17.8 16.1.7 Normalización e indexación
| Etapa | ID | Acción |
|---|---|---|
| 1 | NORM-01 | Preserve the exact Unicode source literal. |
| 2 | NORM-02 | Apply Unicode NFKC to the search representation, not to the stored display form. |
| 3 | NORM-03 | Case-fold for retrieval. |
| 4 | NORM-04 | Create an optional accent-insensitive key. |
| 5 | NORM-05 | Normalize whitespace and punctuation for candidate matching. |
| 6 | NORM-06 | Retain token boundaries and original hyphenation as features. |
| 7 | NORM-07 | Detect exact duplicates within the same concept and locale. |
| 8 | NORM-08 | Detect cross-concept homonyms without merging them. |
| 9 | NORM-09 | Resolve locale fallback using BCP 47 matching policy. |
| 10 | NORM-10 | Emit review candidates with explanation and confidence. |
17.8.1 Claves separadas
displayLiteral = "Talón–punta"
canonicalUnicode = NFKC(displayLiteral)
caseFoldKey = "talón–punta"
accentFoldKey = "talon punta"
tokenKey = ["talon", "punta"]La búsqueda puede usar claves flexibles; la edición y exportación siempre conservan la forma original.