glossary of terms and acronyms used in the translation and linguistic quality control industry
A term used to describe the extent to which the quantity and quality of information present in the source text are maintained in the target text.
The task of modifying a text to make it suitable for a different language variant, region or country. In assessments, adaptation is an intentional deviation from the source version to conform to local usage or context, which needs to be carefully assessed in view of the primary goal of ensuring comparability across language versions.
Process that takes as input two language versions of one text (source and target text) and produces a bilingual file or asset where each source segment is associated with its translation. Typically the aim of this process is to produce a TM.
A literal translation of a translation. Helps reviewers who do not speak the target language to determine whether the meaning of the translation is accurate.
A word used to describe text in which there is both right-to-left and left-to-right content, e.g. text in Arabic script that contains mathematical formulas.
A term used to describe the extent to which a set of explicit standards, rules or criteria are met.
computer-assisted translation (CAT) tool
A specialized program that helps a translator to increase productivity and obtain a higher degree of consistency and quality, by leveraging on special translation-boosting functionalities and language assets such as translation memories and termbases.
Style restricted by rules, such as approved terminology, restricted vocabulary or simplified grammar. Its purpose is to create a source text that is easy to translate.
A translation that has not yet received the final approval by the translator or that has not yet been checked by the reviser.
Activity performed by the translator or another expert that involves making appropriate amendments to fix issues that were previously overlooked by the author or the translator. If performed by the translator after completing the draft translation and before delivering it, it may refer to two differentiated tasks: revision and review.
A feature of a text that requires correction.
A TM match that is identical to the new source segment to be translated.
A term used to describe the extent to which the (target) text reads naturally in the target language. Ideally, it should not be noticeable that a translated has not initially been written in the target language. When this is not the case, at cApStAn we say that the text has a translanese flavour.
A TM match that is not identical to the new source segment to be translated but is sufficiently similar so that it is potentially helpful in translating the new segment
Important language asset containing terminology (terms and expressions in the source language and their target equivalent(s) in one or more target languages). The glossary can be imported into a termbase to be integrated in the work carried out in a CAT tool. It is used to ensure consistent and correct terminology usage, both by allowing the translator to insert the suitable term and by feeding term consistency QA checks.
The activity of directing the development of software or multimedia products taking into account its multilingual nature to allow localization of its content without affecting the source code. For example, internationalization entails enabling the easy extraction of translatable text to external resource files, correct encoding of the different target versions, optimizing the source material, etc.
At cApStAn we use the terms ‘internationalization’ and ‘localization’ in reference also to survey and assessment instruments.
A feature of a text that requires some kind of action.
Resource containing data that can be used for reference or leverage during a localization project, e.g. translation memory, termbase, style guide, segmentation rules, etc.
The pair of languages comprising the source language and the target language between which a translator works.
Proofreading of a text without comparison of target segments with source segments. At cApStAn, we use “proof-reading” to describe this task.
The application of knowledge about language to the development of language technology solutions.
Professional that provides linguistic services, normally taking care of one or several language tasks in a localization project. Also called linguist.
One of the several tasks a localization project consists of, e.g. translation, reconciliation, adaptation, verification, review, etc.
Direction of a language that is written and read starting on the left and following towards the right.
Practice of reusing previously translated segments stored in a TM in the execution of new translations. Also, a measure of the extent to which the workload of a new translation can be reduced by using a TM.
See language expert.
linguistic quality assurance (LQA)
Set of processes designed to ensure that multiple language versions of a document meet standards of linguistic, cultural and functional equivalence.
linguistic quality control (LQC)
Set of processes designed to verify whether the pre-defined quality standards are met; and to propose corrective action if they are not.
The process of translating a software user interface or other multimedia content into a language variant and adapting it to a specific culture/region or country conventions.
At cApStAn we use the term ‘localization’ in reference also to survey and assessment instruments.
Set of technical activities that support the localization process, including internationalization, file preparation, testing, resizing, compilation, and other software-related tasks.
machine translation (MT)
Automatic translation performed by a software program.
Originally, checking a text before publication to spot any typesetting errors. In the translation industry, often used as synonym of revision. At cApStAn, final proof-reading is a revision of the target version’s linguistic correctness, without comparison to the source.
