Texts can now also be translated automatically. The automatic translation of texts from the source language, the source language, into the target language is with the help of translation software possible. Machine translation is a branch of computer linguistics and computer science. There are numerous translation engines on the internet. These include google translate, babelfish, leo, pons, mymemorie, and others.

Historical background

The way was marked decisively translation software by the interest of the military. The first project of a translation software was a translation program of the US military, which translated from Russian to English. So it was possible to get a first impression of the content of Russian documents without consulting a translator or interpreter. The software was very popular in the US military, although the program was of very poor quality. The research on translation software came to a standstill in 1966 for almost 20 years because the project was considered unrealizable. The renewed research began in the 1980s. Co-operation between government, universities and electronics companies led to the first PC programs. Japan thus became the world leader in machine translation research. Further programs and research followed, also in Germany. Increasingly statistical methods were used in the 2000s. Google offers such a program with google translate since 2006. In addition, other programs were developed, such as rule-based.

Machine translation is not perfect

The translations, which with the help of a translation software are not perfect. Dealing with language is so obvious to us that we are seldom aware of the diversity of the language. This diversity is problematic for machines and programs. A digital system is based only on ones and zeros. Languages, on the other hand, have deep structures and diversity in rules and grammar. By contrast, the machine translation programs are based on the surface structure. A translation software therefore translates one word after the other. But there are also translation software that deals with the deep structure of the language by analyzing, categorizing and recognizing the structure sentence by sentence. Then the translation software transfers everything into the structure of the target language. Not only the diversity and complexity of the language causes problems for a translation software, but also the ambiguity of words, sentences and grammatical structures. Machines also have no experience and world knowledge.

Before translating:

For the translation software to work properly, a few things should be noted. The translating text should be formulated grammatically correct. In addition, should be dispensed with colloquial language, as it is often not recognized by the programs. Furthermore, attention should be paid to the correct spelling and punctuation. Finally, care must be taken to avoid complex and long sentences in order to translate a text correctly.

The different types of translation software

There are several types of machine translation. With the help of digital dictionaries, for example, individual words can be translated. Online dictionaries often have several entries for a searched word. Thus, the sought term of meaning can be adjusted. There are also word for word translation programs on the Internet. Each word from the source text is translated by a word in the target language. Often, however, this translation makes little sense. Originally, these translation programs were used to interactively determine the content of a text in another language. Meanwhile, there are also full-text translators, which are used more often for such tasks. Full text translators are sentence by sentence translators. However, these are not error-free. Therefore, a text should be edited before translation with a full-text translation software. Nested, or very long sentences should be shortened, so that many single short and simple sentences arise. These can be translated better.