I have a lot of experience in management and information technology. So for me, quality is paramount and I see localization as any other part of software or content delivery — it has to be good quality before it goes to the users. Crowd-sourced translations can come with issues related to quality. I knew it would be a long never-ending task, but I decided to tackle it. Furthermore, with its clear translation guidelines and reviewing process structure, devoting my time to the Localization Lab was and still is a sound manner to approach the problem. I know that the quality translations offered by the Localization Lab reach millions of users and set a standard as far as crowd-sourced localization goes. Not many open source projects have the means to hire professional translators — the Localization Lab is the next best thing.

What are some of the challenges you face when localizing tools into French?

French is one of the languages where technical terminology exists natively. As with many other languages, the use of English terminology spreads in French when there is really no need. For me, one of the challenges is to make sure the proper terminology is being used. I work with certified terminologists on a regular basis and many people like me who strive to offer quality French terminology no matter what their field of expertise may be. So yes, it’s about offering products that French speakers will understand but it’s also about education in a way. I can see French terms used by the Localization Lab being used more and more in translations that have nothing to do with us and I think we contribute to that.

What drives you to localize Internet freedom tools?

The Internet reaches almost everywhere and is one of the major achievements of the human race. It has changed the way we communicate and interact with others. What is a human being without freedom? Internet Freedom tools and digital security apps help to give back freedom in oppressed countries, but more generally they help everyone to protect their freedom of speech, and the privacy of their data. The vast majority of people do not read or understand English. The Localization Lab language teams give people access to these tools in their native languages and help spread freedom on the Internet and in the digital age.

In July, Cambodians took to the polls in an election that favored incumbent Prime Minister Hu Sen. The lead up to the general election saw crackdowns on the CNRP opposition party, civil society organizations, independent press outlets and multiple international NGOs. Recently, Hu Sen’s government announced a new directive against “fake news” which would require websites to register with the country’s information ministry, leading to jail time and fines for users who are found guilty of spreading fake information. This latest move is leading to fears that heightened state surveillance will mean more attacks on freedom of expression.

Localization Lab interviewed Khmer-speaking community members who shared insights into the need for digital security tools that fit the Cambodian context and what developers can do to make their tools more accessible.

How concerned are everyday people about what the current political situation means for their overall freedoms?

It varies a lot. Many people know that the situation is serious but they are afraid to stand up to it, and other people simply don’t care. The older generations have lived through the instability of the past and they are scared about what could happen in the country if more instability arises. The younger generation seems split down the middle: Some are more concerned with their everyday lives than with the human rights situation while others are working to change what is happening.

How common is it for everyday people to be concerned about digital security?

Many Cambodians are not very knowledgeable about the digital world, so it is still not very common for everyday people to think about digital security. In general, the concept of digital security is quite new. People are more familiar with the term ICT, so it is easier for people to understand if “digital security” projects are described as “ICT” projects.

However, more and more people are seeing news articles covering phone tapping, website hacking, Facebook accounts compromised, and recently the PM said “Only 8 Minutes to Locate Cyber Insult on Facebook”.

If they do have digital security and censorship concerns, who are they threat modelling against?

Some people might threat model against hackers or some typical online money scams, but the majority are threat modelling against local authorities and the government. The government announced the creation of a “Cyber War Team” a number of years ago which was tasked with monitoring online activity that caused so-called “instability”. Also, with the support of the Chinese government, they are going to install CCTV cameras all over the capital which is supposed to monitor traffic violations, but some people are not so sure if that is the real reason behind its use.

What are the most common methods that people use to protect their privacy and security online? Are there any non-technical methods being employed (i.e. self-censorship)?

For people who have more awareness of digital security techniques, they try to only use secure tools that can protect their data and also have some personal rules for themselves like always turn off the location on their phones, use strong passwords, never click on suspicious links etc… Sometimes, they stay offline and meet face-to-face.

People have also learned to self-censor. Even if you are not personally scared and decide not to self-censor, your family and friends might pressure you to self-censor because they are scared for you.

What digital security and circumvention tools are most likely to be adopted by Cambodian users and why do you think they are more likely to be adopted?

Tools that are very user friendly are much more likely to be adopted, especially tools that work for people who don’t know how to read or write well in Khmer — so tools that use a lot of icons/symbols and Khmer words that are easy to understand. Something that is similar to Facebook (in terms of ease of use) would work well in the Cambodian context. Cambodian users are very familiar with Facebook and Facebook messenger, it is something they use daily. They are comfortable with the interface. If there were a tool that was similar to Facebook, but secure, open source and didn’t ask for personal info when users joined — Cambodians would use it.

