Lost in translation
Consumenten beschermen door informatie te censureren

Lost in Translation

Can you do independent research with colored glasses?

A peek behind the curtain of building an internationally compatible tool, from an American intern at a Dutch digital civil rights organisation.

A large part of my work this summer at Bits of Freedom has been focused on the internationalization of a website still in progress, My Data Done Right. On the face of it, My Data Done Right’s goal is simple: to allow users to send requests for data to organisations. It does this by generating a letter or email for the user to then send to the organisation. Making this a tool that can be expanded for all of the EU, not just the Netherlands, comes with some interesting challenges.

Translation troubles

Some seem obvious: we need a (changing) database of organisations from every country and the website needs to be available in every EU language. However, the implementation of these brought up some issues. As Bits of Freedom only uses open-source technology, we could not, for example, simply use Google Maps’ search function to find different organisations. Instead, we are compiling our own database, relying on volunteers throughout Europe. In addition, we are also relying on volunteers to help translate the content for our website. Automated translation tools like Google Translate, which can automatically translate an entire website, are problematic for two reasons. First, it is not open-source, so it is against our policy to use. Second, machine translations commonly have errors that can only be caught by a human reader. This is why Bits of Freedom decided to use translation volunteers instead. To find volunteers we rely on our European Digital Rights (EDRi) network, the umbrella organisation we helped found in 2002.

Weblate as facilitator

Yet even with volunteers, the problem of translation is not solved. The website is still in flux, and the textual content of the website is stored with the code that creates it. So we had to find a translation tool that would collaborate well with our version control tool for the website’s code. Ensuring that the tool facilitates communication between translators and website developers, without creating issues for the website and while being easy to use for our volunteers, has been its own process. Weblate, an open-source translation facilitator, has been our solution. In addition, on the website we have implemented checks to ensure that no partial translations are accidentally uploaded. For the translation tool, we have customized Weblate to fit our needs as much as possible. For example, allowing many different languages to be added and regularly syncing the translations with the website.

This spurred some spirited lunch-time discussions among the team: should we be using the American-English spelling “organization” or the British-English’s “organisation”?

Spelling issues

In addition to these larger issues, some smaller language issues popped up as well. For example, the GDPR does not just apply to companies, but to all sorts of groups. So rather than use the word company throughout the site, we use the word organisation. This spurred some spirited lunch-time discussions among the team: should we be using the American-English spelling “organization” or the British-English’s “organisation”? I, an American, had not even considered the British spelling. Other coders as well had already used the American spelling, with a z, in the code. After several discussions and some back-and-forth, we settled on the British spelling.

Finding a format

Since the tool generates letters, we also had to consider the address formats. There is no common address format across the European Union; there is not even a common postal code. The Netherlands uses letters and numbers, while other countries uses different lengths and only numbers. Some countries prefer the postal code before the town name; some after. For our website to return accurate addresses, we have to consider each and every one of these cases, not just for generating the address for the user, but also when a user is submitting an address for a company not in the database. Yet Bits of Freedom simply does not have the resources to consider every nuance in these cases, and at least for the beta version of the product we are just using the most common address format found in the EU.

An American perspective

Coming from the United States, these issues were new to me. On projects I’ve worked on in the past, English was the default and sole language. I hadn’t considered the intricacies involved in making a project accessible to people from so many varied backgrounds. While initially frustrating, overall these translation challenges have made this project more rewarding to me. In addition, it has left me with a greater appreciation for every website I come across that has multiple languages. In the case of My Data Done Right, to me the frustration caused by translation is overshadowed by the reason we chose to internationalize it in the first place: to empower residents across Europe to control their data.

I had my own issues translating into Dutch culture. For example, mentioning casually to the team that I’d dropped by a coffee shop on my way to work raised some eyebrows (I’d really meant café).

Translating into Dutch culture

In addition to the internationalization issues I saw during the course of the project, I had my own issues translating into Dutch culture. For example, mentioning casually to the team that I’d dropped by a coffee shop on my way to work raised some eyebrows (I’d really meant café). Or consider the Dutch emails and messages; more often than not, I would use Google Translate; (ironically, as we did not use the tool for My Data Done Right). I was surprised by how helpful Google Translate was with reading messages from my colleagues. My inability to speak Dutch did block me from attending some events, like Bits of Freedom’s cases in the Dutch courts. Yet I found that with technology like Google Translate I could understand almost anything virtually, even if some of the phrases ended up awkward in English. And in person, Bits of Freedom employees spoke English at lunch and events to ensure I could participate.

Finishing my internship

I will be leaving soon, but after my departure Bits of Freedom employees and volunteers will continue to work hard to get My Data Done Right off the ground. I’m looking forward to seeing it launched. By coming to work at Bits of Freedom this summer, I was expecting to get a glimpse into the workings of a digital civil rights organisation. I did, but in addition I also ended up learning about the challenges of creating projects for an increasingly connected world. For that experience and many more, I am grateful to Bits of Freedom for hosting me this summer.

We are still looking for volunteers to help finish My Data Done Right! If you would like to volunteer with help translating or with gathering organisation information, please send an email to david.korteweg@bof.nl.

Geef een reactie

Het e-mailadres wordt niet gepubliceerd.

Deze website gebruikt Akismet om spam te verminderen. Bekijk hoe je reactie-gegevens worden verwerkt.

Help mee en steun ons

Door mijn bijdrage ondersteun ik Bits of Freedom, dat kan maandelijks of eenmalig.