By convention the default locale of all applications is US English. This is of course very imperialistic and evil and the Americans are indeed forcing their culture upon the rest of the world. But in the end we need to have a default for the ‘C locale’ and it was decided to stick with the language used in the place where most of the modern computing actually originated. Using Latin would have been a bit awkward and even Esperanto isn’t entirely culturally neutral as well.
One could argue that Americans—no, I’m not going to write USanians—derive a large advantage from the fact that the default locale is their English variant. All software is understandable for them right from the beginning. They never have to wait for translations. However, in this piece I would like to argue that actually it is a disadvantage.
Why would it be a disadvantage to Americans that all software automatically suits their customs and follows their local quirks? Well, for that I would like to do a game of compare and contrast. Mostly contrast. You see, the US English strings are the only texts written by the developers themselves.
Development attracts people who like to develop, not people who like to write. They do not necessarily come from the United States, often are not native English speakers and many of them can’t see the use of arguing about -ize vs -ise, or have own opinions about it. The consequence is that the US English strings are written by people whose primary interest is writing code, not human language. This is detrimental to the quality of the texts, the suitability of the chosen phrases and spelling and grammar in general.
Translation teams, however, attract people who are interested in language. In the world of perfect localisation, all typography nuts, grammar enthusiasts and spelling bees will join together to form a team with Super Language Powers. This means that the people who will write the text you see every day on your computer are fond of language, know how to use it and have experience to say it, if you speak any language or dialect other than US English.
All languages—except US English—have a corrective filter between the developer’s work and the end-user. There is one community that oversees all use of language in the product. Translation teams often work with word lists, style guides and selection of contributors based on their quality. This allows them to guarantee quality, make sure that all text on the system follows the same conventions and warrant consistency across the desktop. You can correct for overuse or underuse of capitalisation, distinct between the computer and user in events by using different verb conjugations and so on.
Consistency is an important issue. For example, a computer can have a screen, a display and a view. These words are near synonyms, but the X server uses them to distinguish between three different things. It is hard enough for a user to understand what the system is talking about already, it becomes even harder when words can have different meanings in different applications. When there is no central organisation of the terminology, this does happen. Translators could correct for this by adapting the translations to the context, but Americans are out of luck.
There are no people looking after the typography, grammar, capital use and readability of US English. But there is more to localisation: translation teams also make sure that the system is using the correct date format, currency, decimal delimiter and so on. Each country has its own conventions here. No one has the job of nitpicking about the American conventions, so they’re missing a watchful eye here as well.
The result is that the US English desktop can often be inconsistent in style, word choice and spelling. This makes our product less appealing to Americans and to other people using the US English version. If we want to pursue perfection, we should not miss this out.
It has also consequences for the translations. The translations are translations of the original English texts. Although I did say earlier that translators can correct for inconsistency and bad wording, they don’t always do. It is a lot of work to manually check the context of each and every string, many translators just stick to translating every word with the same phrase. Badly used capitals and dots will often find their way into translations as well. Vague US English results in vague translations.
Improving the quality of US English will mean large improvements for all languages if it is done properly, by sending patches with corrections to the developers. I am convinced that we need an American ‘localisation team’, consisting of all American typography nuts, grammar enthusiasts and spelling bees who want to contribute to Ubuntu and FOSS in general. They could work together with other projects to establish conventions and methodically go through all applications to check whether they comply with these conventions.
We cannot ask from all the localisers to understand programming language and patching systems. However, with the current state of technology, I am afraid that writing patches directly for the code is the only option. In the long term, something like a POT editor and a reverse POT generator could improve things.
It would also require infrastructure. Many languages have project-agnostic communities for translation in FOSS that provide various language-related services; examples are the French Traduc and the Dutch OpenTaal. These relations are often not formalised, but they are really helpful in making sure everyone is somewhat following similar rules. As part of FreeDesktop an American initiative could be started, which could keep a list of the standard meanings and uses of words.
The solutions above are just ideas for ways to deal with a problem that we should give much more attention than we have done so far. Admittedly, it is easy for me to talk from the sideline, knowing that I—being a native speaker of the Dutch language, not of US English—will never be doing much of the work I propose. But I do hope that some people will be inspired by this piece and do something with it.
Those poor Americans deserve localisation too!