This concept of comfort is expressed well in the story of an Afrikaans speaking computer programmer in South Africa. He refused to use software in Afrikaans. Since he had always used computers in English he found English easier and more natural to use on computers, Afrikaans to him seemed foreign in this context. However, he used bank atms in Afrikaans. Why Because he has always done banking in Afrikaans and struggled with words like account, statement, withdrawal and deposit while using an atm. Afrikaans in this context was so much easier for him to understand. Comfort was the motivator and is in many cases where people translate computer software.
The second motivation is that local language software actually makes a difference to people’s lives. Unesco and other researchers promote mother tongue education as it has a number of quantifiable benefits that lead to better education of children. This includes participation in the learning process and the learning of fundamental concepts 4.
4 See in this book : Marcel Diki-Kidiri, Cyberspace and Mother Tongue Education.
Dwayne Bailey The same findings can easily be extrapolated to software. In the same way that local language education leads to better marks, better retention of concepts and student participation we could expect local language computing to lead to more active and involved users. Users who know the position of menu items needed to execute a task are disempowered compared to a user who can read the entry and make an association based on their understanding of the language of the interface. Anecdotal stories of training conducted in the Western Cape province of South Africa using English software but providing all training in Xhosa showed a marked improvement in participation and results from the students.
The previous two reasons related to the computer and the users, but there are a number of other valid reasons for localising software that have more to do with the development of the languages. Foremost is the development of skills in the language.
It’s no secret that the first localizations into any language are poor.
Localisers will deny this, especially commercial localisers since they have been paid to produce good translation. The low quality has little to do with the lack of translation skills of those employed to do the localization.
They are poor for the following reasons. Firstly, you are using translators who have no experience in software translations. Second, there is usually no developed terminology in the domain. Third, general style guides are not available in the language.
But the act of localising helps to develop the skills of translators so that they can translate computer software. The act of localization develops the language. We develop the language when we coin terms for words like wordprocessor and spreadsheet.
Without style guides and terminology we get inconsistent translations.
Many people approach the issue of poor quality of localization by developing these first. Unfortunately anyone who hasn’t localised and who develops a style guide or terminology will get much of it wrong. Which goes back to our thesis that the first localization will be bad. But using knowledge and developing style guides based on using the language leads to better and better translations.
This is one place in which open source comes to the fore. With slow release cycles there is the very real concern that poor terminology choice and style is accepted as correct. Many languages have examples where the translators Dwayne Bailey followed English style closely without much thought to the style of the language itself. The easiest of these to spot is that of Title Capitals used in software. While English does this regularly most languages use sentence case. Yet software is often translated, even in these languages, using Title Case. Where open source wins is that these mistakes can be corrected very quickly as the style and terminology emerges for the languages.
Localization helps to promote the status of a language and change the perception of the language. In many minority languages, the topic of translation often creates heated debate about whether software should be translated. The debate has two perspectives : “We don’t need translation, we should use it in English, French, etc”. or “Our language cannot be used for computer programs”. Interestingly both arguments against localization prove to be false. While most computer users in a given language will use software in a dominant language, it is precisely because they speak these languages that they don’t have any trouble. Their reflection of a lack of need is expressed based on their own experience not based on the experience of the majority of speakers in their language.
The fact that a language does not have a word for a computer concept is easily addressed by creating these missing words. The reality is that many words needed in translation already exist. Words like proxy, password and authorise will already exist in a language. Thus many words perceived as being about computers have much wider application and have been used for centuries.
Usually the debate about whether a language can handle software is rendered moot with the release of the first translation. Whether it is usable or useful will be determined by the longterm uptake of the software. But the idea that the language cannot be used for software is proved wrong.
Why is open source a better environment than commercial software for the promotion and development of languages Localization requires a given set of skills which can be learnt in training but can really only be honed by practice. While efforts like Microsoft’s llp program are creating localised software they are not safe platforms on which to create skills. When commercial companies hire localisers they expect their software to be localised well. For a new language these Dwayne Bailey localization skills and resources do not exist. A professional translator is a good substitute for an experienced localiser, but they are not the best localisers.
Why do we say that open source is a safe environment Does it have lower standards It is a safe environment because in Open Source mistakes are tolerated and quickly fixed. Quality is enhanced by acknowledging that mistakes will happen and allowing them to be corrected easily.
Both open source and commercial translators try to ensure quality through these additional steps in their translation process. Firstly, translations are reviewed by a second translator. Second, an in-product review is performed by translators using the actual product. Performing a review assumes that the translators know the idiosyncrasies in their language when translating software. Our argument is that for a language that has never been localised, nobody knows these issues. These have to be learnt.
A review by someone with no knowledge of these issues is a superficial review. An in-product review by a person who does not use localised software and is often not even an expert in the translated product, is of questionable value.
