t13n v0.3.1

T13N: Transliteration Made Right

Welcome the the T13N - a transliteration library for Cyrilic languages written for Javascript applications.

What is it for?

Transliteration can be made for different purposes, such as:

• Friendly URLs for Content Management Systems that publish content using Cyrilic languages;
• Longer SMS;
• Passport names;
• Text interpretations for languages that has a dedicated alternative latin alphabets etc.

Status

IN DEVELOPMENT / PHASE 0 PREVIEW

Implementation Checklist

Implementation is split on so-called "phases" for better prioritization.

Phase Zero: Belarusian-To-Latin (BGN/PCGN 1979)

• Define a basic set of rules for each letter;
• Define a set of flags calculated for each letter for better context;g
• Define alternative variations for some letters that require it (like 'г');
• Support the most basic in-between-words separators (dash, underscore) for URL creation support and resolve "similar" symbols ("’" into "'");
• Ignore already available latin symbols and digits;
• Extend configurations via settings;
• Pack everything as v0.1

Phase One: Other Belarusian-To-Latin variations ("Latinka", ICAO, ISO 9)

• Switch to Typescript;
• Schematize a language JSON;
• Reorganize code to support other variations of one language;
• Add Belarusian Latin alphabet ("Łacinka");
• Add ICAO standard;
• Add ISO 9 standard;

Phase Two: Ukrainian-To-Latin

• Reorganize code to support multiple languages;
• Add Ukrainian alphabet and transliteration rules;

Phase Three: Russian-To-Latin

• Add Russian alphabet and transliteration rules.

(Other languages to be supported later on)

Ruleset & Dictionary

Every transformation rule is explicit and described in a so-called Ruleset It's a compilation of rule that explains transliteration behavior of the script. It may be compact and descriptive at the same time, depending on needs.

A result of Ruleset compilation is a Dictionary, that's used for pre-processing analysis and later transliteration.

There are three types of Rules which can possibly be used: