Language Translation Software:
Still in Infancy – but Promises are High

People translate foreign languages into their own for centuries. However, language translation software was invented only in the 20th century when people started to mechanize a translation process.

As early as 1933, the Russian inventor Petr Smirnov-Troiansky patented “a machine for selecting and printing words in translation from one language to the other.” March of 1947 is considered to be the birthday of machine translation. At that time, Warren Weaver, a scientist from the Rockefeller Foundation, wrote a letter to the cyberneticist Norbert Wiener and first suggested using computers for translation.

Weaver outlined his ideas in a special memorandum of 1949, which became a stimulus to further research activities. Public authorities and businesses started to give special attention to and invested heavily in this field. In 1954, the first result was demonstrated jointly by Georgetown University and IBM – an IBM computer translated smoothly more than sixty Russian sentences into English. The same year, the first machine translation experiment was run in the Soviet Union.

Today, machine translation has a 50-year history of ups and downs. The 1990s became a major milestone. The decade brought swift progress in the information and telecommunications sectors, including personal computers, programming aids and techniques, and Internet that created boundless opportunities for international communication. However, new opportunities appeared on the scene along with new challenges, and a language barrier was one of the most formidable of them. That spurred demand for language translation software products.

Rule-based: The translation system uses a set of grammar rules for each language to analyze words, grammar, and punctuation and produce a translation. Statistical: The translation system has millions of previously translated texts fed into its database in the source and target languages. The system produces translations by looking for statistical relationships between the words in the database. Example-based: The translation system uses a set of sentences in the source language and their corresponding translations in the target language as examples to translate other similar sentences by analogy.

Systems for automatic translation

Language translation software for automatic translation has been developed on the basis of three approaches: rule-based (traditional), statistical and example-based. To improve the output quality of machine translation systems, combinations of these approaches are used too.

In the attempt to satisfy all categories of potential users, developers of language translation software have developed different automatic translation products:

- for general public (e.g. Systran Web Translator),

- for corporations (e.g. @promt NET Professional),

- for professional translators (e.g. Personal Translator PT2008 Professional),

- for websites to translate webpages (e.g. InterTran Web Site Translation Server),

- for use with various mobile/hand-held devices (e.g. Honyaku Kore-Ippon for Zaurus).

In addition, there are some systems to translate speech (e.g. Multi-language Translator).

All these systems can be bilingual or multilingual. An online language translation service is another way to use machine translation. Often, this service is provided free of charge. Among the most popular providers of the free online translation service are Babel Fish, Google Translate, and FreeTranslation.

Translation memory: A software tool that enables professional translators to create and store their translations for later reuse. Translator workstation: A system that enables professional translators to access dictionaries, glossaries, and terminological resources and includes such features as multilingual word processing and translation memories. Software localization tool: A tool to translate and adapt computer software manuals to the cultural context of countries where this software will be used.

Translation support tools

Support tools form another class of language translation software. They include: electronic dictionaries (e.g. Lingvo), translation memories (e.g. Wordfast), translator workstations (e.g. Transit), software localization tools (e.g. SDL Passolo), terminology management systems (e.g. MTX), as well as pre-editing and alignment tools.

Language translation software is still in its infancy. Current machine translation systems still produce output of low quality. Nevertheless, language translation software has been improved over the last decades due to continued research and development efforts and has become an important support tool for communication and translation.

Thus, with the invention of translation memories, the productivity of professional translators increased notably; machine translation systems that are limited to specific domains and use controlled language are helpful for large multinational companies and organizations (such as the European Commission) with large volumes of documentation.

It is unlikely that automatic systems will be able to produce high quality translations in the foreseeable future, but they will improve, their role in the information exchange will increase, and we will see their new practical applications.

Further reading:

Major Translation Tools In Translation Industry

Spoken Language Translation Tool: Speech Recognition and Translation

A list of free translation software

Didn't find what you were looking for? Use this search feature to find it.

Return from Language Translation Software Page
to Home Page

__________________________________________________________________ Website owner: Irina Lychak, self-employed freelance linguist, Russian translator, Ukrainian translator, Kiev (Kyiv), Ukraine