Translation & Localization Glossary
Plain-English definitions of the terms you will encounter when translating a website or internationalizing a product. Each entry explains both the concept and why it matters in practice.
i18n (Internationalization)
The practice of designing software so it can be adapted to different languages and regions without requiring code changes. The abbreviation comes from the 18 letters between the 'i' and 'n' in 'internationalization'. A properly internationalized app externalizes all user-facing text, formats dates and numbers according to locale rules, and avoids hardcoded assumptions about text direction or string length.
See: React i18n approaches →l10n (Localization)
The process of actually adapting a product for a specific locale — translating text, adjusting date and number formats, switching to local currency, and handling RTL layout if needed. The '10' stands for the 10 letters between 'l' and 'n' in 'localization'. Where i18n is the engineering foundation, l10n is the content and cultural work built on top of it.
See: Localization cost guide →Locale
A combination of language and region code, such as fr-FR (French as spoken in France) or fr-CA (French as spoken in Canada). Locales matter because the same language varies significantly by region — vocabulary, spelling, currency, date formats, and even punctuation conventions can differ. The format follows the IETF BCP 47 standard (language-REGION).
See: Multilingual SEO guide →hreflang
An HTML link attribute that tells search engines which language version of a page to show to users in each locale. It looks like <link rel="alternate" hreflang="fr" href="https://example.com/fr/page">. Without hreflang, a search engine may serve the wrong language version to users — or consolidate all versions into one, destroying multilingual SEO. It is one of the most important technical requirements for international sites.
See: hreflang implementation guide →RTL (Right-to-Left)
The text direction used by Arabic, Hebrew, Persian, and Urdu. RTL layouts flip the entire page: text aligns to the right, margins and padding are mirrored, flex direction reverses, and UI elements like back-buttons and breadcrumbs swap sides. Correctly supporting RTL requires CSS logical properties (margin-inline-start instead of margin-left) rather than physical ones.
See: RTL implementation guide →Translation Memory (TM)
A database of previously translated sentence pairs. When a segment of text that was translated before appears again — exactly or approximately — the TM suggests the existing translation. This ensures consistency across a product, prevents the same sentence from being translated differently in two places, and reduces cost and time because matched segments do not need re-translation.
See: Lingvit features →Machine Translation (MT)
Translation performed automatically by software rather than a human. Output quality varies by provider, language pair, domain, and source text, so important product, legal, and commerce copy should be reviewed before publishing.
See: How to translate a website →Neural Machine Translation (NMT)
A subset of machine translation that uses deep learning neural networks to model language. NMT dramatically outperforms older statistical approaches by understanding context across a full sentence rather than translating word-by-word. It handles fluency, idioms, and ambiguity far better, and is the current standard for production translation systems.
See: Translation approaches compared →CAT Tool (Computer-Assisted Translation)
Software that helps human translators work faster by combining translation memories, terminology glossaries, and project management into a single interface. The translator sees the source text alongside their translation and receives suggestions from the TM. Common CAT tools include OmegaT (open source), SDL Trados, and Phrase. CAT tools complement rather than replace human judgment.
See: Translation tool comparisons →Glossary (in localization)
A curated list of terms paired with their approved translations. A glossary ensures that brand names, product names, legal terms, and technical vocabulary are translated consistently — or intentionally left untranslated. Without one, different translators or MT engines may render the same term in five different ways across a product.
See: Lingvit glossary feature →Source Language
The original language of the content being translated. For most international products, the source language is English, since English-first development is the default in global software. The source language is the authoritative version — all translations derive from it and return to it when content is updated.
See: All supported languages →Target Language
The language that content is being translated into. A single localization project typically has one source language and multiple target languages. Target language selection is driven by market data — which languages your users speak, where your product is growing, and which regions have regulatory or support requirements for local-language content.
See: All supported languages →String
In software localization, a string is any piece of user-facing text: a button label, an error message, a tooltip, a navigation item, a form placeholder, or a page title. The localization workflow extracts strings from code, sends them for translation, and re-inserts the translated versions. String management is the core challenge of i18n — keeping translations in sync as the product changes.
See: Documentation →Pseudo-localization
A testing technique that replaces source-language characters with visually similar accented or expanded versions before real translation begins. For example, 'Hello' might become 'Ĥéļļö Ŵörļd' — still readable in English but visually distinct. This reveals UI layout bugs caused by text expansion, missing font support, or hardcoded element widths, without waiting for real translations.
See: When i18n libraries make sense →MutationObserver
A browser API that watches for changes to the DOM and fires a callback whenever nodes are added, removed, or modified. Translation widgets use MutationObserver to detect and translate text that is injected into the page after initial load — essential for single-page apps and any content loaded dynamically via JavaScript. Without it, a widget would only translate the initial HTML and miss all subsequent content.
See: React & SPA integration guide →FIGS
An abbreviation for French, Italian, German, and Spanish — the four major Western European languages. FIGS is shorthand in the localization industry for the first expansion set most global software products target after English. It covers four large, high-GDP markets and shares enough linguistic infrastructure (Latin alphabet, similar text expansion ratios) to be tackled together.
See: All supported languages →Related guides
Keep reading
- All 30+ supported languages — explore which locales Lingvit covers, including RTL languages.
- How to translate a website — a practical step-by-step guide applying many of the terms above.