The contemporary users of machine translation can be divided into two groups: individual unprofessional users who use online translation services to acquire a translation of different texts for their own use and then professional users who employ machine translation systems often incorporated in some kind of computer-aided (assisted) translation tool to deliver professionally translated texts or documents to their customers. Such a use of MT requires a human input known as post-editing of pre-translated text.
2.3 Early machine translation systems
Information regarding the machine translation systems and translation tools (Apertium, Česílko, MemSource, and SDL Trados Studio) were acquired from their respective official web pages. The resource of information about EUROTRA was the report “Anatomy of Eurotra: a multilingual machine translation system” written by Anne De Roeck in 1981 and the official web page of the European Union www.europa.eu.
Apertium is an open-source machine translation platform used for translations from one language to another. It uses dictionaries and grammar rules to deliver consistent and comprehensible sentences in target languages. The Czech language is currently not among the released language pairs (April 2016), however, the pairs Czech–Polish and Czech–Serbian are stated to be in the phase of preparation. (www.apertium.org)
Česílko is a project of the Institute of Formal and Applied Linguistics on Faculty of Mathematics and Physics of Charles University in Prague and has been developed since 1998. It was originally designed as a tool for fast and efficient translations from a source language into target languages (and it should work with mutually related languages). The system currently contains five language pairs (Czech–Polish, Czech–Lithuanian, Czech–Macedonian, Czech–Lower Sorbian, and Czech–Slovak) with only one working language pair Czech into Slovak. (https://lindat.mff.cuni.cz/services/cesilko/demo.php)
EUROTRA (1978–1992) was an ambitious research and development project founded and funded by the European Commission that wanted to build on the successes of the older machine translation system SYSTRAN (which was a commercial project). The aim of EUROTRA project was to create a machine translation system that could translate from and into the growing number (six at the time of the foundation: German, Danish, English, French, Italian, and Dutch) of official languages of the European Union. One of the founding principles of the European Union is that every EU citizen has the right to use any of the official languages in correspondence with EU institutions and has the right to read the published texts in any of the official languages. Also the webpages of the EU are available in all the official languages (24 languages currently) and all the general information regarding the EU policies is provided in all the official languages. This all presents an enormously extensive amount of written texts and documents that require translation into 23 other languages.
2.4. Translation tools that make use of machine translation
David Čaněk (the founder of MemSource Technologies) says in the interview with Radek Pavlovič for the journal Euro that the European Union wasted its opportunity to become a leader in the field of translation technologies. It had all the required prerequisites: a multinational and multilingual environment, large numbers of linguists concentrated in a “small” space, a sufficient amount of funds, and sufficient volumes of texts that needed (and still need) to be translated. Instead, the leaders of IT industry from the United States (Google and Microsoft) drew on their technologies, expertise, and thousands of customers and millions of users.
The use of machine translation today is applicable particularly to technical and scientific texts as the ratio of repetitions and highly similar texts in this field is higher than in any other field of translation. Arturo Trujillo states in his book Translation Engines: Techniques for Machine Translation that:
Although MT systems are useful when translating texts with a high degree of repetition, they are especially applicable when translating new versions of a document that has already been translated. A common feature of such documents (e.g. computer manuals, budgets, financial reports) is that some sentences differ from others only in elements that can be easily copied or adapted to the TL text. (68)
MemSource Technologies is a software company founded in 2010 by David Čaněk in Prague, Czech Republic.
MemSource is a world-widely used computer-aided translation tool at the moment. It stores all resources, projects, translatable files, glossaries or term bases, and translation memories in the cloud. It includes a complete translation environment with an editor (options to use a web editor or desktop MemSource editor which can be downloaded for free and has to be installed in a translator’s computer) and a cloud. The incorporated user interface is well arranged and simple to use, text is separated into segments which are easy to work with. MemSource provides the possibilities of using translation memories, term bases, and also machine translation. The translated file is located in the translator’s computer, all other sources are in MemSource cloud and the editor connects to them. (There is a web-based editor as well.)
The complete workflow of a translation project can be done in MemSource, it provides an environment for managing the project (entering the order into system, assigning to a translator, translation by a professional translator, review, analyses, sending the files back to a customer, closing the order), translated files, vendor databases, translation memories, and glossaries. The editors include functions for spelling check, and also the possibility to preview the translated file. MemSource does not offer its own MT system but allows translators to connect to 3rd-party MT systems (for example Google Translate, Microsoft Translator, Systran, Microsoft Translator Hub, Apertium, Asia Online, SDL BeGlobal, and SDL Language Cloud, to name just a few).
Many translation agencies and localization companies employ this presently popular translation environment. MemSource Technologies states that approximately 200 million words are translated (human-translated) every month in MemSource by professional translators (April 2016).
SDL Trados Studio
SDL Trados Studio is another computer-aided tool that offers the use of machine translation. SDL uses rule-based machine translation system. This system combines a set of grammatical rules and a dictionary for each language. The dictionary is filled with detailed information regarding the source and target words and their grammar. Translators owning a license for SDL Trados Studio have the option to fill the empty segments (not containing matches from a translation memory) by automated machine translation of the source text.
SDL Trados Studio provides SDL Language Cloud and SDL BeGlobal MT services and also the option to connect to 3rd-party MT systems, such as Google Translate.
Translation environment tools
Jost Oliver Zetzsche, a localization and translation consultant and certified English-to-German translator, uses a new name for this kind of translation tools – Translation environment tools (TEnT). Such tools provide more functions than simple translation tools. In his article “What Makes a Translation Environment Tool a Good TEnT?” on the web www.translatorscafe.com Zetzsche describes and lists the features of such a tool:
a seamless input, a working spell-checker, grammar checks, autotext and autocorrect functions, tracking changes, comments, inline codes (markers that remember formatting), smart quotes of particular language, non-printing characters (spaces, tabs, line breaks), and WYSIWYG (what-you-see-is-what-you-get – the ability to see translation in its context and layout).