Différences

Ci-dessous, les différences entre deux révisions de la page.

--- en:guide_de_parametrage [2025/09/30 16:59] – [Focus on the customizable online journey] admin
+++ en:guide_de_parametrage [2025/11/04 16:59] (Version actuelle) – [Enable User Data Entry] admin
@@ Ligne 286: / Ligne 286: @@
 The entered last name is compared with both the birth name and the usual name.
-🔧 If you still want to perform such checks while redirecting the user to the ID360 identification URL—without letting them modify the expected data—an API method is available to push this data into the system (see: Developer Guide).
+🔧 If you still want to perform such checks while redirecting the user to the ID360 identification URL—without letting them modify the expected data, an API method is available to push this data into the system (see: Developer Guide).
+===== String Comparison in ID360 =====
 🔧 The algorithm used is a fuzzy matching.
+ID360 uses a **three-step approach** to compare strings.
+The goal is to enable reliable comparisons even when strings contain accents, special characters, or linguistic variations.
+==== 1. Unidecode Step ====
+The first step is to **remove accents and diacritical marks** to standardize the comparison.
+This allows words to be compared as if they were written using only standard ASCII characters.
+**Example:**
+  * élodie → elodie
+Thus, *élodie* and *elodie* will be considered **identical** at this stage.
+==== 2. Generic Transliteration Step ====
+Once accents are removed, ID360 applies a **generic transliteration**.
+This process converts non-ASCII characters (such as letters specific to certain languages) into their most common Latin alphabet equivalents.
+The original language of the word is **not** taken into account here — only a standard transliteration is applied.
+**Examples:**
+  * weiß → weiss
+  * düker → duker
+This step makes it possible to harmonize strings originating from different languages without knowing the linguistic context.
+==== 3. Precise Transliteration Step ====
+The final step performs a **contextual transliteration**, taking into account the **original language of the text**.
+For identity documents (passports, national ID cards, etc.), the language used corresponds to the **issuing country of the document**.
+As a result, some letters are transformed differently depending on the language.
+**Examples:**
+  *For Germany (DE)* :
+  * Jörg → Joerg
+  * Düker → Dueker
+  *For France (FR)* :
+  * Jörg → Jorg
+  * Düker → Duker
+This language-aware transliteration provides a **comparison more faithful to the document’s context**.
+==== Notes ====
+  * Specific transliteration rules depend on the **country of origin**.
+  * European countries are generally less affected by these differences, but **Germany** is a notable exception, as its rules are frequently used in official documents.
+  * Other countries may apply their own conventions depending on their civil registry standards or national requirements.
+==== Summary ====
+ID360 compares strings through a progressive process:
+  - **Character normalization** (removal of accents)
+  - **Universal transliteration** (generic conversion of non-ASCII characters)
+  - **Contextual transliteration** (conversion adapted to the document’s language)
+This method ensures a **robust, consistent, and linguistically accurate comparison**.