Supported languages

randino generates in nine languages, each with its own pools, its own name order and its own romanization. Every generator takes a language, and mixes all of the ones it supports when you leave it out.

Code	Language	Native	Person names	Words and nicknames	Sentences
en	English	English	✅	✅	✅
ko	Korean	한국어	✅	✅	✅
ja	Japanese	日本語	✅	✅	✅
zh	Chinese	中文	✅	✅	✅
it	Italian	Italiano	✅	✅	✅
de	German	Deutsch	✅	✅	✅
ru	Russian	Русский	✅	✅	✅
es	Spanish	Español	✅	✅	✅
vi	Vietnamese	Tiếng Việt	✅	✅	✅

The codes are string literals, and they are also available at runtime as NAME_LANGUAGES and WORD_LANGUAGES.

import { NAME_LANGUAGES, WORD_LANGUAGES } from 'randino';

NAME_LANGUAGES; // ['en', 'ko', 'ja', 'zh', 'it', 'de', 'ru', 'es', 'vi']
WORD_LANGUAGES; // ['en', 'ko', 'ja', 'zh', 'vi', 'es', 'it', 'de', 'ru']

The codes are the members of two enums, and the lists are also available as nameLanguages and wordLanguages.

import 'package:randino/randino.dart';

NameLanguage.ko.name; // 'ko'
nameLanguages; // every NameLanguage, in presentation order
wordLanguages; // every WordLanguage, in presentation order

The codes are Literal types, so a checker rejects a code that does not exist, and the tuples are also available at runtime as NAME_LANGUAGES and WORD_LANGUAGES.

from randino import NAME_LANGUAGES, WORD_LANGUAGES

NAME_LANGUAGES  # ('en', 'ko', 'ja', 'zh', 'it', 'de', 'ru', 'es', 'vi')
WORD_LANGUAGES  # ('en', 'ko', 'ja', 'zh', 'vi', 'es', 'it', 'de', 'ru')

Person names

Each language decides three things: which part of a name comes first, whether there is a middle part at all, and how the native form becomes an English pronunciation.

Code	Name order	Middle name	Example
en	given first	yes	Paisley Lewis
ko	family first	no	김태윤 → Kim Taeyun
ja	family first	no	山口直人 → Yamaguchi Naoto
zh	family first	no	赵勇轩 → Zhao Yongxuan
it	given first	yes	Giorgia Mancini
de	given first	yes	Johanna Wolf
ru	given first	yes (patronymic)	Иван Семёнов → Ivan Semyonov
es	given first	yes	Gonzalo Martín → Gonzalo Martin
vi	family first	yes	Đặng Quân → Dang Quan

Korean, Japanese and Chinese have no middle part, so the middle-name option is ignored for them rather than inventing one. There is a helper that answers that question directly: nameSupportsMiddleName.

Romanization

Romanization is the English pronunciation of the native form, not a translation, and each script gets there its own way:

• Latin scripts drop their diacritics: Pérez → Perez, Müller → Muller, Đỗ → Do.
• Cyrillic is transliterated character by character: Семёнов → Semyonov.
• Korean follows the Revised Romanization of Korean, including the sound changes between syllables — 석민 is Seongmin and not Seokmin — with the conventional spelling for surnames, so 김 is Kim rather than Gim.
• Japanese and Chinese carry the reading on each character, so 佐藤 is Sato and 王 is Wang.

English is the one language where romanizing changes nothing, because the names are already in the Latin alphabet.

Words and nicknames

The word pools cover every one of the nine, and so do randWord, the twenty-nine themed generators and randNickname. What differs between the languages is where the modifier goes and whether it changes shape beside the noun, and each language says both in its own frames and agreement rules.

Code	Language	Shapes	Example
ko	Korean	modifier, trailing word	멋진사자, 고양이꼬리
en	English	modifier, trailing word	MistyOwl, CatTail
ja	Japanese	modifier (attributive form)	青いライオン, 星の影
zh	Chinese	modifier, 的 before a verb	快乐熊猫, 奔跑的狮子
vi	Vietnamese	modifier after the noun	mèo xanh, đuôi mèo
es	Spanish	modifier after, agreeing	gato azul, luna dorada
it	Italian	modifier after, agreeing	gatto azzurro, luna dorata
de	German	modifier before, declining	blauer Wal, blaue Katze
ru	Russian	modifier before, declining	синий кит, синяя рыба

The shapes belong to the language, so Vietnamese writes its own: the modifier behind the noun, the possessed thing in front of its owner. A language that inflects tags each noun with its gender and lists the endings a modifier changes, so Spanish writes gato dorado beside luna dorada without either form being stored twice. A noun with no singular at all, such as ножницы, gafas or Jeans, is tagged plural instead, because no singular modifier can stand beside it.

Every language the name generator knows has word pools too, so language means the same nine codes on every generator in the package.

Japanese and Chinese reach a second noun only through の and 的. A bare noun-noun compound needs a particle more often than not in either language, and reads as garbled once the base word is abstract.

Sentences

Every language writes sentences too, and each declares the shapes its own grammar can carry. What differs is where the verb stands, how a phrase is marked, and which shapes the language can write at all.

Code	Language	Shapes it declares	A sentence
ko	Korean	every one, verb last	검은 고양이가 숲에서 잠잔다.
ja	Japanese	every one, verb last	ハンバーガーが暗礁で冷める。
zh	Chinese	every one, place before the verb	巨乌贼又品尝朱红椰汁。
en	English	every one, verb second	The angler cleans the towel in the balcony.
vi	Vietnamese	every one, verb second	Lông mi run trong khe núi.
es	Spanish	every one, verb second	La vía láctea brilla en la roca pobre.
it	Italian	every one, place without article	Il birraio audace dimentica l'eclissi.
de	German	no object, no place	Im Frühling blüht eine Chrysantheme noch.
ru	Russian	no object, no place	Спутанный юпитер едва светлеет.

German and Russian are narrower for the same reason. Both would put an object in the accusative and a place in another case again, and each case changes the noun's own ending or the article in front of it, so they do not declare the shapes that need one. Asking for one falls back to the closest they have. Sentences covers this in full.