The i18n glossary, minus the jargon

Fallback Locale

The locale the translation layer uses when a key is missing in the active one, keeping the UI functional while coverage catches up.

Hardcoded String

User-facing text embedded directly in code or templates, invisible to the translation pipeline and stuck in one language.

Locale

A locale identifies a language plus regional conventions, such as en_US or pt_BR, and drives translations, formats, and collation.

Pluralization

Handling grammatical number correctly across languages, whose plural rules range from one form to several distinct categories.

Source Language

The language content is authored in and translated from, serving as the reference version that every target language tracks.

String Extraction

Collecting translatable strings out of source code into translation files or a TMS, by static scanning of calls like __().

String Interpolation & Placeholders

Inserting dynamic values into translations through named placeholders instead of concatenation, which breaks under real grammar.

Target Language

Any language content is translated into. Each target locale carries its own coverage, conventions, difficulty, and review pipeline.

Translation Key

The stable identifier that maps a piece of UI text to its translations, such as dashboard.welcome, decoupling code from copy.

i18n (Internationalization)

Internationalization: engineering software so it can support any language and region without code changes for each new locale.

l10n (Localization)

Localization: adapting an internationalized product to one specific locale, covering translation, formats, imagery, and conventions.

BiDi (Bidirectional Text)

Text mixing right-to-left and left-to-right content, ordered by the Unicode Bidirectional Algorithm and prone to subtle bugs.

CLDR (Common Locale Data Repository)

The Unicode Consortium dataset of locale conventions: plural rules, date patterns, number formats, and translated names.

Character Encoding & UTF-8

How text becomes bytes. UTF-8 encodes all of Unicode, is ASCII-compatible, and is the assumed encoding of the modern web stack.

Date & Time Localization

Rendering dates and times per locale: component order, separators, month names, clock style, week start, and time zone handling.

ICU MessageFormat

A standard syntax for translatable messages with plurals, gender, and typed placeholders, implemented by the ICU library.

ISO 639-1

The standard two-letter language codes, en, es, de, used as the primary subtag in language tags and as Laravel locale names.

Language Tag (BCP 47)

The IETF standard for language tags like en-US or zh-Hant, composing language, script, and region subtags into one identifier.

Number Formatting

Locale-aware rendering of numbers and currency: decimal separators, digit grouping, percent style, and currency placement.

PO File (Portable Object)

The human-readable translation file format of the gettext system, pairing msgid source strings with msgstr translations.

RTL (Right-to-Left)

Support for scripts written right to left, such as Arabic and Hebrew, requiring mirrored layouts and direction-aware CSS.

XLIFF

The XML Localization Interchange File Format, an OASIS standard for passing translatable content between tools and vendors.

gettext

The GNU localization system built around extracting source strings into PO files, a long-standing standard in the C and PHP worlds.

CAT Tool (Computer-Assisted Translation)

The editing environment professional translators work in, combining segment-by-segment editing with translation memory and termbases.

Contextual Translation

Translating with knowledge of where and how a string is used, since short UI strings are ambiguous without their context.

Fuzzy Matching

Matching new source text against similar, not identical, past translations so near-repeats need review instead of retranslation.

Glossary / Termbase

A curated list of approved terminology, product names, and do-not-translate entries that keeps wording consistent across languages.

LQA (Localization Quality Assurance)

Reviewing localized software for language and presentation defects: mistranslations, truncation, broken placeholders, wrong formats.

Localization Style Guide

Per-language writing rules for translators covering tone, formality, casing, units, and how to handle names and placeholders.

Machine Translation (MT)

Software-generated translation, from neural MT engines to LLMs, typically paired with human review for anything user-facing.

Pseudo-Localization

A testing technique that replaces text with accented, expanded fake translations to expose i18n bugs before real translation starts.

TMS (Translation Management System)

A platform that centralizes translation work: storing strings, tracking coverage, coordinating translators, and syncing with code.

Transcreation

Creative adaptation rather than direct translation, used for slogans and marketing copy where effect matters more than wording.

Translation Coverage

The share of translation keys that have a value in each target locale, the primary health metric of a localization pipeline.

Translation Memory (TM)

A database of previously translated segments that suggests or reuses past translations, cutting cost and keeping wording consistent.

Word Count (Localization)

The source-text volume measure used to scope and price translation work, often weighted by translation memory match rates.

CI/CD Localization

Wiring translation checks and syncs into the delivery pipeline, so missing keys and broken placeholders fail builds, not users.

Continuous Localization

Running translation as an always-on process alongside development, instead of a batch phase at the end of each release cycle.

Lazy Loading Translations

Loading only the translation groups or locales a request actually needs, instead of shipping every string to every user.

Over-the-Air (OTA) Translations

Delivering translation updates to a running application at runtime, so fixing copy never requires a deploy, build, or release.