ICU MessageFormat

Handles the complex plural rules of different languages using CLDR data. The plural argument type selects the correct message variant based on a numeric value and its plural category (e.g., one, few, many, other).

• Example: {count, plural, =0 {No items} one {# item} other {# items}}
• The # token is automatically replaced with the numeric value, formatted with locale-appropriate number formatting.
• Ordinal selection (selectordinal) handles categories like one, two, few for ranking strings (e.g., "1st", "2nd", "3rd").

The select argument type enables gender inflection and other arbitrary keyword-based selections. It matches a provided string value against a set of explicit cases.

• Example: {gender, select, male {He} female {She} other {They}} will reply soon.
• Unlike plurals, select does not interpret the value; it performs a strict string match against the provided keys.
• This mechanism is essential for languages with grammatical gender that affects surrounding adjectives and verbs.

Provides built-in argument types for locale-aware formatting of dates, times, and numbers without requiring pre-formatted strings.

• Numbers: {price, number, ::currency/USD} formats a value as currency with the correct symbol and decimal separator.
• Dates: {dueDate, date, ::yyyyMMMMdd} outputs a date using locale-specific month names.
• Duration: {duration, duration, ::h:mm} formats a time interval.
• These formatters use skeletons (like ::currency/USD) for precise, locale-independent control over the output pattern.

MessageFormat supports arbitrary nesting of argument types, enabling the resolution of highly complex grammatical dependencies within a single translatable string.

• A plural rule can contain a select argument, which can itself contain another plural rule.
• Example: {count, plural, one {{gender, select, male {He has one item} female {She has one item} other {They have one item}}} other {They have # items}}
• This composability allows linguists to model the full complexity of grammatical agreement without requiring developers to concatenate fragmented strings.

A robust escaping syntax ensures that special characters used by the MessageFormat parser can appear as literal text in the output.

• Single quotes ' denote literal text: This '{curly}' is literal outputs This {curly} is literal.
• A single quote itself is escaped by doubling it: ''.
• Within a quoted sequence, all characters—including braces { } and hash symbols #—are treated as plain text and not as syntax tokens.
• This is critical for including apostrophes in languages like French and English without breaking the parser.

The ICU standard allows implementations to define custom formatter functions beyond the built-in number, date, and time types.

• A custom formatter is invoked using the syntax: {value, formatterName, arg1, arg2}.
• This enables domain-specific formatting like ordinal numbers ({pos, ordinal}) or spellout ({num, spellout}).
• Implementations like intl-messageformat in JavaScript provide hooks for registering user-defined formatters, making the syntax extensible for proprietary formatting logic.

The software engineering discipline of designing codebases to be locale-independent, making ICU MessageFormat implementation possible. i18n involves:

• String externalization: Moving all user-facing text into resource files
• Locale-agnostic logic: Ensuring no hardcoded assumptions about date formats, number separators, or text direction
• Plural rule handling: Architecting data models to support languages with multiple plural forms (e.g., Arabic has six) ICU MessageFormat is the runtime expression of a well-executed i18n strategy.

The programmatic process of presenting data according to regional conventions, which ICU MessageFormat orchestrates through its select, plural, and selectordinal argument types. This encompasses:

• Number formatting: {price, number, ::currency/USD} renders $1,234.56 or 1.234,56 € based on locale
• Date formatting: {dueDate, date, ::full} produces "Monday, January 15, 2024" or "lundi 15 janvier 2024"
• List formatting: Handling Oxford commas and conjunction rules per language MessageFormat unifies these concerns into a single, parsable syntax.

A centralized platform that automates translation workflows and manages linguistic assets. TMS integration with ICU MessageFormat is critical because:

• Translators encounter structured placeholders like {companyName} that must remain untouched
• Plural forms require translators to provide all variant strings (one, few, many, other)
• Gender inflection demands context-aware translation for {gender, select, male {he} female {she}} Modern TMS platforms parse MessageFormat syntax to present translators with guided, segmented views that prevent syntax corruption.

The ability to correctly display mixed-direction scripts within the same paragraph—essential when ICU MessageFormat strings combine Arabic/Hebrew (RTL) with English placeholders (LTR). Challenges include:

• Embedded variables: {userName} must render in correct visual order when surrounded by RTL text
• Neutral characters: Punctuation and numbers require explicit directional markers
• Mirroring: Parentheses and brackets must flip orientation in RTL contexts ICU MessageFormat itself is direction-agnostic, but the rendering engine consuming its output must implement the Unicode Bidirectional Algorithm (UBA) to avoid garbled text.

A pre-translation testing technique that simulates localization by transforming source strings with:

• Character expansion: Adding tildes and accents (e.g., "Hello" → "[!!! Ĥéļļö !!!]")
• Length inflation: Padding strings by 30-40% to simulate German or Finnish text expansion
• Delimiter wrapping: Adding brackets to detect string concatenation bugs Pseudolocalization stress-tests ICU MessageFormat implementations by verifying that placeholder integrity survives transformation and that UI layouts accommodate expanded plural variants without truncation or overflow.