Manual or automated checks aimed to spot any errors that the translator or any other language expert might have overlooked, normally related to accuracy, compliance or language criteria.
The extent to which a text or a translation satisfies stated and/or implied needs or expectations. At cApStAn we prefer to refer to ‘fitness for purpose’, a more concrete concept.
The task of merging two independent translations to obtain a high quality ‘reconciled’ version. Typically, the reconciler takes the best elements from each translation rather than selecting one of the two translations.
The process of re-reading a draft translation to spot language or fluency issues and apply appropriate amendments.
Reading a text to identify errors, inconsistencies, incorrect grammar and punctuation, poor or inappropriate style, and, in the case of a translation, conformance with the source text, and making appropriate changes and corrections to the text. In general, the number of revision stages is proportional to the demands on the text quality: a translation intended for publication may, for example, be revised by the translator and by one or two third parties (e.g. the author, a subject expert, a second translator, an editor), whereas an internal memo may not require any revision after translation. (What exactly revising and editing entail and how they differ is the subject of much debate. What is important is that the person commissioning the work communicates clearly what is expected of the editor.)
Direction of a language that is written and read starting on the right and following towards the left.
One of the many fragments into which the whole of a text is split for easier handling, corresponding to the selected unit of translation: either the sentence or the paragraph/cell.
The task of splitting a text into segments. It is an important task because leverage and productivity depend on it.
Pattern-based rules used to define how the source text must be segmented when preparing files for translation. These rules are used to achieve segmentation beyond the paragraph marks, normally defining what is a boundary between two sentences and what is not (e.g. after an abbreviation).
Segmentation Rules eXchange (SRX)
XML-based standard format used to define and exchange segmentation rules between different CAT tools.
Language in which a text is originally written, or the language of the source text in a translation process.
The text to be translated.
The task of customizing part of an assignment (or of a file, often the source text) to make it more suitable for translation into a given target language or family of target languages. This can affect features of the source text or translation notes.
Language in which the target text is produced.
target text The translation, i.e. the result of the translation process.
Process of extracting terminological units from a source text and, optionally, their target equivalents from a text in the target language. The aim of the extraction is to build a glossary or termbase to enhance the translation process or implement QA checks.
A terminological database that is integrated in a CAT tool, typically allowing the translator to easily check the target version of a source term and often to insert the target term quickly, without having to type. Also called glossary or dictionary.
A source segment stored in a TM that has a degree of similarity higher than a pre-defined threshold (e.g. 75%) with a new source segment to be translated; the stored translation of the matching segment is proposed as a suggestion to translate the new source segment.
Advance translation of a draft source text by a panel of linguists representing the target language families, performed with a view to identifying potential translation or adaptation issues. The feedback from these linguists is collated to produce a translatability report, which may propose alternative wording to circumvent problems or translation/adaptation notes to assist in solving them.
translation memory (TM)
Database or file containing translation units, which is used to provide leverage in new translation projects.
Translation Memory eXchange (TMX)
XML-based open standard format designed to allow easy exchange of translation memories between localization tools or translation vendors with no critical loss of data.
Segment pair consisting of the source segment and the target segment.
Translation, Review, Adjudication, Prestesting, Documentation. A translation and adaptation model designed by Harkness et al, used in survey translation to produce high quality team translations.
Character encoding standard which, unlike ASCII, uses 8 or 16 bit character encoding, making possible the representation of any character sets (e.g. Latin, Cyrillic, Arabic, Japanese, Chinese).
Segment by segment comparison of a target text with a source text to (i) identify and document deviations; (ii) check compliance with a set of guidelines; and (iii) suggest corrective action as needed. At cApStAn, verification is performed by a linguist trained to document edits using cApStAn’s set of standardized verifier intervention categories.
The number of words in a text, used as a measure to calculate costs and timelines.
Software program that can be used to translate or edit XLIFF files, displaying only the translatable content.
XML Localization Interchange File Format (XLIFF)
XML-based open standard for the exchange of localization data, including elements and attributes for alternative translations, comments, segment states, etc.