Open Source has two benefits for the development of the language. It has a lively community that can easily give feedback about terminology, style and other translation issues. Release cycles in open source vary from 6-months, while commercial software is more like 3 years. This allows new translations to appear quickly. In contrast the llp translation of Windows xp into Zulu, once it was reviewed and released, was not updated. The next Zulu version was Windows Vista. Mistakes will happen and open source creates a proper environment for testing where users can give feedback that can result in changes in the product in a short space of time. This prevents bad mistakes from becoming entrenched in the language.
Open source benefits languages through the creation of open resources.
A number of commercial companies have added terminology and style guide development as a prelude to translating into new language. This is a good step as these two resources are very important for localization.
However, these resources are easily created for well established languages but fit badly for emerging languages. For established languages the resources exist in some form or another and simply need to be repurposed for the job at hand. Spelling and grammar rules and bilingual dictionaries Dwayne Bailey can be used to create the resources needed for localization. Now, imagine creating a style guide when there is no general guide in a language and no experience in localization. It is a useful starting point but it is definitely a work in progress and needs to be thought of as that. Terminology creation has the same issue. Gathering a number of language experts together to coin 6000 terms without really understanding their use in software will always be problematic.
But probably of most concern for an emerging language is the availability of these resources once created. In a case in South Africa, Microsoft created 6000 computer terms in various languages. These were reviewed by the Pan South African Language Board (PanSALB) and approved. However, the list of words were not generally available to the public or to other localisers. This was shortsighted on the part of the language board, as it would have been easy to require that these resources be made generally available since the approval made use of state resources. To Microsoft’s credit these terminology lists (and style guides) have been made more generally available, but it’s taken a number of years.
In contrast open source by its very nature must create open resources. A terminology list or style guide created for an open source translation will be released under an open license allowing others to reuse and improve the guide as needed.
By translating open source it is possible for translators to hone their skills, develop new skills and develop new markets. In an environment where dominant languages are accepted there is no need for translation skills. In an environment in which users make use of localised software there is a growing need for more localization. By localising open source software it is possible to stimulate a demand from users for local language software, which leads to a demand from software producers and thus creates new work for the translation industry.
While this market is growing, open source allows professional translators to build their skills. They get to actively translate without the business risk of poor localization. They get to work with localization software, follow formal localization processes and receive feedback from localization engineers. All of these are skills that they can reuse in commercial localization.
Dwayne Bailey In a university or training environment, open source is a valuable tool with which to teach localization principles to students. It presents a real localization environment, as opposed to a sterile simulation. And students are very aware that their work will be used by real people in real software.
Open source and open content can grow a language community that can achieve much more than individual people and companies. Open source is predominantly a volunteer movement. This is important when it comes to local languages. When people are aware that their time and effort can impact their language then it ensures that their language has a future.
If five people are paid to translate Wikipedia at one article a day for a year then we’d have one thousand new articles at the end of the year. If fifty volunteers translate one Wikipedia article a month it would result in six hundred articles in that year. The speed is not as great, but the effort is more sustainable. Each of the fifty volunteer translators has the potential to bring in new people to translate, while the five paid translators won’t be doing any recruiting. Once the payments stop the five translators disappear while the fifty volunteers are more likley to continue. A combined approach would be very powerful as it would allow professional translators to lay the groundwork for volunteer translators. After all it is much more exciting to add your one translated article to a corpus of one thousand articles than it would be to add your one to the existing twenty.
In the same way a large group of software localisers can localise much more software than one or two individual translators.
ANLOC : AFRICAN NETWORK FOR LOCALIZATION The African Network for Localization (ANLoc) has been able to apply these principles of open source localization. Within ANLoc we’ve seen open source provide an amazing platform in our ‘localise software’ and ‘training’ programme. The end result has been a number of pieces of software translated into local language. These included Firefox in Northern Sotho, Luganda, Akan and Songhay. Tuxpaint, a children’s drawing program. Abiword, a word processor and vlc, a media player.
During the ANLoc training we used the Pidgin instant messenger client and Virtaal Computer Aided Translation tool to teach localization principles. Both Pidgin and Virtaal are open source application softwares.
Dwayne Bailey Using Pidgin allowed us to easily demonstrate the results of the translation to the participants. One of the participants went on to complete the translation of Pidgin into Swahili, on the way becoming a better localiser while stimulating demand for more localization.
This is Firefox operating in Swahili with the preference dialog open. The user is looking at the Afrikaans Firefox download page. Both languages are supported by our efforts in Anloc.
By being able to demonstrate localization issues in Pidgin we had feedback from a commercial localiser who’d localised Microsoft products, that for the first time he understood some of the issues of localization that he hadn’t before. While you don’t require open source to demonstrate these principles, it is worth realising that none of the trainers developed Pidgin, yet the software was easy to modify to add the translations. The students could see these issues in a real piece of software that they knew would be used by real people.
Материалы этого сайта размещены для ознакомления, все права принадлежат их авторам.
Если Вы не согласны с тем, что Ваш материал размещён на этом сайте, пожалуйста, напишите нам, мы в течении 1-2 рабочих дней удалим его.