ducet

Ducet, short for Default Unicode Collation Element Table, is the baseline data set used by the Unicode Collation Algorithm to define the default ordering of Unicode code points for sorting and comparing text. The table is maintained by the Unicode Consortium and published as part of the Unicode Collation data. It is used by software libraries such as ICU to provide locale-aware collation. Many languages require tailoring, where locale-specific rules adjust the default ordering; such tailorings are built on top of DUCET and may alter primary, secondary, or tertiary weights to reflect local alphabet conventions.

The DUCET assigns to each code point a sequence of weights that determine its place in sort order: the primary weight reflects the base character, the secondary weight diacritics, and the tertiary weight case and other minor differences. For combining marks, contractions, and other special cases, the table contains additional entries to define relative ordering. The result is a language- and script-agnostic baseline against which locale-specific rules are applied.

In practice, DUCET serves as a starting point or reference implementation for Unicode collation. It is widely used in libraries and databases that implement the Unicode Collation Algorithm. Because real-world expectations differ by language, collations are generally tailored to specific locales, so the DUCET alone is rarely the final sort order used in applications.