Accent Marks

Sources and methodology

The reference tools draw on shared character records and separately verified typing guidance. This page explains what the sources establish, how the data is prepared, and where the limits of a claim matter.

Unicode character data

The generated character dataset is pinned to Unicode 17.0.0. Character identities, official names, general categories, and direct canonical decompositions come from the Unicode Character Database. Normalization and property data provide separate checks on the relationships displayed by the site.

The current collection includes uppercase and lowercase precomposed Latin letters whose full canonical decomposition starts with one ASCII Latin base letter followed by combining diacritic marks. A similar-looking letter is not enough: compatibility-only forms and letters without that relationship are excluded. This scope determines which letter pages exist.

HTML character references

Decimal and hexadecimal HTML references are derived from verified Unicode scalar values. A named reference is stored only when the HTML Standard's entity data maps its semicolon-terminated name exactly to the character. Names are never guessed from a Unicode character name.

When several names map to one character, generation preserves an existing valid selection. Otherwise it selects the shortest name and uses lexical order to break ties. This is a consistent display choice, not a claim that one alias is the official or canonical name.

Normalization and decomposition

Direct canonical decomposition is the mapping recorded for a character in UnicodeData. One of its components may itself decompose further. Full NFD follows those canonical decompositions and the prescribed ordering of combining marks, so it can contain more code points than the direct mapping.

The generator reads NFC and NFD results from Unicode's normalization test data and cross-checks NFD against recursive canonical decomposition. The reference tables retain direct decomposition, NFC, and NFD as separate fields. None of these sequences is presented as a keyboard instruction.

Diacritic relationships

A character is associated with every combining diacritic mark in its verified NFD sequence. The generator checks mark categories against UnicodeData and the Diacritic property against PropList. Multiple marks remain separate relationships, even when they belong to one precomposed character.

On diacritic pages, a dotted circle helps show where a combining mark appears. The circle, U+25CC, is display scaffolding only. It is not stored as part of the combining mark's value or included in the precomposed character copied by a button.

Typing-method verification

Typing claims use official platform documentation first, official application documentation for application behavior, and standards or maintainer documentation where appropriate. Each published guide identifies the environment supported by its evidence.

  • Word Alt+X is an application-specific Unicode conversion method, not a Windows legacy Alt code.
  • Linux guidance is scoped to the documented GNOME and GTK input environment, not every Linux application.
  • Mobile behavior is scoped to the documented keyboard. The current iPhone and iPad guide covers Apple's onscreen keyboard.
  • A character's code point does not establish a typing method. Unresolved shortcut claims are omitted.

The following references illustrate this evidence hierarchy. The typing and Unicode guides provide each guide's full source list, instructions, and limitations.

Generated and manually curated data

A developer-run generator retrieves the pinned Unicode files and HTML entity data, validates the complete candidate set, and writes typed character and diacritic records. Its check mode compares regenerated output with the stored files without changing them. The site build uses those stored records; visiting a reference page does not run the generator.

Platform and application guide evidence is manually curated separately. The Unicode generator does not scrape platform support sites or generate shortcut claims. Guide examples reuse verified character records while keeping input instructions tied to their own sources.

Corrections and uncertainty

Uncertain claims are omitted rather than filled from memory. Source conflicts should be resolved before publication where possible; generation stops on detected data conflicts instead of silently replacing conflicting facts. Technical changes are checked against primary sources.

Verification dates identify when a reference was checked, not a guarantee that an external page will never change. Consult a guide's scope and linked documentation when applying it to a different environment. Citing a publisher does not imply endorsement or affiliation with AccentMarks.org.

To report a factual error, use the Contact page and include the affected page and a supporting source when available.