1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" 2"https://www.w3.org/TR/html4/loose.dtd"> 3<html> 4<head> 5 <meta name="generator" content= 6 "HTML Tidy for HTML5 for Apple macOS version 5.6.0"> 7 <meta http-equiv="Content-Type" content= 8 "text/html; charset=utf-8"> 9 <meta http-equiv="Content-Language" content="en-us"> 10 <link rel="stylesheet" href= 11 "../reports.css" type="text/css"> 12 <title>UTS #35: Unicode LDML: General</title> 13 <style type="text/css"> 14 <!-- 15 .dtd { 16 font-family: monospace; 17 font-size: 90%; 18 background-color: #CCCCFF; 19 border-style: dotted; 20 border-width: 1px; 21 } 22 23 .xmlExample { 24 font-family: monospace; 25 font-size: 80% 26 } 27 28 .blockedInherited { 29 font-style: italic; 30 font-weight: bold; 31 border-style: dashed; 32 border-width: 1px; 33 background-color: #FF0000 34 } 35 36 .inherited { 37 font-weight: bold; 38 border-style: dashed; 39 border-width: 1px; 40 background-color: #00FF00 41 } 42 43 .element { 44 font-weight: bold; 45 color: red; 46 } 47 48 .attribute { 49 font-weight: bold; 50 color: maroon; 51 } 52 53 .attributeValue { 54 font-weight: bold; 55 color: blue; 56 } 57 58 li, p { 59 margin-top: 0.5em; 60 margin-bottom: 0.5em 61 } 62 63 h2, h3, h4, table { 64 margin-top: 1.5em; 65 margin-bottom: 0.5em; 66 } 67 --> 68 </style> 69</head> 70<body> 71 <table class="header" width="100%"> 72 <tr> 73 <td class="icon"><a href="https://unicode.org"><img alt= 74 "[Unicode]" src="../logo60s2.gif" 75 width="34" height="33" style= 76 "vertical-align: middle; border-left-width: 0px; border-bottom-width: 0px; border-right-width: 0px; border-top-width: 0px;"></a> 77 <a class="bar" href= 78 "https://www.unicode.org/reports/">Technical Reports</a></td> 79 </tr> 80 <tr> 81 <td class="gray"> </td> 82 </tr> 83 </table> 84 <div class="body"> 85 <h2 style="text-align: center">Unicode Technical Standard #35</h2> 86 <h1>Unicode Locale Data Markup Language (LDML)<br> 87 Part 2: General</h1> 88 <!-- At least the first row of this header table should be identical across the parts of this UTS. --> 89 <table border="1" cellpadding="2" cellspacing="0" class="wide"> 90 <tr> 91 <td>Version</td> 92 <td>38</td> 93 </tr> 94 <tr> 95 <td>Editors</td> 96 <td>Yoshito Umaoka (<a href= 97 "mailto:yoshito_umaoka@us.ibm.com">yoshito_umaoka@us.ibm.com</a>) 98 and <a href="tr35.html#Acknowledgments">other CLDR 99 committee members</a></td> 100 </tr> 101 </table> 102 <p>For the full header, summary, and status, see <a href= 103 "tr35.html">Part 1: Core</a></p> 104 <h3><i>Summary</i></h3> 105 <p>This document describes parts of an XML format 106 (<i>vocabulary</i>) for the exchange of structured locale data. 107 This format is used in the <a href= 108 "https://unicode.org/cldr/">Unicode Common Locale Data 109 Repository</a>.</p> 110 <p>This is a partial document, describing general parts of the 111 LDML: display names & transforms, etc. For the other parts 112 of the LDML see the <a href="tr35.html">main LDML document</a> 113 and the links above.</p> 114 <h3><i>Status</i></h3> 115 116 <!-- NOT YET APPROVED 117 <p> 118 <i class="changed">This is a<b><font color="#ff3333"> 119 draft </font></b>document which may be updated, replaced, or superseded by 120 other documents at any time. Publication does not imply endorsement 121 by the Unicode Consortium. This is not a stable document; it is 122 inappropriate to cite this document as other than a work in 123 progress. 124 </i> 125 </p> 126 END NOT YET APPROVED --> 127 <!-- APPROVED --> 128 <p><i>This document has been reviewed by Unicode members and 129 other interested parties, and has been approved for publication 130 by the Unicode Consortium. This is a stable document and may be 131 used as reference material or cited as a normative reference by 132 other specifications.</i></p> 133 <!-- END APPROVED --> 134 135 <blockquote> 136 <p><i><b>A Unicode Technical Standard (UTS)</b> is an 137 independent specification. Conformance to the Unicode 138 Standard does not imply conformance to any UTS.</i></p> 139 </blockquote> 140 <p><i>Please submit corrigenda and other comments with the CLDR 141 bug reporting form [<a href="tr35.html#Bugs">Bugs</a>]. Related 142 information that is useful in understanding this document is 143 found in the <a href="tr35.html#References">References</a>. For 144 the latest version of the Unicode Standard see [<a href= 145 "tr35.html#Unicode">Unicode</a>]. For a list of current Unicode 146 Technical Reports see [<a href= 147 "tr35.html#Reports">Reports</a>]. For more information about 148 versions of the Unicode Standard, see [<a href= 149 "tr35.html#Versions">Versions</a>].</i></p> 150 <!-- This section of Parts should be identical in all of the parts of this UTS. --> 151 <h2><a name="Parts" href="#Parts" id="Parts">Parts</a></h2> 152 <p>The LDML specification is divided into the following 153 parts:</p> 154 <ul class="toc"> 155 <li>Part 1: <a href="tr35.html#Contents">Core</a> (languages, 156 locales, basic structure)</li> 157 <li>Part 2: <a href="tr35-general.html#Contents">General</a> 158 (display names & transforms, etc.)</li> 159 <li>Part 3: <a href="tr35-numbers.html#Contents">Numbers</a> 160 (number & currency formatting)</li> 161 <li>Part 4: <a href="tr35-dates.html#Contents">Dates</a> 162 (date, time, time zone formatting)</li> 163 <li>Part 5: <a href= 164 "tr35-collation.html#Contents">Collation</a> (sorting, 165 searching, grouping)</li> 166 <li>Part 6: <a href= 167 "tr35-info.html#Contents">Supplemental</a> (supplemental 168 data)</li> 169 <li>Part 7: <a href= 170 "tr35-keyboards.html#Contents">Keyboards</a> (keyboard 171 mappings)</li> 172 </ul> 173 <h2><a name="Contents" href="#Contents" id="Contents">Contents 174 of Part 2, General</a></h2> 175 <!