However, no matter what is introduced, if there is no training about new tools, people won’t even know they exist. For example, Signal is not popular at the moment and many people prefer to use popular communication tools. What is the point of using Signal when your friends are staying on different channels?

What are some of the biggest obstacles for adoption of these methods on a wider scale?

There are many obstacles to the adoption of these tools in a Cambodian context. First of all, language is a big problem. Young Cambodians are really the only ones who can understand English, so when a tool is only in English it takes a long time for people to adopt the technology. You can’t just put these tools out there and hope people will catch on by themselves because understanding everything on their own is just impossible.

Another issue concerns people who can’t read or write. There is still a sector of the population with high rates of illiteracy, so it is difficult for people to use digital security tools and be safe when they can’t understand the interface. Simply put — if people can’t read, they really can’t use the tools securely.

The terminology used in the tools can also be a pain because it often doesn’t make sense when translated into Khmer. The tools really need to be localized not just translated directly.

The last issue concerns overall adoption. Many people see digital security as something that is only for activists or for frontline defenders. They don’t understand that digital security is important for everyone. There needs to be better messaging so people understand that digital security threats can affect anyone.

What can developers do to make these tools more accessible for people who can’t read or write?

Developers should create interfaces that are less complex with a lot of easy to understand visuals. Perhaps voice narration functions which are available in local languages could be useful for certain tools. Quick start pop-up screens with animations might be helpful for first time users.

In your estimate, what sector of the population is currently benefiting from the use of these tools? How can developers reach the other people in need?

At this moment, I think that only people who are activists or human rights defenders are benefiting from the use of these tools. Perhaps a small number of people who have direct contact with those human rights defenders (like their immediate family) are benefiting as well.

There is a need to really identify the people at risk, provide more training, practice and encouragement to use the tools. This sounds simple, but it is hard to do.

Developers can reach more people by making these tools easier to use, localizing them more and, if possible, working closer with the people who deal directly with users. For a small country like Cambodia, it is very hard to get in touch with developers and, the truth is, we need these tools the most.

According to Human Rights Watch, Turkmenistan is “among the world’s most repressive and closed countries” where President Gurbanguly Berdymukhamedov has held power since 2006 when Saparmurat Niyazov, Turkmenistan’s previous president, died. Since Berdymukhamedov’s rise to the presidency in 2006, Turkmenistan has maintained one of the worst “Democracy Scores” according to Freedom House’s 2017 Nations in Transit Report and has been labeled “Not Free” over more than 10 consecutive years in Freedom House’s annual Freedom of the Press report, which scores the legal, political and economic environments supporting a free press and freedom of expression.

Getting information about users of circumvention and digital security tools in Turkmenistan has been notoriously difficult which is why we recently chatted with Localization Lab contributors from Turkmenistan in order to better understand what current limitations on Internet freedom and freedom of expression mean for Turkmens, how they perceive censorship and surveillance and how they access information.

What are the primary languages used in Turkmenistan and what is the literacy rate?

The official language of Turkmenistan is Turkmen and that is the primary language used in primary and higher education. Since 1991, Turkmen officially moved from being written in Cyrillic Script to Latin Script. While English has recently been pushed to replace Russian as a new “language of inter-ethnic communication,” there were an estimated 900,000 Russian speakers in Turkmenistan as of 2012, most in urban areas. English is not widely spoken.

Literacy rates are high in Turkmenistan are very high, estimated at around 99.6% among Turkmens over age 15 by UNICEF in 2013.

Do Turkmens generally have a strong technical background?

The majority of Turkmens do not have significant technical knowledge. Individuals use smart phones primarily to access social media and messaging.

What is the state of Internet infrastructure in Turkmenistan?

Turkmenistan has some of the slowest and most expensive Internet access in the world and is managed by TurkmenTelekom who has a full monopoly over Internet access in the country. Unlimited Internet plans in Turkmenistan currently range from $38.50 to $71 per month with speeds between 250 kbps to 2 mbps dependent on plan, however TurkmenTelekom advertises these plans as only available through December 31, 2017. Outside of the temporary discount, unlimited plans offering speeds of 128 kbps up to 1 mbps range from $28 to $85 per month. To put this cost to the individual in context, according to the CIA Factbook, the average monthly salary for a Turkmen was $616 in 2010.

