Glossary of Translation Terms, Part 2
Welcome to the second part of our glossary of the most common translation terms. In the previous part, we looked at general terms that come in handy when ordering a translation, and today we will focus on more specific terms from the field of translation technologies.
Translation memory (TM)
A database that stores sentences, paragraphs, or text segments that have been translated in the past. Each translation memory entry consists of the original text and its translation. Translation memories are used in translation programs such as memoQ. They automatically suggest sentences, paragraphs, or segments from previous translations to the translator when they are translating a new text. Thus, the translator does not have to re-translate text that has already been translated.
A database containing terminology. Each entry in the terminology database consists of a term and its translation, but can also contain other data, such as a definition, examples of use, grammatical data (verb, noun, etc.), and others. Terminology databases are used in translation programs such as memoQ.
Quality Assurance (QA)
The key process for ensuring that a text is flawless. Every translation should go through this process before submission. It is usually performed with translation tools that check if there are any machine-detectable errors in the translation. The QA tool compares the translation with the original and automatically detects errors such as double spaces, mismatching numbers, inconsistently translated segments, etc.
A tool that is part of a translation program (such as memoQ) that enables us to find specific words or phrases in the active translation memory, i.e. in texts that have already been translated. Words and phrases are displayed in context and help the translator during the translation process.
When the text is imported into a translation tool (e.g. memoQ), it is broken down into individual segments according to the defined segmentation rules. A segment usually equals a sentence, but it is also possible to specify different segmentation rules.
A bitext format used for the exchange of data between CAT tools.
When a text gets imported into a CAT tool, the tool compares how much of the text completely (full match) or partially (fuzzy match) matches within the text to be translated or with the translation memory. Thanks to the match analysis, it is easier to plan how much time the translator will spend translating, as words that fully or partially match can be translated faster than new words.
The process by means of which we can create a translation memory from existing texts (originals and translations). Alignment is performed in a translation tool that divides the text into segments and finds matches between the segments of the original and the translation. Matching segments are then exported to the translation memory and can be of great help during translation.