CodePark{gender, select,
female {{name} shared {count, plural,
=0 {no photos} one {# photo} other {# photos}} with her team}
male {{name} shared {count, plural,
=0 {no photos} one {# photo} other {# photos}} with his team}
other {{name} shared {count, plural,
=0 {no photos} one {# photo} other {# photos}} with their team}
}
ICU MessageFormat is a standard syntax for translatable messages. One string holds the whole sentence, with placeholders and rules for every variant. Libraries like react-intl (FormatJS) and many translation tools read it.
The parts in the example:
{name}is a simple placeholder for a value.{gender, select, female {…} male {…} other {…}}picks a branch by a value.otheris the required fallback.{count, plural, …}picks a branch by the language's plural category.=0matches the exact number 0, before the plural categories are checked.#is replaced by the number, formatted for the locale.
Why it is better than building the sentence in code:
- The translator sees the full sentence. Word order changes by language, and the translator can move
{name}anywhere. - Each language adds the plural branches it needs, such as
fewandmanyfor Polish, without code changes. - Gender and grammar agreement stay inside the translation, where they belong.
- The code passes only data:
{ name, gender, count }.
One tip from the example: put whole sentences inside each branch, not small pieces. That gives translators full sentences to work with.
- ICU MessageFormat keeps a whole sentence and all its variants in one string.
{name}is a placeholder,selectpicks a branch by value andpluralpicks by plural category.=0matches an exact number, and#shows the formatted number.- Translators can reorder placeholders and add the plural forms their language needs.
- The code only passes data, so no sentence is built by concatenation.