For more information:
Turkmenistan — Information and Communication Technology
Turkmenistan recognized as the country with the most expensive Internet rates
Turkmenistan: Nations in Transit 2016
Turkmenistan: Want Internet Speed? Yours for Just $7K

What are the popular social media and messaging services in Turkmenistan?

• Facebook
• Twitter
• VK (also known as Vkontakte)
  VK has about 90% penetration in Turkmenistan.
  Two popular pages in this platform are: Turkmenistan Limited and Turkmenistan Limited #1
• OK (also known as Ondoklassniki)
• Telegram
• WhatsApp
• LINE
• imo

How are individuals affected by censorship and surveillance in Turkmenistan?

Turkmens have a general understanding that everything is being monitored in the digital sphere.

Criticism of the government is heavily punished creating fear and self-censorship. Even Turkmens living and studying outside the country fear “family punishment,” or that their online behaviors from outside the country will result in consequences for their relatives still living in Turkmenistan.

What are some of the sites and services that are blocked in Turkmenistan?

Sites that are blocked change regularly and individuals will switch tools or use circumvention methods based on what is available.

Social Media and Messaging:

Facebook
Twitter
WhatsApp
Telegram
VK (Blocked on occasion for unknown reasons)
LINE (Sending images over LINE was recently blocked which may be a reason for the recent popularity of imo as an alternative)

Media:

Альтернативные Новости Туркменистана
Chronicles of Turkmenistan

Browsers:

Opera
Firefox
Puffin

Do people use circumvention methods?

Yes, due to the extensive reach of state censorship, use of circumvention tools is popular across Turkmenistan regardless of your political, educational or technical background.

What are some of the methods individuals use to access blocked content?

People generally prefer to use VPNs and proxy applications in order to access blocked content as opposed to using browsers like Tor Browser.

As VPNs and proxy tools gain popularity, they are blocked by the government, resulting in a cat and mouse game. There is a constant scramble to find new VPN options and many are marketed on social media sites like VK. Who is behind this regular influx of circumvention tools is uncertain. A related question is who is behind open advertising of certain of these circumvention tools on social media sites like VK considering that promoting digital security and circumvention tools can result in consequences from the state.

A couple of censored news sites will post content on Facebook, including Альтернативные Новости Туркменистана and Chronicles of Turkmenistan. Individuals may fear that accessing certain censored sites is risky, even while using a VPN, however accessing Facebook via circumvention tools is very common and individuals will even like the Facebook pages of censored sites.

What kinds of content are people accessing through circumvention tools?

Youth (under 30 years) are the largest consumers of online services and make up about 60% of Turkmenistan’s population. Their online priorities are primarily networking and social media, chatting, and accessing information about education and migration abroad.

Accessing blocked media and blocked civil society or human rights content is not a high priority.

*According to Psiphon, there are about 70,000 users of Psiphon in Turkmenistan. The most visited sites by those users are YouTube, Google and VK (Russian social media network).

Are there consequences for using circumvention tools?

People fear using a circumvention tool can be incriminating and that using one on e regular basis could result in government retribution. For this reason, individuals will often use a proxy tool or VPN only to access blocked content and then turn it off for all other uncensored activities.

Individuals who advertise and share information about circumvention and security tools run the risk of facing consequences from the government.

What is the importance of Open Source and Free Software in Turkmenistan?

There is not a broad understanding of what Open Source technology is or what its benefits are. Along those same lines, there is no understanding of the potential threat of using proprietary tools.

Software that is free of charge is incredibly important in Turkmenistan however. As a closed country, international payment systems do not work in Turkmenistan and Turkmen bank cards do not work in the Google Play Store and Apple AppStore. If applications are not free for download, they are only accessible to the minority who have access to a foreign debit credit card.

Are there local human rights and civil society groups in Turkmenistan?

While there is an activist and NGO community in Turkmenistan, it is a very small, trust-based network. The size and scope of local organizations is inhibited by government restrictions on accessing foreign funding.

There are several international and regionally focussed organizations that work within Turkmenistan however, including: Human Rights Watch, Amnesty International, Arzuw Foundation, Crude Accountability and Eurasia Foundation.

How could digital security and circumvention tools potentially be shared with the public in Turkmen?

There are several international NGOs (listed above) that work in Turkmenistan that might be able to assist with outreach and networking locally in Turkmenistan, including connecting with the small community of local organizations. It is also possible to utilize popular social media and messaging tools to do outreach.

LINE and imo: Market through popular LINE and imo channels
Facebook and VK: Advertize and do outreach through media sites that have been censored but have a Facebook or VK presence.