Translation and localization glossary

Terms A to Z

Back-translation

Translating a finished translation back into the source language by a separate translator, so the meaning can be compared with the original. Common for patient material in clinical studies. Medical translation.

CAT tool

Computer-assisted translation software, such as Trados, memoQ or Phrase. Translators work sentence by sentence while the tool protects formatting, reuses translation memory and checks terminology. It is not machine translation. Tools and formats.

Co-edition

A foreign-language edition of a book laid out to match the original page for page, often with all text on the black plate so color images are printed once for every language. For publishers.

Consecutive interpreting

Interpreting where the speaker pauses and the interpreter repeats each part in the other language. Suits meetings, interviews and appointments. Interpreting.

Data annotation

Adding labels to text, audio, images or video so they can be used to train or test a machine learning model, such as tagging names in a sentence or drawing boxes around objects. Data annotation.

DTP (desktop publishing)

Laying out translated text in the original design files, such as InDesign, so the translated document looks like the original. Multilingual DTP.

EPUB 3

The standard file format for eBooks, used by Apple Books, Kobo and most libraries. It can be reflowable or fixed layout and supports accessibility features. eBook conversion.

Glossary (termbase)

An approved list of key terms and their translations in each language, used so every document says the same thing the same way.

Inter-annotator agreement

A measure of how often different annotators give the same label to the same item. Low agreement usually means the guidelines need clarifying. For AI teams.

Localization

Adapting a product or content for users in another language and market. It covers translation, but also date, number and currency formats, images, layout and testing. Website and software localization.

LLM evaluation

Rating a large language model's answers against a rubric, for accuracy, helpfulness, safety and how natural they sound in each language. LLM evaluation.

Machine translation post-editing (MTPE)

A professional translator editing machine-translated text until it meets an agreed quality level. Suitable for some content, not for marketing or anything legally binding.

OCR (optical character recognition)

Converting scanned pages or images into editable text, so they can be translated or searched. OCR and data entry.

Pseudo-localization

Replacing interface text with longer, accented test text before translation, to find layout and encoding problems early. Software translation.

Red teaming

Writing prompts designed to make an AI model fail or produce harmful output, so the problems can be fixed before release. Red teaming.

RLHF

Reinforcement learning from human feedback. People rank or rewrite a model's answers, and those preferences are used to train it. RLHF data.

Right-to-left (RTL)

Scripts written from right to left, such as Arabic, Hebrew, Farsi and Urdu. Layouts are usually mirrored as well. Arabic translation.

SCORM

A standard that lets e-learning courses run in a learning management system and report completion and scores. Versions 1.2 and 2004 are the most common. E-learning localization.

Simultaneous interpreting

Interpreting at the same time as the speaker, usually from a booth or a remote platform, so a meeting or conference runs at normal speed. Interpreting.

Subtitles and captions

Subtitles translate speech for viewers who don't know the language. Captions show speech and sounds in the same language, for viewers who can't hear it. Subtitling.

Transcreation

Adapting marketing copy for another market so it has the same effect, rewriting idioms, wordplay and references where a literal translation would fall flat. Transcreation.

Translation memory (TM)

A database of source sentences and their approved translations. CAT tools use it to reuse previous work, which keeps wording consistent and lowers the cost of repeated text.

Typesetting

Setting text into pages for print or digital publication, with attention to spacing, hyphenation, line breaks and page balance. Typesetting.

xAPI

A newer e-learning standard than SCORM that records learning activity in more detail, including activity outside a learning management system.

XLIFF

An XML file format for exchanging translatable text between tools, used by many CAT tools, authoring tools and software platforms.