• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1<?xml version="1.0" encoding="UTF-8" ?>
2<!DOCTYPE supplementalData SYSTEM "../../common/dtd/ldmlSupplemental.dtd">
3<!--
4Copyright © 1991-2013 Unicode, Inc.
5CLDR data files are interpreted according to the LDML specification (http://unicode.org/reports/tr35/)
6For terms of use, see http://www.unicode.org/copyright.html
7-->
8<supplementalData>
9	<version number="$Revision: 13787 $"/>
10	<transforms>
11		<transform source="ps" target="ps_Latn" variant="BGN" direction="forward" draft="contributed" alias="Pashto-Latin/BGN ps-Latn-t-ps-m0-bgn">
12			<tRule><![CDATA[
13#
14########################################################################
15# BGN/PCGN 1968 System
16#
17# This system was adopted in 1968 for the romanization of Pashto
18# geographic names in Afghanistan. Persian names in Afghanistan are
19# romanized in accordance with the Romanization System for Persian
20# (BGN/PCGN 1958 System), shown on pages 87-92).
21#
22# Originally prepared by Michael Everson <everson@evertype.com>
23########################################################################
24#
25# MINIMAL FILTER: Pashto-Latin
26#
27
28:: [ءآابةتثجحخدذرزسشصضطظعغفقكلمنهوىئَُِّْ٠١٢٣٤٥٦٧٨٩ټپځڅچډړږژښگڰڼیۍې] ;
29:: NFD (NFC) ;
30#
31#
32########################################################################
33
34
35#
36########################################################################
37#
38# Define All Transformation Variables
39#
40########################################################################
41#
42
43$alef = ’;
44$ayin = ‘;
45$disambig =  ̱ ;
46#
47#
48# Use this $wordBoundary until bug 2034 is fixed in ICU:
49# http://bugs.icu-project.org/cgi-bin/icu-bugs/transliterate?id=2034;expression=boundary;user=guest
50#
51
52$wordBoundary =  [^[:L:][:M:][:N:]] ;
53#
54#
55########################################################################
56
57[:Nd:]{٫}[:Nd:] ↔ [:Nd:]{','}[:Nd:] ; # ARABIC DECIMAL SEPARATOR
58[:Nd:]{٬}[:Nd:] ↔ [:Nd:]{'.'}[:Nd:] ; # ARABIC THOUSANDS SEPARATOR
59٫ ↔ ',' $disambig ; # ARABIC DECIMAL SEPARATOR
60٬ ↔ '.' $disambig ; # ARABIC THOUSANDS SEPARATOR
61#  ٭ ↔ ; # ARABIC FIVE POINTED STAR // no need to transliterate
62، ↔ ',' ; # ARABIC COMMA
63؛ ↔ ';' ; # ARABIC SEMICOLON
64؟ ↔ '?' ; # ARABIC QUESTION MARK
65٪ ↔ '%' ; # ARABIC PERCENT SIGN
66٠ ↔ 0 $disambig ; # ARABIC-INDIC DIGIT ZERO
67١ ↔ 1 $disambig ; # ARABIC-INDIC DIGIT ONE
68٢ ↔ 2 $disambig ; # ARABIC-INDIC DIGIT TWO
69٣ ↔ 3 $disambig ; # ARABIC-INDIC DIGIT THREE
70٤ ↔ 4 $disambig ; # ARABIC-INDIC DIGIT FOUR
71٥ ↔ 5 $disambig ; # ARABIC-INDIC DIGIT FIVE
72٦ ↔ 6 $disambig ; # ARABIC-INDIC DIGIT SIX
73٧ ↔ 7 $disambig ; # ARABIC-INDIC DIGIT SEVEN
74٨ ↔ 8 $disambig ; # ARABIC-INDIC DIGIT EIGHT
75٩ ↔ 9 $disambig ; # ARABIC-INDIC DIGIT NINE
76۰ ↔ 0 ; # EXTENDED ARABIC-INDIC DIGIT ZERO
77۱ ↔ 1 ; # EXTENDED ARABIC-INDIC DIGIT ONE
78۲ ↔ 2 ; # EXTENDED ARABIC-INDIC DIGIT TWO
79۳ ↔ 3 ; # EXTENDED ARABIC-INDIC DIGIT THREE
80۴ ↔ 4 ; # EXTENDED ARABIC-INDIC DIGIT FOUR
81۵ ↔ 5 ; # EXTENDED ARABIC-INDIC DIGIT FIVE
82۶ ↔ 6 ; # EXTENDED ARABIC-INDIC DIGIT SIX
83۷ ↔ 7 ; # EXTENDED ARABIC-INDIC DIGIT SEVEN
84۸ ↔ 8 ; # EXTENDED ARABIC-INDIC DIGIT EIGHT
85۹ ↔ 9 ; # EXTENDED ARABIC-INDIC DIGIT NINE
86
87#
88########################################################################
89#
90# Rules moved to front to avoid masking
91#
92########################################################################
93
94
95#
96########################################################################
97#
98# BGN Page 89 Rule 4
99#
100# The character sequences كه , زه , سه , and گه may be romanized k·h, z·h,
101# s·h, and g·h in order to differentiate those romanizations from the
102# digraphs kh, zh, sh, and gh.
