larvitgeodata
Version:
Geo data, primarily ISO territories, languages etc. Data fetched mostly from CLDR.
341 lines (335 loc) • 13.8 kB
text/xml
<?xml version="1.0" encoding="UTF-8" ?>
<!DOCTYPE supplementalData SYSTEM "../../common/dtd/ldmlSupplemental.dtd">
<!--
Copyright © 1991-2013 Unicode, Inc.
CLDR data files are interpreted according to the LDML specification (http://unicode.org/reports/tr35/)
For terms of use, see http://www.unicode.org/copyright.html
-->
<supplementalData>
<version number="$Revision: 11914 $"/>
<transforms>
<transform source="Turkmen" target="Latin" direction="forward" variant="BGN" draft="provisional">
<comment>
########################################################################
# BGN/PCGN 1979 System
#
# The BGN/PCGN system for Turkmen was designed for use in
# romanizing names written in the Turkmen alphabet.
# The Turkmen alphabet contains five letters not present
# in the Russian alphabet: Җҗ, Ңң, Өө, Үү, and Әә.
#
# The Turkmen Cyrillic Alphabet as defined by the BGN (Page 103):
#
# АБВГДЕЁЖҖЗИЙКЛМНҢОӨПРСТУҮФХЦЧШЩЪЫЬЭӘЮЯ
# абвгдеёжзҗийклмнңоөпрстуүфхцчшщъыьэәюя
#
# Originally prepared by Michael Everson <everson@evertype.com>
########################################################################
#
# MINIMAL FILTER: Turkmen-Latin
#
</comment>
<!-- <tRule>:: [АБВГДЕЁЖҖЗИЙКЛМНҢОӨПРСТУҮФХЦЧШЩЪЫЬЭӘЮЯабвгдеёжзҗийклмнңоөпрстуүфхцчшщъыьэәюя] ;</tRule> -->