-- START Generated TOC: CheckHtmlFiles --> 176 <ul class="toc"> 177 <li>1 <a href="#Display_Name_Elements">Display Name 178 Elements</a> 179 <ul class='toc'> 180 <li>1.1 <a href="#locale_display_name_algorithm">Locale Display Name Algorithm</a></li> 181 <li>1.2 <a href="#locale_display_name_fields">Locale Display Name Fields</a></li> 182 </ul> 183 </li> 184 <li>2 <a href="#Layout_Elements">Layout Elements</a></li> 185 <li>3 <a href="#Character_Elements">Character Elements</a> 186 <ul class="toc"> 187 <li>3.1 <a href="#Exemplars">Exemplars</a> 188 <ul class="toc"> 189 <li>3.1.1 <a href="#ExemplarSyntax">Exemplar 190 Syntax</a></li> 191 <li>3.1.2 <a href= 192 "#Restrictions">Restrictions</a></li> 193 </ul> 194 </li> 195 <li>3.2 <a href="#Character_Mapping">Mapping</a></li> 196 <li>3.3 <a href="#IndexLabels">Index Labels</a></li> 197 <li>3.4 <a href="#Ellipsis">Ellipsis</a></li> 198 <li>3.5 <a href="#Character_More_Info">More 199 Information</a></li> 200 <li>3.6 <a href="#Character_Parse_Lenient">Parse 201 Lenient</a></li> 202 </ul> 203 </li> 204 <li>4 <a href="#Delimiter_Elements">Delimiter 205 Elements</a></li> 206 <li>5 <a href="#Measurement_System_Data">Measurement System 207 Data</a> 208 <ul class="toc"> 209 <li>5.1 <a href="#Measurement_Elements">Measurement 210 Elements (deprecated)</a></li> 211 </ul> 212 </li> 213 <li>6 <a href="#Unit_Elements">Unit Elements</a> 214 <ul class="toc"> 215 <li>6.1 <a href="#Unit_Preference_and_Conversion">Unit Preference and Conversion Data</a></li> 216 <li>6.2 <a href="#Unit_Identifiers">Unit Identifiers</a></li> 217 <li>6.3 <a href="#Example_Units">Example Units</a></li> 218 <li>6.4 <a href="#compound-units">Compound Units</a></li> 219 <li>6.5 <a href="#Unit_Sequences">Unit Sequences (Mixed Units)</a></li> 220 <li>6.6 <a href="#durationUnit">durationUnit</a></li> 221 <li>6.7 <a href="#coordinateUnit">coordinateUnit</a></li> 222 <li>6.8 <a href= "#Territory_Based_Unit_Preferences">Territory-Based Unit 223 Preferences</a></li> 224 </ul> 225 </li> 226 <li>7 <a href="#POSIX_Elements">POSIX Elements</a></li> 227 <li>8 <a href="#Reference_Elements">Reference 228 Element</a></li> 229 <li>9 <a href="#Segmentations">Segmentations</a> 230 <ul class="toc"> 231 <li>9.1 <a href="#Segmentation_Inheritance">Segmentation 232 Inheritance</a></li> 233 <li>9.2 <a href="#Segmentation_Exceptions">Segmentation 234 Suppressions</a></li> 235 </ul> 236 </li> 237 <li>10 <a href="#Transforms">Transforms</a> 238 <ul class="toc"> 239 <li>10.1 <a href="#Inheritance">Inheritance</a> 240 <ul class="toc"> 241 <li>10.1.1 <a href="#Pivots">Pivots</a></li> 242 </ul> 243 </li> 244 <li>10.2 <a href="#Variants">Variants</a></li> 245 <li>10.3 <a href="#Transform_Rules_Syntax">Transform 246 Rules Syntax</a> 247 <ul class="toc"> 248 <li>10.3.1 <a href="#Dual_Rules">Dual Rules</a></li> 249 <li>10.3.2 <a href="#Context">Context</a></li> 250 <li>10.3.3 <a href="#Revisiting">Revisiting</a></li> 251 <li>10.3.4 <a href="#Example">Example</a></li> 252 <li>10.3.5 <a href="#Rule_Syntax">Rule 253 Syntax</a></li> 254 <li>10.3.6 <a href="#Transform_Rules">Transform 255 Rules</a></li> 256 <li>10.3.7 <a href= 257 "#Variable_Definition_Rules">Variable Definition 258 Rules</a></li> 259 <li>10.3.8 <a href="#Filter_Rules">Filter 260 Rules</a></li> 261 <li>10.3.9 <a href="#Conversion_Rules">Conversion 262 Rules</a></li> 263 <li>10.3.10 <a href= 264 "#Intermixing_Transform_Rules_and_Conversion_Rules">Intermixing 265 Transform Rules and Conversion Rules</a></li> 266 <li>10.3.11 <a href="#Inverse_Summary">Inverse 267 Summary</a></li> 268 </ul> 269 </li> 270 </ul> 271 </li> 272 <li>11 <a href="#ListPatterns">List Patterns</a> 273 <ul class="toc"> 274 <li>11.1 <a href="#List_Gender">Gender of Lists</a></li> 275 </ul> 276 </li> 277 <li>12 <a href="#Context_Transform_Elements">ContextTransform 278 Elements</a> 279 <ul class="toc"> 280 <li>Table: <a href= 281 "#contextTransformUsage_type_attribute_values">Element 282 contextTransformUsage type attribute values</a></li> 283 </ul> 284 </li> 285 <li>13 <a href="#Choice_Patterns">Choice Patterns</a></li> 286 <li>14 <a href="#Annotations">Annotations and Labels</a> 287 <ul class="toc"> 288 <li>14.1 <a href="#SynthesizingNames">Synthesizing 289 Sequence Names</a></li> 290 <li>14.2 <a href="#Character_Labels">Annotations 291 Character Labels</a></li> 292 <li>14.3 <a href="#Typographic_Names">Typographic 293 Names</a></li> 294 </ul> 295 </li> 296 <li>15 <a href="#Grammatical_Features">Grammatical Features</a> 297 <ul class="toc"> 298 <li>15.1 <a href="#Gender" >Gender</a></li> 299<li>15.2 <a href="#Case">Case</a></li> 300</ul> 301 </li> 302 <li>16 <a href="#Grammatical_Derivations">Grammatical Derivations</a> 303 <ul class="toc"><li>16.1<a href="#gender_compound_units">Deriving the Gender of Compound Units</a></li> 304 <li>16.2 <a href="#plural_compound_units">Deriving the Plural Category of Unit Components</a></li> 305 <li>16.3 <a href="#case_compound_units">Deriving the Case of Unit Components</a></li> 306 </ul> 307 </li> 308 <li> </li> 309 </ul> 310 <h2>1 <a name="Display_Name_Elements" href= 311 "#Display_Name_Elements" id="Display_Name_Elements">Display 312 Name Elements</a></h2> 313 <p class="dtd"><!ELEMENT localeDisplayNames ( alias | ( 314 localeDisplayPattern?, languages?, scripts?, territories?, 315 subdivisions?, variants?, keys?, types?, transformNames?, 316 measurementSystemNames?, codePatterns?, special* ) )></p> 317 <p>Display names for scripts, languages, countries, currencies, 318 and variants in this locale are supplied by this element. They 319 supply localized names for these items for use in 320 user-interfaces for various purposes such as displaying menu 321 lists, displaying a language name in a dialog, and so on. 322 Capitalization should follow the conventions used in the middle 323 of running text; the <contextTransforms> element may be 324 used to specify the appropriate capitalization for other 325 contexts (see <i>Section 12 <a href= 326 "#Context_Transform_Elements">ContextTransform Elements</a></i> 327 ). Examples are given below.</p> 328 <blockquote> 329 <p class="note"><b>Note:</b> The "<span style= 330 "color: blue">en</span>" locale may contain translated names 331 for deprecated codes for debugging purposes. Translation of 332 deprecated codes into other languages is discouraged.