Dwayne offers tips on getting started in localisation

By Frederick Noronha

Dwayne Bailey (34) lives in Pretoria and comes from Cape Town, South Africa. In the world of Free Software and Open Source, Dwayne is well known for preaching the localisation gospel.

Says Dwayne: "Actually localisation is everything that makes the computer work for you in your locale (country and language). Translating the computer interfaces is by far the biggest task and ongoing. But its not the complete picture. Keyboards, fonts, locales, date systems, rendering, bidi are all part of localisation."

Dwayne has been involved with a major translation project in South Africa. As he put it when we last spoke, "At Translate.org.za we're localising Free and Open Source Software into 11 South African languages. One is English, it's quite an easy one (smiles). The others are Afrikaans, Zulu, Xhosa, Venda, Tsonga, Tswana, Siswati, Northern Sotho, Southern Sotho, Ndebele." [2]

But this time, at Africa Source 2, we focussed on another issue. What's the best route for a small-to-medium language to take, if it wants to enter the world of computing?

Says Dwayne: "What I'm discovering now is that the first things to do (for any language) is to get in place the basic infrastructure. One reason is so that they can start, the other is so that they can write and do things in their language."

Three things need to be checked, he explains.

Firstly, whether you can type your language, and have a keyboard to do so. Then, you need to be able to see the text, so font-rendering is the issue here. The third important task is to define a locale for language and country.

(In simple language, this is a configuration file that defines whether your language uses the metric or imperial system, what are the names in your language for the days of the weeks and months of the year, what calendar system do you use, and so on...)

After this, what?

Then, says Dwayne, you're really free to do translations. That's when you need to really start thinking about what you want to achieve. Even a small group can achieve change, he believes.

"The reason why you need to define goals is to ensure you don't kid yourself about what you're doing. If you're localising [GNU]Linux into Bulgarian, and you say this is being done so that Bulgarians can use a computer in their language, when in fact they all use Windows, then you need objectives that are based in reality," he explains.

Your objectives could be anything. "I'm happy if someone says I'm translating Thunderbird, the email client, into Afrikaans for my 100 year old granny," jokes Dwayne, whose specs and lean figure makes him look much more serious than he is.

The import thing is to understand what your goals are like.

"It's all about focus. Your objective helps to focus where you put in your energies. In our organisation, we found we were being pushed from pillar to post. Everyone wanted us to translate something, and there was always a good reason," he narrates.

But what works well for a small language?

"My feeling is that you have to translate stuff that focuses on the end-user. And that narrows down your scope. End-users are the people that are mostly going to benefit from whatever you do," he stresses.

Dwayne also suggests that localisation teams look at cross-platform tools, i.e. those that work on multiple operating systems such as GNU/Linux, Windows or Apple Mac. "Even within that, I would define (the more useful tools as) a sub-set: anything to do with communication is probably the most important thing to work on," he explains.

Has he emerged wiser from his experiences of past work? "If I was going to do things again, I would prioritise it in the following order: email client, instant messaging and a word processor. When you look at (the whole of) Open Office, it's a relatively large project. But you can (begin by) translating only the word processor."

Can localisation in computing really make a difference to a language's future?

Dwayne narrates the "interesting story" of the Venda language, which has some 700,000 speakers. "We needed to translate it. I found there were some extra characters needed -- beyond the Latin characters. So I investigated how to make a keyboard, and made it just for fun. I made some fonts."

One translator in the team was a linguist. He talked about 'mechanical imperialism', where the deficiencies of the computer were changing the way people could write a language. That is still a problem.

This worked itself out in odd ways: a professional translator employed by the South African parliament to translate into Venda, couldn't type all the characters needed. So they would type things out, and then add the 'missing' characters by hand. "Which is completely sad. Valuable information which could be created for the language is completely lost. That's a demonstration that simple things could do amazing things for a language," says Dwayne.

Isn't it sometimes an uneasy relationship between techies and linguists, both of whom need a stable partnership to make a translation project a success?

Says Dwayne: "Whereever there are issues about techies and translators, the reality is that techies don't appreciate translation. They appreciate translation when it looks good in a press release; but in their behaviour there's not that kind of care. But having said that, there are certain projects where there's a growing respect for the localisers. The key thing there is usually to have a representative for the localisers (in tech teams). They often act as the go-between."

Check the TranslateWiki [3], a great starting point for localisation of your language. It offers links to: The WordForge project; The Translate Toolkit (a toolkit to convert between various different translation formats); Pootle (a portal that will enable you to manage your translation project, do web-based translation and offline translation); the Localisation Guide (a guide to how to start and run a localisation project) and a glossary of translation terms.

Once programmers see localisers as valuable members of the community and once localisers see responsive programmers both sides begin to see a very healthy relationship, argues Dwayne. "Something that can take us onto 100 languages."

Here are some figures that remind us of how serious the task is: There are about 239 languages in Africa with more than one million speakers. Free/Libre and Open Source Software covers about 50-60 languages in total. In Africa currently covered are 10! Says Dwayne: "Localisation is a task that many people can do and clearly we need lots more people to contribute."

[1] http://en.wikipedia.org/wiki/Software_localization [2] http://www.tacticaltech.org/node/237 [3] http://translate.sourceforge.net/wiki/