<tRule>:: NFD (NFC) ;</tRule>
<comment>
#
########################################################################
</comment>
<comment>
########################################################################
#
# Define All Transformation Variables
#
########################################################################
#
</comment>
<tRule>$prime = ʹ ;</tRule>
<tRule>$doublePrime = ʺ ;</tRule>
<tRule>$upperConsonants = [БВГДЖҖЗЙКЛМНҢПРСТФХЦЧШЩЪЬ] ;</tRule>
<tRule>$lowerConsonants = [бвгджҗзйклмнңпрстфхцчшщъь] ;</tRule>
<tRule>$consonants = [$upperConsonants $lowerConsonants] ;</tRule>
<tRule>$upperVowels = [АЕЁИОӨУҮЫЭӘЮЯ] ;</tRule>
<tRule>$lowerVowels = [аеёиоөуүыэәюя] ;</tRule>
<tRule>$vowels = [$upperVowels $lowerVowels] ;</tRule>
<tRule>$lower = [$lowerConsonants $lowerVowels] ;</tRule>
<comment>
#
# Use this $wordBoundary until bug 2034 is fixed in ICU:
# http://bugs.icu-project.org/cgi-bin/icu-bugs/transliterate?id=2034;expression=boundary;user=guest
#
</comment>
<tRule>$wordBoundary = [^[:L:][:M:][:N:]] ;</tRule>
<comment>
#
########################################################################
</comment>
<comment>
########################################################################
#
# Start of Alphabetic Transformations
#
########################################################################
#
</comment>
<tRule>А → A ; # CYRILLIC CAPITAL LETTER A</tRule>
<tRule>а → a ; # CYRILLIC SMALL LETTER A</tRule>
<tRule>Б → B ; # CYRILLIC CAPITAL LETTER BE</tRule>
<tRule>б → b ; # CYRILLIC SMALL LETTER BE</tRule>
<tRule>В → W ; # CYRILLIC CAPITAL LETTER VE</tRule>
<tRule>в → w ; # CYRILLIC SMALL LETTER VE</tRule>
<tRule>Г → G ; # CYRILLIC CAPITAL LETTER GHE</tRule>
<tRule>г → g ; # CYRILLIC SMALL LETTER GHE</tRule>
<tRule>Д → D ; # CYRILLIC CAPITAL LETTER DE</tRule>
<tRule>д → d ; # CYRILLIC SMALL LETTER DE</tRule>
<comment>
#
########################################################################
#
# BGN Page 104 Rule 1:
#
# The character e should be romanized ye initially, after the vowel
# characters a, e, ё, и, о, ө, у, ү, ы, э, ю, and я, and after й, ъ, and ь.
# In all other instances, it should be romanized e.
#
########################################################################
#
</comment>
<tRule>Е}[[$upperVowels - [Ә]] [ЙЪЬ]] → YE ; # CYRILLIC CAPITAL LETTER IE</tRule>
<tRule>Е}[[$lowerVowels - [ә]] [йъь]] → Ye ; # CYRILLIC CAPITAL LETTER IE</tRule>
<tRule>$wordBoundary{Е → Ye ; # CYRILLIC CAPITAL LETTER IE</tRule>
<tRule>Е → E ; # CYRILLIC CAPITAL LETTER IE</tRule>
<tRule>е}[[$upperVowels - [Ә]] [$lowerVowels - [ә]] [ЙйЪъЬь]] → ye ; # CYRILLIC SMALL LETTER IE</tRule>