</p> 333 </blockquote> 334 <p>Where present, the display names must be unique; that is, 335 two distinct code would not get the same display name. (There 336 is one exception to this: in time zones, where parsing results 337 would give the same GMT offset, the standard and daylight 338 display names can be the same across different time zone 339 IDs.)</p> 340 <p>Any translations should follow customary practice for the 341 locale in question. For more information, see [<a href= 342 "tr35.html#DataFormats">Data Formats</a>].</p> 343 <p class="element2"><localeDisplayPattern></p> 344 <p class="dtd"><!ELEMENT localeDisplayPattern ( alias | 345 (localePattern*, localeSeparator*, localeKeyTypePattern*, 346 special*) ) ></p> 347 <p>For compound language (locale) IDs such as "pt_BR" which 348 contain additional subtags beyond the initial language code: 349 When the <languages> data does not explicitly specify a 350 display name such as "Brazilian Portuguese" for a given 351 compound language ID, "Portuguese (Brazil)" from the display 352 names of the subtags.</p> 353 <p>It includes three sub-elements:</p> 354 <ul> 355 <li>The <localePattern> element specifies a pattern 356 such as "{0} ({1})" in which {0} is replaced by the display 357 name for the primary language subtag and {1} is replaced by a 358 list of the display names for the remaining subtags.</li> 359 <li>The <localeSeparator> element specifies a pattern 360 such as "{0}, {1}" used when appending a subtag display name 361 to the list in the <localePattern> subpattern {1} 362 above. If that list includes more than one display name, then 363 <localeSeparator> subpattern {1} represents a new 364 display name to be appended to the current list in {0}. 365 <em>Note: Before CLDR 24, the <localeSeparator> element 366 specified a separator string such as ", ", not a 367 pattern.</em></li> 368 <li>The <localeKeyTypePattern> element specifies the 369 pattern used to display key-type pairs, such as "{0}: 370 {1}"</li> 371 </ul> 372 <p>For example, for the locale identifier 373 zh_Hant_CN_co_pinyin_cu_USD, the display would be "Chinese 374 (Traditional, China, Pinyin Sort Order, Currency: USD)". The 375 key-type for co_pinyin doesn't use the localeKeyTypePattern 376 because there is a translation for the key-type in English:</p> 377 <blockquote> 378 <p><type type="pinyin" key="collation">Pinyin Sort 379 Order</type></p> 380 </blockquote> 381 382 <h3>1.1 <a href="#locale_display_name_algorithm" name="locale_display_name_algorithm">Locale Display Name Algorithm</a></h3> 383 384 <p>A locale display name LDN is generated for a locale identifer L in the following way. First, canonicalize the locale identifier as per <strong><a href="tr35.html#Canonical_Unicode_Locale_Identifiers">Part 1, Section 3.2.1 Canonical Unicode Locale Identifiers</a></strong>. That will put the subtags in a defined order, and replace aliases by their canonical counterparts. (That defined order is followed in the processing below.) </p> 385 <p>Then follow each of the following steps for the subtags in L, building a base name LDN and a list of qualifying strings LQS.</p> 386 <p>Where there is a match for a subtag, disregard that subtag from L and add the element value to LDN or LQS as described bbelow. If there is no match for a subtag, use the fallback pattern with the subtag subtag instead.</p> 387 <p>Once LDN and LQS are built, return the following based on the length of LQS. </p> 388 <table class='simple'> 389 <tr> 390 <td>0</td> 391 <td>return LDN</td> 392 </tr> 393 <tr> 394 <td>1</td> 395 <td>use the <localePattern> to compose the result LDN from LDN and LQS[0], and return it.</td> 396 </tr> 397 <tr> 398 <td>>1</td> 399 <td>use the <localeSeparator> element value to join the elements of the list into LDN2, then use the <localePattern> to compose the result LDN from LDN and LDN2, and return it.</td> 400 </tr> 401 </table> 402 403 <p>The processing can be controled via the following parameters.</p> 404 <ul> 405 <li>CombineLanguage: boolean 406 <ul> 407 <li>Example: the CombineLanguage = true, picking the bold value below.</li> 408 <li> <language type="nl">Dutch</language></li> 409 <li><strong><language type="nl_BE">Flemish</language></strong></li> 410 </ul> 411 </li> 412 <li>PreferAlt: map from element to preferrred alt value, picking the bold value below. 413 <ul> 414 <li>Example: the PreferAlt contains {"language"="short"}:</li> 415 <li><language type="az">Azerbaijani</language></li> 416 <li><strong><language type="az" alt="short">Azeri</language></strong></li> 417 </ul> 418 </li> 419 </ul> 420 <p> In addition, the input locale display name could be minimized (see <a href="tr35.html#Likely_Subtags">Part 1: Section 4.3 Likely Subtags</a>) before generating the LDN. Selective minimization is often the best choice. For example, in a menu list it is often clearer to show the region if there are any regional variants. Thus the user would just see ["Spanish"] for es if the latter is the only supported Spanish, but where es-MX is also listed, then see ["Spanish (Spain)", "Spanish (Mexico)"].</p> 421 <hr> 422 <p><strong>Processing types of locale identifier subtags</strong> </p> 423 <p>When the display name contains "(" or ")" characters (or full-width equivalents), replace them "[", "]" (or full-width equivalents) before adding.</p> 424 <ol> 425 <li><strong>Language. </strong>Match the L subtags against the type values in the <language> elements. Pick the element with the most subtags matching. If there is more than one such element, pick the one that has subtypes matching earlier. If there are two such elements, pick the one that is alphabetically less. Set LBN to that value. Disregard any of the matching subtags in the following processing. 426 <ul> 427 <li>If CombineLanguage is false, only choose matches with the language subtag matching.</li> 428 </ul> 429 </li> 430 <li><strong>Script, Region, Variants.</strong> Where any of these subtags are in L, append the matching element value to LQS.