Unicode Combining Classes

The Canonical Combining Class (ccc) is a numeric property assigned to Unicode characters—primarily combining marks—that tells text-processing and rendering systems how those marks should be ordered when multiple marks are applied to the same base character. When a sequence such as a letter plus several accents is encoded, the marks might not initially appear in a consistent order. The ccc provides a standard way to reorder them into a canonical order, ensuring that equivalent sequences are treated the same way internally and displayed correctly. This is especially important for Unicode normalization, where text is converted into a consistent form for comparison, searching, and storage.

A combining class of 0 is assigned to characters that do not participate in reordering. These include most base characters (letters, digits, symbols) and some combining marks that act as starters in a sequence. Marks with nonzero values are reordered relative to one another based on their ccc values: lower numbers come first, and higher numbers follow. This does not change the logical meaning of the text—it simply standardizes the sequence. For example, if two accents are applied to the same letter, their order in the encoded string may vary depending on input method, but normalization will reorder them according to their combining classes so that equivalent text has a consistent binary representation.

The numeric values of the ccc also loosely correspond to the typical visual placement of marks. For instance, certain classes are associated with marks that appear below a base character, others above, and some that overlay or attach in more complex positions. While rendering engines ultimately decide exact placement using font data and layout rules, the combining class ensures that marks are processed in a stable and predictable order before rendering occurs. Without this system, sequences of multiple combining marks could produce inconsistent results across platforms or fail to match during text comparisons, even if they look identical to the user.