<tRule>$wordBoundary{е → ye ; # CYRILLIC SMALL LETTER IE</tRule>
<tRule>е → e ; # CYRILLIC SMALL LETTER IE</tRule>
<comment>
#
########################################################################
#
# End of Rule 1
#
########################################################################
#
</comment>
<tRule>Ё} $lower → Yo ; # CYRILLIC CAPITAL LETTER IO</tRule>
<tRule>Ё → YO ; # CYRILLIC CAPITAL LETTER IO</tRule>
<tRule>ё → yo ; # CYRILLIC SMALL LETTER IO</tRule>
<tRule>Ж} $lower → Zh ; # CYRILLIC CAPITAL LETTER ZHE</tRule>
<tRule>Ж → ZH ; # CYRILLIC CAPITAL LETTER ZHE</tRule>
<tRule>ж → zh ; # CYRILLIC SMALL LETTER ZHE</tRule>
<comment>
#
########################################################################
#
# BGN Page 104 Rule 2
#
# The character sequences зх, нг, сх, and цх may be romanized z·h,
# n·g, s·h, and ts·h in order to differentiate those romanizations form
# the digraphs zh, ng, sh, and the letter sequence tsh, which are used
# to render the characters ж, ң, ш, and the character sequence тш.
#
########################################################################
#
</comment>
<tRule>ЗХ → Z·H ; # CYRILLIC CAPITAL LETTER ZE</tRule>
<tRule>Зх → Z·h ; # CYRILLIC CAPITAL LETTER ZE</tRule>
<tRule>зх → z·h ; # CYRILLIC SMALL LETTER ZE</tRule>
<tRule>З → Z ; # CYRILLIC CAPITAL LETTER ZE</tRule>
<tRule>з → z ; # CYRILLIC SMALL LETTER ZE</tRule>
<comment>
#
########################################################################
#
# End Rule 2
#
########################################################################
#
</comment>
<tRule>И → И ; # CYRILLIC CAPITAL LETTER I</tRule>
<tRule>и → и ; # CYRILLIC SMALL LETTER I</tRule>
<tRule>Й → Y ; # CYRILLIC CAPITAL LETTER I</tRule>
<tRule>й → y ; # CYRILLIC SMALL LETTER I</tRule>
<tRule>К → K ; # CYRILLIC CAPITAL LETTER KA</tRule>
<tRule>к → k ; # CYRILLIC SMALL LETTER KA</tRule>
<tRule>Л → L ; # CYRILLIC CAPITAL LETTER EL</tRule>
<tRule>л → l ; # CYRILLIC SMALL LETTER EL</tRule>
<tRule>М → M ; # CYRILLIC CAPITAL LETTER EM</tRule>
<tRule>м → m ; # CYRILLIC SMALL LETTER EM</tRule>
<comment>
#
########################################################################
#
# BGN Page 104 Rule 2
#
# нг becomes n·g
#
########################################################################
#
</comment>
<tRule>НГ → N·G ; # CYRILLIC CAPITAL LETTER EN</tRule>
<tRule>Нг → N·g ; # CYRILLIC CAPITAL LETTER EN</tRule>
<tRule>нг → n·g ; # CYRILLIC SMALL LETTER EN</tRule>
<tRule>Н → N ; # CYRILLIC CAPITAL LETTER EN</tRule>
<tRule>н → n ; # CYRILLIC SMALL LETTER EN</tRule>
<comment>
#
########################################################################
#
# End Rule 2
#
########################################################################
#
</comment>
<tRule>Ң} $lower → Ng ; # CYRILLIC CAPITAL LETTER EN WITH DESCENDER</tRule>
<tRule>Ң → NG ; # CYRILLIC CAPITAL LETTER EN WITH DESCENDER</tRule>
<tRule>ң → ng ; # CYRILLIC SMALL LETTER EN WITH DESCENDER</tRule>
<tRule>О → O ; # CYRILLIC CAPITAL LETTER O</tRule>
<tRule>о → o ; # CYRILLIC SMALL LETTER O</tRule>
<tRule>Ө → Ö ; # CYRILLIC CAPITAL LETTER BARRED O</tRule>
<tRule>ө → ö ; # CYRILLIC SMALL LETTER BARRED O</tRule>
<tRule>П → P ; # CYRILLIC CAPITAL LETTER PE</tRule>
<tRule>п → p ; # CYRILLIC SMALL LETTER PE</tRule>
<tRule>Р → R ; # CYRILLIC CAPITAL LETTER ER</tRule>
<tRule>р → r ; # CYRILLIC SMALL LETTER ER</tRule>
<comment>
#
########################################################################
#
# BGN Page 104 Rule 2
#
# сх becomes s·h
#
########################################################################
#
</comment>
<tRule>СХ → S·H ; # CYRILLIC CAPITAL LETTER ES</tRule>
<tRule>Сх → S·h ; # CYRILLIC CAPITAL LETTER ES</tRule>
<tRule>сх → s·h ; # CYRILLIC SMALL LETTER ES</tRule>
<tRule>С → S ; # CYRILLIC CAPITAL LETTER ES</tRule>
<tRule>с → s ; # CYRILLIC SMALL LETTER ES</tRule>
<comment>
#
########################################################################
#
# End Rule 2
#
########################################################################
#
</comment>
<tRule>Т → T ; # CYRILLIC CAPITAL LETTER TE</tRule>
<tRule>т → t ; # CYRILLIC SMALL LETTER TE</tRule>