</li> 431 <li><strong>T extensions. </strong>Get the value of the key="h0" type="hybrid" element, if there is one; otherwise the value of the <key type="t"> element. Next get the locale display name of the tlang. Join the pair using localePattern> and append to the LQS. Then format and add display names to LQS for any of the remaining tkey-tvalue pairs as described below.</li> 432 <li><strong>U extensions. </strong>If there is an attribute value A, process the key-value pair <"u", A> as below and append to LQS. Then format and add display names for each of the remaining key-type pairs as described below.</li> 433 <li><strong>Other extensions. </strong>There are currently no such extensions defined. Until such time as there are formats defined for them, append each of the extensions’s subtags to LQS.</li> 434 <li><strong>Private Use extensions. </strong>Get the value </li> 435 </ol> 436 <p><strong>Formatting T/U Key-Value pairs as display names</strong></p> 437 <ol> 438 <li> If there is a match for the key/value, then append the element value and return.</li> 439 <li>Otherwise, get the display name for the key, using the subtag if not available.</li> 440 <li>Format special values. As usual, if lacking data, use the subtag(s). 441 <ol> 442 <li>key="kr": (REORDER_CODE) assume the value is a script code, and get its display name.</li> 443 <li>key="dx": (SCRIPT_CODE) assume the value is a script code, and get its display name.</li> 444 <li>key="vt": (CODEPOINTS, deprecated) the value is a list of code points. Set the value display name to it, after replacing [-_] by space.</li> 445 <li>key="x0": (PRIVATE_USE) the value is a list of subtags. No formatting available, so use the subtag(s).</li> 446 <li>key="sd": (SUBDIVISION_CODE) use the subdivision data to find the display name.</li> 447 <li>key="rg": (RG_KEY_VALUE): handle as with key="sd"</li> 448 </ol> 449 </li> 450 <li>Then use the value of the <localeKeyTypePattern> element to join the key display name and the value display name, and append the result to LQS. </li> 451 </ol> 452 <p><strong>Examples of English locale display names</strong></p> 453 <table class='simple'> 454 <tr> 455 <th>Locale identifier</th> 456 <th>Locale display name</th> 457 </tr> 458 <tr> 459 <td>es</td> 460 <td>Spanish</td> 461 </tr> 462 <tr> 463 <td>es-419</td> 464 <td>Spanish (Latin America)</td> 465 </tr> 466 <tr> 467 <td>es-Cyrl-MX</td> 468 <td>Spanish (Cyrillic, Mexico)</td> 469 </tr> 470 <tr> 471 <td>en-Latn-GB-fonipa-scouse</td> 472 <td>English (Latin, United Kingdom, IPA Phonetics, Scouse)</td> 473 </tr> 474 <tr> 475 <td>en-u-nu-thai-ca-islamic-civil</td> 476 <td>English (Calendar: islamic-civil, Thai Digits)</td> 477 </tr> 478 <tr> 479 <td>hi-u-nu-latn-t-en-h0-hybrid</td> 480 <td>Hindi (Hybrid: English, Western Digits)</td> 481 </tr> 482 <tr> 483 <td>en-u-nu-deva-t-de</td> 484 <td>English (Transform: German, Devanagari Digits)</td> 485 </tr> 486 <tr> 487 <td>fr-z-zz-zzz-v-vv-vvv-u-uu-uuu-t-ru-Cyrl-s-ss-sss-a-aa-aaa-x-u-x</td> 488 <td>French (Transform: Russian [Cyrillic], uu: uuu, a: aa-aaa, s: ss-sss, v: vv-vvv, x: u-x, z: zz-zzz)</td> 489 </tr> 490 </table><br> 491 <h3>1.2 <a href="#locale_display_name_fields" name="locale_display_name_fields">Locale Display Name Fields</a></h3> 492 493 <p class="element2"><languages></p> 494 <p>This contains a list of elements that provide the 495 user-translated names for language codes, as described in 496 <i><a href= 497 "tr35.html#Unicode_Language_and_Locale_Identifiers">Section 3, 498 Unicode Language and Locale Identifiers</a></i>.</p> 499 <blockquote> 500 <pre><language type="<span style= 501 "color: blue">ab</span>"><span style= 502 "color: blue">Abkhazian</span></language> 503<language type="<span style= 504"color: blue">aa</span>"><span style= 505"color: blue">Afar</span></language> 506<language type="<span style= 507"color: blue">af</span>"><span style= 508"color: blue">Afrikaans</span></language> 509<language type="<span style= 510"color: blue">sq</span>"><span style= 511"color: blue">Albanian</span></language> 512</pre> 513 </blockquote> 514 <p>There should be no expectation that the list of languages 515 with translated names be complete: there are thousands of 516 languages that could have translated names. For debugging 517 purposes or comparison, when a language display name is 518 missing, the Description field of the language subtag registry 519 can be used to supply a fallback English user-readable 520 name.</p> 521 <p>The type can actually be any locale ID as specified above. 522 The set of which locale IDs is not fixed, and depends on the 523 locale. For example, in one language one could translate the 524 following locale IDs, and in another, fall back on the normal 525 composition.</p> 526 <table border="1" cellpadding="4" cellspacing="0"> 527 <tr> 528 <th width="33%">type</th> 529 <th width="33%">translation</th> 530 <th width="34%">composition</th> 531 </tr> 532 <tr> 533 <td width="33%">nl_BE</td> 534 <td width="33%">Flemish</td> 535 <td width="34%">Dutch (Belgium)</td> 536 </tr> 537 <tr> 538 <td width="33%">zh_Hans</td> 539 <td width="33%">Simplified Chinese</td> 540 <td width="34%">Chinese (Simplified)</td> 541 </tr> 542 <tr> 543 <td width="33%">en_GB</td> 544 <td width="33%">British English</td> 545 <td width="34%">English (United Kingdom)</td> 546 </tr> 547 </table> 548 <p>Thus when a complete locale ID is formed by composition, the 549 longest match in the language type is used, and the remaining 550 fields (if any) added using composition.</p> 551 <p>Alternate short forms may be provided for some languages 552 (and for territories and other display names), for example.