Machine translation technology began to be explored in the 1950s, but the first commercial machine translation system appeared in 1991, with web applications appearing a few years later.
Machine translation technology now ranges from free online translation, to on-the-go mobile phone apps to customisable, professional software packages.
There are two major benefits of machine translation. First, it is a much cheaper alternative to human translation. Second, it is much quicker, allowing on-the-spot translations.
However, there is a lack of nuance in these translations, as a human translator can bring in-depth knowledge of the nuances of a language. Machine translation is a direct translation without any localisation process, which is why it can sometimes produce badly ordered sentences that are difficult to read.
The variable quality of machine translation does not matter in ephemeral interactions such as instant messages or ordering an item on a menu. However, mistranslations cannot be excused when it comes to high quality content that is needed in a professional environment.
A well known machine translation failure was the Chinese café that wanted a sign in both English and Chinese. The Chinese text, 餐厅, means “dining hall,” but the English sign read as “Translate Server Error.”
The translation software produced an error message, but because the software’s user did not know English, they assumed the error message was the translation, which is why it went on the sign.
This is a humorous example of a translation blunder, but machine translation failure can lead to serious problems, such as damaged reputations and the loss of customers.
Here at Lingo24 we directly control the development of the engines we market. We offer a range of services, customised to fit each client’s needs.
Our customised engines significantly outperform generic Machine Translation (MT) programs. Each one is built to fit a client’s exact requirements, including industry sector, subject matter, preferred terminology and style.
We can build a high-quality MT engine within two weeks, provided there is a good supply of data available. If you do not have any data, we recommend our generic engines which cover a wide range of domains and subject areas. In this case, we can build a new engine for a specific domain and language pair in less than six weeks.
Want to find out more? Just get in touch with our team!