<tRule>У → U ; # CYRILLIC CAPITAL LETTER U</tRule>
<tRule>у → u ; # CYRILLIC SMALL LETTER U</tRule>
<tRule>Ү → Ü ; # CYRILLIC CAPITAL LETTER STRAIGHT U</tRule>
<tRule>ү → ü ; # CYRILLIC SMALL LETTER STRAIGHT U</tRule>
<tRule>Ф → F ; # CYRILLIC CAPITAL LETTER EF</tRule>
<tRule>ф → f ; # CYRILLIC SMALL LETTER EF</tRule>
<tRule>Х → H ; # CYRILLIC CAPITAL LETTER HA</tRule>
<tRule>х → h ; # CYRILLIC SMALL LETTER HA</tRule>
<comment>
#
########################################################################
#
# BGN Page 104 Rule 2
#
# цх becomes ts·h
#
########################################################################
#
</comment>
<tRule>ЦХ → TS·H ; # CYRILLIC CAPITAL LETTER GHE</tRule>
<tRule>Цх → Ts·h ; # CYRILLIC CAPITAL LETTER GHE</tRule>
<tRule>цх → ts·h ; # CYRILLIC SMALL LETTER GHE</tRule>
<tRule>Ц} $lower → Ts ; # CYRILLIC CAPITAL LETTER TSE</tRule>
<tRule>Ц → TS ; # CYRILLIC CAPITAL LETTER TSE</tRule>
<tRule>ц → ts ; # CYRILLIC SMALL LETTER TSE</tRule>
<comment>
#
########################################################################
#
# End Rule 2
#
########################################################################
#
</comment>
<tRule>Ч} $lower → Ch ; # CYRILLIC CAPITAL LETTER CHE</tRule>
<tRule>Ч → CH ; # CYRILLIC CAPITAL LETTER CHE</tRule>
<tRule>ч → ch ; # CYRILLIC SMALL LETTER CHE</tRule>
<comment>
#
########################################################################
#
# Implied rule from BGN Russian-Latin transliteration (Page 94 Note 3.6).
#
# шч becomes sh·ch
#
########################################################################
#
</comment>
<tRule>ШЧ → SH·CH ; # CYRILLIC CAPITAL LETTER SHA</tRule>
<tRule>Шч → Sh·ch ; # CYRILLIC CAPITAL LETTER SHA</tRule>
<tRule>шч → sh·ch ; # CYRILLIC SMALL LETTER SHA</tRule>
<tRule>Ш} $lower → Sh ; # CYRILLIC CAPITAL LETTER SHA</tRule>
<tRule>Ш → SH ; # CYRILLIC CAPITAL LETTER SHA</tRule>
<tRule>ш → sh ; # CYRILLIC SMALL LETTER SHA</tRule>
<tRule>Щ} $lower → Shch ; # CYRILLIC CAPITAL LETTER SHCHA</tRule>
<tRule>Щ → SHCH ; # CYRILLIC CAPITAL LETTER SHCHA</tRule>
<tRule>щ → shch ; # CYRILLIC SMALL LETTER SHCHA</tRule>
<comment>
#
########################################################################
#
# End Implied rule
#
########################################################################
#
</comment>
<tRule>Ъ → $doublePrime ; # CYRILLIC CAPITAL LETTER HARD SIGN</tRule>
<tRule>ъ → $doublePrime ; # CYRILLIC SMALL LETTER HARD SIGN</tRule>
<comment>
#
########################################################################
#
# BGN Page 104 Note 3
#
# The character Ыы may be romanized Ɨɨ instead of Yy, if so desired.
#
########################################################################
#
</comment>
<tRule>Ы → Y ; # CYRILLIC CAPITAL LETTER YERU</tRule>
<tRule>ы → y ; # CYRILLIC SMALL LETTER YERU</tRule>
<comment>
#
# Alternative rule to implement the option described here. To apply
# uncomment the following by removing the '#' mark at the start of the
# line and insert before the two rule lines above.
#
#Ы → Ɨ ; # CYRILLIC CAPITAL LETTER YERU
#ы → ɨ ; # CYRILLIC SMALL LETTER YERU
#
########################################################################
#
# End BGN Page 104 Note 2
#
########################################################################
#
</comment>
<tRule>Ь → $prime ; # CYRILLIC CAPITAL LETTER SOFT SIGN</tRule>
<tRule>ь → $prime ; # CYRILLIC SMALL LETTER SOFT SIGN</tRule>
<tRule>Э → E ; # CYRILLIC CAPITAL LETTER E</tRule>
<tRule>э → e ; # CYRILLIC SMALL LETTER E</tRule>
<tRule>Ә → Ä ; # CYRILLIC CAPITAL LETTER SCHWA</tRule>
<tRule>ә → ä ; # CYRILLIC SMALL LETTER SCHWA</tRule>
<tRule>Ю} $lower → Yu ; # CYRILLIC CAPITAL LETTER YU</tRule>
<tRule>Ю → YU ; # CYRILLIC CAPITAL LETTER YU</tRule>
<tRule>ю → yu ; # CYRILLIC SMALL LETTER YU</tRule>
<tRule>Я} $lower → Ya ; # CYRILLIC CAPITAL LETTER YA</tRule>
<tRule>Я → YA ; # CYRILLIC CAPITAL LETTER YA</tRule>
<tRule>я → ya ; # CYRILLIC SMALL LETTER YA</tRule>
<comment>
#
########################################################################
</comment>
</transform>
</transforms>
</supplementalData>