</p> 553 <blockquote> 554 <pre><language type="<span style= 555 "color: blue">az</span>"><span style= 556 "color: blue">Azerbaijani</span></language> 557<language type="<span style= 558"color: blue">az</span>" alt="<span style= 559"color: blue">short</span>"><span style= 560"color: blue">Azeri</span></language> 561<language type="<span style= 562"color: blue">en_GB</span>"><span style= 563"color: blue">British English</span></language> 564<language type="<span style= 565"color: blue">en_GB</span>" alt="<span style= 566"color: blue">short</span>"><span style= 567"color: blue">U.K. English</span></language> 568<language type="<span style= 569"color: blue">en_US</span>"><span style= 570"color: blue">American English</span></language> 571<language type="<span style= 572"color: blue">en_US</span>" alt="<span style= 573"color: blue">short</span>"><span style= 574"color: blue">U.S. English</span></language> 575</pre> 576 </blockquote> 577 <p class="element2"><scripts></p> 578 <p>This element can contain an number of script elements. Each 579 script element provides the localized name for a script code, 580 as described in <i><a href= 581 "tr35.html#Unicode_Language_and_Locale_Identifiers">Section 3, 582 Unicode Language and Locale Identifiers</a></i> (see also 583 <i>UAX #24: Script Names</i> [<a href= 584 "https://www.unicode.org/reports/tr41/#UAX24">UAX24</a>]). For 585 example, in the language of this locale, the name for the Latin 586 script might be "Romana", and for the Cyrillic script is 587 "Kyrillica". That would be expressed with the following.</p> 588 <blockquote> 589 <pre><script type="<span style= 590 "color: blue">Latn</span>"><span style= 591 "color: blue">Romana</span></script> 592<script type="<span style= 593"color: blue">Cyrl</span>"><span style= 594"color: blue">Kyrillica</span></script> 595</pre> 596 </blockquote> 597 <p>The script names are most commonly used in conjunction with 598 a language name, using the <localePattern> combining 599 pattern, and the default form of the script name should be 600 suitable for such use. When a script name requires a different 601 form for stand-alone use, this can be specified using the 602 "stand-alone" alternate:</p> 603 <blockquote> 604 <pre><script type="<span style= 605 "color: blue">Hans</span>"><span style= 606 "color: blue">Simplified</span></script> 607<script type="<span style= 608"color: blue">Hans</span>" alt="<span style= 609"color: blue">stand-alone</span>"><span style= 610"color: blue">Simplified Han</span></script> 611<script type="<span style= 612"color: blue">Hant</span>"><span style= 613"color: blue">Traditional</span></script> 614<script type="<span style= 615"color: blue">Hant</span>" alt="<span style= 616"color: blue">stand-alone</span>"><span style= 617"color: blue">Traditional Han</span></script> 618</pre> 619 </blockquote> 620 <p>This will produce results such as the following:</p> 621 <ul> 622 <li>Display name of language + script, using 623 <localePattern>: “Chinese (Simplified)”</li> 624 <li>Display name of script alone, using 625 <localePattern>: “Simplified Han”</li> 626 </ul> 627 <p class="element2"><territories></p> 628 <p>This contains a list of elements that provide the 629 user-translated names for territory codes, as described in 630 <i><a href= 631 "tr35.html#Unicode_Language_and_Locale_Identifiers">Section 3, 632 Unicode Language and Locale Identifiers</a></i>.</p> 633 <blockquote> 634 <pre><territory type="<span style= 635 "color: blue">AD</span>"><span style= 636 "color: blue">Andorra</span></territory> 637<territory type="<span style= 638"color: blue">AF</span>"><span style= 639"color: blue">Afghanistan</span></territory> 640<territory type="<span style= 641"color: blue">AL</span>"><span style= 642"color: blue">Albania</span></territory> 643<territory type="<span style= 644"color: blue">AO</span>"><span style= 645"color: blue">Angola</span></territory> 646<territory type="<span style= 647"color: blue">DZ</span>"><span style= 648"color: blue">Algeria</span></territory> 649<territory type="<span style= 650"color: blue">GB</span>"><span style= 651"color: blue">United Kingdom</span></territory> 652<territory type="<span style= 653"color: blue">GB</span>" alt="<span style= 654"color: blue">short</span>"><span style= 655"color: blue">U.K.</span></territory> 656<territory type="<span style= 657"color: blue">US</span>"><span style= 658"color: blue">United States</span></territory> 659<territory type="<span style= 660"color: blue">US</span>" alt="<span style= 661"color: blue">short</span>"><span style= 662"color: blue">U.S.</span></territory> 663</pre> 664 </blockquote> 665 <p class="element2"><variants></p> 666 <p>This contains a list of elements that provide the 667 user-translated names for the <i>variant_code</i> values 668 described in <i><a href= 669 "tr35.html#Unicode_Language_and_Locale_Identifiers">Section 3, 670 Unicode Language and Locale Identifiers</a></i> .</p> 671 <blockquote> 672 <pre><variant type="<span style= 673 "color: blue">nynorsk</span>"><span style= 674 "color: blue">Nynorsk</span></variant> 675</pre> 676 </blockquote> 677 <p class="element2"><keys></p> 678 <p>This contains a list of elements that provide the 679 user-translated names for the <i>key</i> values described in 680 <i><a href= 681 "tr35.html#Unicode_Language_and_Locale_Identifiers">Section 3, 682 Unicode Language and Locale Identifiers</a></i>.</p> 683 <blockquote> 684 <pre><key type="<span style= 685 "color: blue">collation</span>"><span style= 686 "color: blue">Sortierung</span></key></pre> 687 </blockquote> 688 <p>Note that the <strong>type</strong> values may use aliases. Thus if the locale u-extension key "co" does not match, then the aliases have to be tried, using the bcp47 XML data:</p> 689 <blockquote> 690 <p> <key name="<strong>co</strong>" description="…" alias="<strong>collation</strong>"></p> 691 </blockquote> 692 <p class="element2"><types></p> 693 <p>This contains a list of elements that provide the 694 user-translated names for the <i>type</i> values 695 described in <i><a href= 696 "tr35.html#Unicode_Language_and_Locale_Identifiers">Section 3, 697 Unicode Language and Locale Identifiers</a></i> . Since the 698 translation of an option name may depend on the <i>key</i> it 699 is used with, the latter is optionally supplied.</p> 700 <blockquote> 701 <pre><type type="<span style= 702 "color: blue">phonebook</span>" key="<span style= 703 "color: blue">collation</span>"><span style= 704 "color: blue">Telefonbuch</span></type> 705</pre> 706 </blockquote> 707 <p>Note that the <strong>key</strong> and <strong>type</strong> values may use aliases. Thus if the locale u-extension key "co" does not match, then the aliases have to be tried, using the bcp47 XML data.