103#
104########################################################################
105#
106
107كه → k·h ; # ARABIC LETTER KAF + HEH
108زه → z·h ; # ARABIC LETTER ZAIN + HEH
109سه → s·h ; # ARABIC LETTER SEEN + HEH
110گه → g·h ; # ARABIC LETTER GAF + HEH
111#
112#
113########################################################################
114#
115# End Rule 4
116#
117########################################################################
118
119
120
121#
122########################################################################
123#
124# BGN Page 91 Rule 7
125#
126# Doubles consonant sounds are represented in Arabic script by
127# placing a shaddah ( ّ ) over a consonant character. In romanization
128# the letter should be doubled. [The remainder of this rule deals with
129# the definite article and is lexical.]
130#
131########################################################################
132#
133
134بّ → bb ; # ARABIC LETTER BEH + SHADDA
135پّ → pp ; # ARABIC LETTER PEH + SHADDA
136تّ → tt ; # ARABIC LETTER TEH + SHADDA
137ټّ → ṯṯ ; # ARABIC LETTER TEH WITH RING + SHADDA
138ثّ → s̄s̄ ; # ARABIC LETTER THEH + SHADDA
139جّ → jj ; # ARABIC LETTER JEEM + SHADDA
140چّ → chch ; # ARABIC LETTER TCHEH + SHADDA
141ّڅ → tsts ; # ARABIC LETTER HAH WITH THREE DOTS ABOVE + SHADDA
142ّځ → dz ; # ARABIC LETTER HAH WITH HAMZA ABOVE + SHADDA
143حّ → ḥḥ ; # ARABIC LETTER HAH + SHADDA
144خّ → khkh ; # ARABIC LETTER KHAH + SHADDA
145دّ → dd ; # ARABIC LETTER DAL + SHADDA
146ّډ → ḏḏ ; # ARABIC LETTER DAL WITH RING + SHADDA
147ذّ → z̄z̄ ; # ARABIC LETTER THAL + SHADDA
148رّ → rr ; # ARABIC LETTER REH + SHADDA
149ّړ → ṟṟ ; # ARABIC LETTER REH WITH RING + SHADDA
150زّ → zz ; # ARABIC LETTER ZAIN + SHADDA
151ژّ → zhzh ; # ARABIC LETTER JEH + SHADDA
152ّژ → z͟hz͟h ; # ARABIC LETTER REH WITH DOT BELOW AND DOT ABOVE + SHADDA
153سّ → ss ; # ARABIC LETTER SEEN + SHADDA
154شّ → shsh ; # ARABIC LETTER SHEEN + SHADDA
155ّښ → s͟hs͟h ; # ARABIC LETTER SEEN WITH DOT BELOW AND DOT ABOVE
156صّ → ṣṣ ; # ARABIC LETTER SAD + SHADDA
157ضّ → ḍḍ ; # ARABIC LETTER DAD + SHADDA
158طّ → ṭṭ ; # ARABIC LETTER TAH + SHADDA
159ظّ → ẓẓ ; # ARABIC LETTER ZAH + SHADDA
160عّ → $ayin $ayin ; # ARABIC LETTER AIN + SHADDA
161غّ → ghgh ; # ARABIC LETTER GHAIN + SHADDA
162فّ → ff ; # ARABIC LETTER FEH + SHADDA
163قّ → qq ; # ARABIC LETTER QAF + SHADDA
164كّ → kk ; # ARABIC LETTER KAF + SHADDA
165ّگ → gg ; # ARABIC LETTER GAF + SHADDA
166ّڰ → gg ; # ARABIC LETTER GAF WITH RING + SHADDA
167لّ → ll ; # ARABIC LETTER LAM + SHADDA
168مّ → mm ; # ARABIC LETTER MEEM + SHADDA
169نّ → nn ; # ARABIC LETTER NOON + SHADDA
170ّڼ → ṉṉ ; # ARABIC LETTER NOON WITH RING + SHADDA
171هّ → hh ; # ARABIC LETTER HEH + SHADDA
172وّ → ww ; # ARABIC LETTER WAW + SHADDA
173ّی → yy ; # ARABIC LETTER FARSI YEH + SHADDA
174ىّ → yy ; # ARABIC LETTER YEH + SHADDA
175#
176#
177########################################################################