</p> 708 <blockquote> 709 <p><key name="<strong>co</strong>" description="…" alias="<strong>collation</strong>"></p> 710 <p> <type name="<strong>phonebk</strong>" description="…" alias="<strong>phonebook</strong>"/></p> 711 </blockquote> 712 <p class="element2"><measurementSystemNames></p> 713 <p>This contains a list of elements that provide the 714 user-translated names for systems of measurement. The types 715 currently supported are "<span style="color: blue">US</span>", 716 "<span style="color: blue">metric</span>", and "<span style= 717 "color: blue">UK</span>".</p> 718 <blockquote> 719 <pre><measurementSystemName type="<span style= 720 "color: blue">US</span>"><span style= 721 "color: blue">U.S.</span></type> 722</pre> 723 </blockquote> 724 <p class="note"><b>Note:</b> In the future, we may need to add 725 display names for the particular measurement units (millimeter 726 versus millimetre versus whatever the Greek, Russian, etc are), 727 and a message format for positioning those with respect to 728 numbers. For example, "{number} {unitName}" in some languages, 729 but "{unitName} {number}" in others.</p> 730 <p class="element2"><transformNames></p> 731 <p> </p> 732 <blockquote> 733 <pre><transformName type="<span style= 734 "color: blue">Numeric</span>"><span style= 735 "color: blue">Numeric</span></type> 736</pre> 737 </blockquote> 738 <p class="element2"><codePatterns></p> 739 <blockquote> 740 <pre><codePattern type="<span style= 741 "color: blue">language</span>"><span style= 742 "color: blue">Language: {0}</span></type> 743</pre> 744 </blockquote> 745 <p class="dtd"><!ELEMENT subdivisions ( alias | ( 746 subdivision | special )* ) ><br> 747 <!ELEMENT subdivision ( #PCDATA )></p> 748 <p>Note that the subdivision names are in separate files, in 749 the subdivisions/ directory. The type values are the fully 750 qualified subdivsion names. For example:</p> 751 <p class="xmlExample"><subdivision type="AL-04">Fier 752 County</subdivision><br> 753 <subdivision type="AL-FR">Fier</subdivision> 754 <!-- in AL-04 : Fier County --><br> 755 <subdivision type="AL-LU">Lushnjë</subdivision> 756 <!-- in AL-04 : Fier County --><br> 757 <subdivision type="AL-MK">Mallakastër</subdivision> 758 <!-- in AL-04 : Fier County --></p> 759 <p>See also <strong>Part 6</strong> <em>Section 2.1.1 <a href= 760 "tr35-info.html#Subdivision_Containment">Subdivision 761 Containment</a></em>.</p> 762 <h2>2 <a name="Layout_Elements" href="#Layout_Elements" id= 763 "Layout_Elements">Layout Elements</a></h2> 764 <p class="dtd"><!ELEMENT layout ( alias | (orientation*, 765 inList*, inText*, special*) ) ></p> 766 <p>This top-level element specifies general layout features. It 767 currently only has one possible element (other than 768 <special>, which is always permitted).</p> 769 <p class="dtd"><!ELEMENT orientation ( characterOrder*, 770 lineOrder*, special* ) ><br> 771 <!ELEMENT characterOrder ( #PCDATA ) ><br> 772 <!ELEMENT lineOrder ( #PCDATA ) ></p> 773 <p>The lineOrder and characterOrder elements specify the 774 default general ordering of lines within a page, and characters 775 within a line. The possible values are:</p> 776 <table> 777 <tr> 778 <th>Direction</th> 779 <th>Value</th> 780 </tr> 781 <tr> 782 <td rowspan="2">Vertical</td> 783 <td>top-to-bottom</td> 784 </tr> 785 <tr> 786 <td>bottom-to-top</td> 787 </tr> 788 <tr> 789 <td rowspan="2">Horizontal</td> 790 <td>left-to-right</td> 791 </tr> 792 <tr> 793 <td>right-to-left</td> 794 </tr> 795 </table> 796 <p>If the value of lineOrder is one of the vertical values, 797 then the value of characterOrder must be one of the horizontal 798 values, and vice versa. For example, for English the lines are 799 top-to-bottom, and the characters are left-to-right. For 800 Mongolian (in the Mongolian Script) the lines are 801 right-to-left, and the characters are top to bottom. This does 802 not override the ordering behavior of bidirectional text; it 803 does, however, supply the paragraph direction for that text 804 (for more information, see <i>UAX #9: The Bidirectional 805 Algorithm</i> [<a href= 806 "https://www.unicode.org/reports/tr41/#UAX9">UAX9</a>]).</p> 807 <p>For dates, times, and other data to appear in the right 808 order, the display for them should be set to the orientation of 809 the locale.</p> 810 <p><inList> (deprecated)</p> 811 <p>The <inList> element is deprecated and has been 812 superseded by the <contextTransforms> element; see 813 <i>Section 12 <a href= 814 "#Context_Transform_Elements">ContextTransform Elements</a></i> 815 .</p> 816 <p>This element controls whether display names (language, 817 territory, etc) are title cased in GUI menu lists and the like. 818 It is only used in languages where the normal display is lower 819 case, but title case is used in lists. There are two 820 options:</p> 821 <pre><inList casing="titlecase-words"></pre> 822 <pre><inList casing="titlecase-firstword"></pre> 823 <p>In both cases, the title case operation is the default title 824 case function defined by Chapter 3 of <i>[<a href= 825 "tr35.html#Unicode">Unicode</a>]</i> . In the second case, only 826 the first word (using the word boundaries for that locale) will 827 be title cased. The results can be fine-tuned by using 828 alt="list" on any element where titlecasing as defined by the 829 Unicode Standard will produce the wrong value. For example, 830 suppose that "turc de Crimée" is a value, and the title case 831 should be "Turc de Crimée". Then that can be expressed using 832 the alt="list" value.</p> 833 <p><inText> (deprecated)</p> 834 <p>The <inList> element is deprecated and has been 835 superseded by the <contextTransforms> element; see 836 <i>Section 12 <a href= 837 "#Context_Transform_Elements">ContextTransform Elements</a></i> 838 .</p> 839 <p>This element indicates the casing of the data in the 840 category identified by the inText type attribute, when that 841 data is written in text or how it would appear in a dictionary. 