178#
179# End Rule 7
180#
181########################################################################
182
183
184
185#
186########################################################################
187#
188# Start of Transformations
189#
190########################################################################
191#
192
193$wordBoundary{ء →  ; # ARABIC LETTER HAMZA
194ء → $alef ; # ARABIC LETTER HAMZA
195$wordBoundary{ا →  ; # ARABIC LETTER ALEF
196آ → $alef ā ; # ARABIC LETTER ALEF WITH MADDA ABOVE
197ب → b ; # ARABIC LETTER BEH
198پ → p ; # ARABIC LETTER PEH
199ت → t ; # ARABIC LETTER TEH
200ټ → ṯ ; # ARABIC LETTER TEH WITH RING
201ة → h ; # ARABIC LETTER TEH MARBUTA
202ث → s̄ ; # ARABIC LETTER THEH
203ج → j ; # ARABIC LETTER JEEM
204چ → ch ; # ARABIC LETTER TCHEH
205څ → ts ; # ARABIC LETTER HAH WITH THREE DOTS ABOVE
206ځ → dz ; # ARABIC LETTER HAH WITH HAMZA ABOVE
207ح → ḥ ; # ARABIC LETTER HAH
208خ → kh ; # ARABIC LETTER KHAH
209د → d ; # ARABIC LETTER DAL
210ډ → ḏ ; # ARABIC LETTER DAL WITH RING
211ذ → z̄ ; # ARABIC LETTER THAL
212ر → r ; # ARABIC LETTER REH
213ړ → ṟ ; # ARABIC LETTER REH WITH RING
214ز → z ; # ARABIC LETTER ZAIN
215ژ → zh ; # ARABIC LETTER JEH
216ږ → z͟h ; # ARABIC LETTER REH WITH DOT BELOW AND DOT ABOVE
217س → s ; # ARABIC LETTER SEEN
218ش → sh ; # ARABIC LETTER SHEEN
219ښ → s͟h ; # ARABIC LETTER SEEN WITH DOT BELOW AND DOT ABOVE
220ص → ṣ ; # ARABIC LETTER SAD
221ض → ẕ ; # ARABIC LETTER DAD
222ط → ṭ ; # ARABIC LETTER TAH
223ظ → ẓ ; # ARABIC LETTER ZAH
224ع → $ayin ; # ARABIC LETTER AIN
225غ → gh ; # ARABIC LETTER GHAIN
226ف → f ; # ARABIC LETTER FEH
227ق → q ; # ARABIC LETTER QAF
228ك → k ; # ARABIC LETTER KAF
229گ → g ; # ARABIC LETTER GAF
230ڰ → g ; # ARABIC LETTER GAF WITH RING
231ل → l ; # ARABIC LETTER LAM
232م → m ; # ARABIC LETTER MEEM
233ن → n ; # ARABIC LETTER NOON
234ڼ → ṉ ; # ARABIC LETTER NOON WITH RING
235و → w ; # ARABIC LETTER WAW
236ه → h ; # ARABIC LETTER HEH
237ٔی → ey ; # ARABIC LETTER FARSI YEH + HAMZA ABOVE
238ی → y ; # ARABIC LETTER FARSI YEH
239ى → y ; # ARABIC LETTER YEH
240ې → e ; # ARABIC LETTER E
241
242َا → ā ; # ARABIC FATHA + ALEF
243َىْ → ay ; # ARABIC FATHA + FARSI YEH + SUKUN
244َى → á ; # ARABIC FATHA + ALEF MAKSURA
245َْۍ → êy ; # ARABIC FATHA + SUKUN + YEH WITH TAIL
246َْ → ê ; # ARABIC FATHA + SUKUN
247َ → a ; # ARABIC FATHA
248
249ِْي → ey ; # ARABIC KASRA + FARSI YEH + SUKUN
250ِي → ī ; # ARABIC KASRA + FARSI YEH
251ِى → ī ; # ARABIC KASRA + YEH
252ِو → ew ; # ARABIC KASRA + WAW
253ِ → i ; # ARABIC KASRA
254
255ُوْ → ow ; # ARABIC DAMMA + WAW + SUKUN
256ُوی → ūy ; # ARABIC DAMMA + WAW + FARSI YEH
257ُو → ū ; # ARABIC DAMMA + WAW
258ُ → u ; # ARABIC DAMMA
259
260ْ →  ; # ARABIC SUKUN
261
262
263#
264#
265########################################################################
266
267			]]></tRule>
268		</transform>
269	</transforms>
270</supplementalData>
271