842 For example :</p> 843 <pre> 844 <inText type="languages">lowercase-words</inText></pre> 845 <p>indicates that language names embedded in text are normally 846 written in lower case. The possible values and their meanings 847 are :</p> 848 <ul> 849 <li>titlecase-words : all words in the phrase should be title 850 case</li> 851 <li>titlecase-firstword : the first word should be title 852 case</li> 853 <li>lowercase-words : all words in the phrase should be lower 854 case</li> 855 <li>mixed : a mixture of upper and lower case is permitted. 856 generally used when the correct value is unknown.</li> 857 </ul> 858 <h2>3 <a name="Character_Elements" href="#Character_Elements" 859 id="Character_Elements">Character Elements</a></h2> 860 <p class="dtd"><!ELEMENT characters ( alias | ( 861 exemplarCharacters*, ellipsis*, moreInformation*, stopwords*, 862 indexLabels*, mapping*, parseLenients*, special* ) ) ></p> 863 <p>The <characters> element provides optional information 864 about characters that are in common use in the locale, and 865 information that can be helpful in picking resources or data 866 appropriate for the locale, such as when choosing among 867 character encodings that are typically used to transmit data in 868 the language of the locale. It may also be used to help reduce 869 confusability issues: see [<a href= 870 "https://www.unicode.org/reports/tr41/#UTR36">UTR39</a>]. It 871 typically only occurs in a language locale, not in a 872 language/territory locale. The stopwords are an experimental 873 feature, and should not be used.</p> 874 <h3>3.1 <a name="Exemplars" href="#Exemplars" id= 875 "Exemplars">Exemplars</a></h3> 876 <p>Exemplars are characters used by a language, separated into 877 different categories. The following table provides a summary, 878 with more details below.</p> 879 <table> 880 <tr> 881 <th scope="col">Type</th> 882 <th scope="col">Description</th> 883 <th scope="col">Examples</th> 884 </tr> 885 <tr> 886 <td>main / standard</td> 887 <td>Main letters used in the language</td> 888 <td style= 889 "font-family: Georgia, 'Times New Roman', Times, serif">a-z 890 å æ ø</td> 891 </tr> 892 <tr> 893 <td><span class="element2">auxiliary</span></td> 894 <td>Additional characters for common foreign words, 895 technical usage</td> 896 <td style= 897 "font-family: Georgia, 'Times New Roman', Times, serif">á à 898 ă â å ä ã ā æ ç é è ĕ ê ë ē í ì ĭ î ï ī ñ ó ò ŏ ô ö ø ō œ ú 899 ù ŭ û ü ū ÿ</td> 900 </tr> 901 <tr> 902 <td><span class="element2">index</span></td> 903 <td>Characters for the header of an index</td> 904 <td style= 905 "font-family: Georgia, 'Times New Roman', Times, serif">A B 906 C D E F G H I J K L M N O P Q R S T U V W X Y Z</td> 907 </tr> 908 <tr> 909 <td>punctuation</td> 910 <td>Common punctuation</td> 911 <td style= 912 "font-family: Georgia, 'Times New Roman', Times, serif">- ‐ 913 – — , ; \: ! ? . … “ ” ‘ ’ ( ) [ ] § @ * / & # † ‡ ′ 914 ″</td> 915 </tr> 916 <tr> 917 <td>numbers</td> 918 <td>The characters needed to display the common number 919 formats: decimal, percent, and currency.</td> 920 <td style= 921 "font-family: Georgia, 'Times New Roman', Times, serif"> 922 [\u061C\u200E \- , ٫ ٬ . % ٪ ‰ ؉ + 0٠ 1١ 2٢ 3٣ 4٤ 5٥ 6٦ 7٧ 923 8٨ 9٩]</td> 924 </tr> 925 </table> 926 <p>The basic exemplar character sets (main and auxiliary) 927 contain the commonly used letters for a given modern form of a 928 language, which can be for testing and for determining the 929 appropriate repertoire of letters for charset conversion or 930 collation. ("Letter" is interpreted broadly, as anything having 931 the property Alphabetic in the [<a href= 932 "https://unicode.org/reports/tr41/#UAX44">UAX44</a>], which also 933 includes syllabaries and ideographs.) It is not a complete set 934 of letters used for a language, nor should it be considered to 935 apply to multiple languages in a particular country. 936 Punctuation and other symbols should not be included in the 937 main and auxiliary sets. In particular, format characters like 938 CGJ are not included.</p> 939 <p>There are five sets altogether: main, auxiliary, 940 punctuation, numbers, and index. The <i>main</i> set should 941 contain the minimal set required for users of the language, 942 while the <i>auxiliary</i> exemplar set is designed to 943 encompass additional characters: those non-native or historical 944 characters that would customarily occur in common publications, 945 dictionaries, and so on. Major style guidelines are good 946 references for the auxiliary set. So, for example, if Irish 947 newspapers and magazines would commonly have Danish names using 948 å, for example, then it would be appropriate to include å in 949 the auxiliary exemplar characters; just not in the main 950 exemplar set. Thus English has the following:</p> 951 <p><exemplarCharacters>[a b c d e f g h i j k l m n o p q 952 r s t u v w x y z]</exemplarCharacters><br> 953 <exemplarCharacters type="auxiliary">[á à ă â å ä ã ā æ ç 954 é è ĕ ê ë ē í ì ĭ î ï ī ñ ó ò ŏ ô ö ø ō œ ú ù ŭ û ü ū 955 ÿ]</exemplarCharacters></p> 956 <p>For a given language, there are a few factors that help for 957 determining whether a character belongs in the auxiliary set, 958 instead of the main set:</p> 959 <ul> 960 <li>The character is not available on all normal 961 keyboards.</li> 962 <li>It is acceptable to always use spellings that avoid that 963 character.</li> 964 </ul> 965 <p>For example, the exemplar character set for en (English) is 966 the set [a-z]. This set does not contain the accented letters 967 that are sometimes seen in words like "résumé" or "naïve", 968 because it is acceptable in common practice to spell those 969 words without the accents. The exemplar character set for fr 970 (French), on the other hand, must contain those characters: 971 [a-z é è ù ç à â ê î ô û æ œ ë ï ÿ]. The main set typically 972 includes those letters commonly "alphabet".</p> 973 <p>The <em>punctuation</em> set consists of common punctuation 974 characters that are used with the language (corresponding to 975 main and auxiliary). Symbols may also be included where they 976 are common in plain text, such as ©. It does not include 977 characters with narrow technical usage, such as dictionary 978 punctuation/symbols or copy-edit symbols. For example, English 979 would have something like the following:</p> 980 <blockquote> 981 - ‐ – —<br> 982 , ; : ! ? . …<br> 983 ' ‘ ’ " “ ” ′ ″<br> 984 ( ) [ ] { } ⟨ ⟩<br> 985 © ® ™ @ & ° ‧ ·/ # % ¶ § * † ‡<br> 986 + − ± × ÷ < ≤ = ≅ ≥ > √<br> 987 </blockquote> 988 <p>The numbers exemplars does not currently include lesser-used 989 characters: exponential notation (3.1 × 10²³, ∞, NAN). Nor does 990 it contain the units or currency symbols such as $, ¥, ₹,… It 991 does contain %, because that occurs in the percent format. It 992 may contain some special formatting characters like the RLM. A 993 full list of the currency symbols used with that locale are in 994 the <currencies> element, while the units can be gotten 995 from the <units> element (both using inheritance, of 996 course).The digits used in each numbering system are accessed 997 in numberingSystems.xml. For more information, see 998 <em><strong>Part 3: <a href= 999 "tr35-numbers.html#Contents">Numbers</a></strong> , Section 1000 2 <a href="tr35-numbers.html#Number_Elements">Number 1001 Elements</a></em>.</p> 1002 <p><em>Examples for zh.xml:</em></p> 1003 <table> 1004 <tr> 1005 <th scope="col">Type</th> 1006 <th scope="col">Description</th> 1007 </tr> 1008 <tr> 1009 <td>defaultNumberingSystem</td> 1010 <td>latn</td> 1011 </tr> 1012 <tr> 1013 <td>otherNumberingSystems/native</td> 1014 <td>hanidec</td> 1015 </tr> 1016 <tr> 1017 <td>otherNumberingSystems/traditional</td> 1018 <td>hans</td> 1019 </tr> 1020 <tr> 1021 <td>otherNumberingSystems/finance</td> 1022 <td>hansfin</td> 1023 </tr> 1024 </table> 1025 <p>When determining the character repertoire needed to support 1026 a language, a reasonable initial set would include at least the 1027 characters in the main and punctuation exemplar sets, along 1028 with the digits and common symbols associated with the 1029 numberSystems supported for the locale (see <i><a href= 1030 "tr35-numbers.html#Numbering_Systems">Numbering 1031 Systems</a></i>).</p> 1032 <p>The <em>index</em> characters are a set of characters for 1033 use as a UI "index", that is, a list of clickable characters 1034 (or character sequences) that allow the user to see a segment 1035 of a larger "target" list. For details see the <a href= 1036 "tr35-collation.html#Collation_Indexes">Unicode LDML: 1037 Collation</a> document. The index set may only contain 1038 characters whose lowercase versions are in the main and 1039 auxiliary exemplar sets, though for cased languages the index 1040 exemplars are typically in uppercase. Characters from the 1041 auxiliary exemplar set may be necessary in the index set if it 1042 needs to properly handle items such as names which may require 1043 characters not included in the main exemplar set.</p> 1044 <p>Here is a sample of the XML structure:</p> 1045 <pre> 1046 <exemplarCharacters type="index">[A B C D E F G H I J K L M N O P Q R S T U V W X Y Z]</exemplarCharacters></pre> 1047 <p>The display of the index characters can be modified with the 1048 Index labels elements, discussed in Section 5.6.4.</p> 1049 <h4>3.1.1 <a name="ExemplarSyntax" href="#ExemplarSyntax" id= 1050 "ExemplarSyntax">Exemplar Syntax</a></h4> 1051 <p>In all of the exemplar characters, the list of characters is 1052 in the <a href="tr35.html#Unicode_Sets">Unicode Set</a> format, 1053 which normally allows boolean combinations of sets of letters 1054 and Unicode properties.</p> 1055 <p>Sequences of characters that act like a single letter in the 1056 language — especially in collation — are included within 1057 braces, such as [a-z á é í ó ú ö ü ő ű {cs} {dz} {dzs} {gy} 1058 ...]. The characters should be in normalized form (NFC). Where 1059 combining marks are used generatively, and apply to a large 1060 number of base characters (such as in Indic scripts), the 1061 individual combining marks should be included. Where they are 1062 used with only a few base characters, the specific combinations 1063 should be included. Wherever there is not a precomposed 1064 character (for example, single codepoint) for a given 1065 combination, that must be included within braces. For example, 1066 to include sequences from the <a href= 1067 "https://unicode.org/standard/where/">Where is my Character?</a> 1068 page on the Unicode site, one would write: [{ch} {tʰ} {x̣} {ƛ̓} 1069 {ą́} {i̇́} {ト゚}], but for French one would just write [a-z é è 1070 ù ...]. When in doubt use braces, since it does no harm to 1071 include them around single code points: for example, [a-z {é} 1072 {è} {ù} ...].</p> 1073 <p>If the letter 'z' were only ever used in the combination 1074 'tz', then we might have [a-y {tz}] in the main set. (The 1075 language would probably have plain 'z' in the auxiliary set, 1076 for use in foreign words.) If combining characters can be used 1077 productively in combination with a large number of others (such 1078 as say Indic matras), then they are not listed in all the 1079 possible combinations, but separately, such as:</p> 1080 <blockquote> 1081 [ॐ ऄ-ऋ ॠ ऌ ॡ ऍ-क क़ ख ख़ ग ग़ घ-ज ज़ झ-ड ड़ ढ ढ़ ण-फ फ़ ब-य 1082 य़ र-ह ़ ँ-ः ॑-॔ ऽ ् ॽ ा-ॄ ॢ ॣ ॅ-ौ] 1083 </blockquote> 1084 <p>The exemplar character set for Han characters is composed 1085 somewhat differently. It is even harder to draw a clear line 1086 for Han characters, since usage is more like a frequency curve 1087 that slowly trails off to the right in terms of decreasing 1088 frequency. So for this case, the exemplar characters simply 1089 contain a set of reasonably frequent characters for the 1090 language.</p> 1091 <p>The ordering of the characters in the set is irrelevant, but 1092 for readability in the XML file the characters should be in 1093 sorted order according to the locale's conventions. The main 1094 and auxiliary sets should only contain lower case characters 1095 (except for the special case of Turkish and similar languages, 1096 where the dotted capital I should be included); the upper case 1097 letters are to be mechanically added when the set is used. For 1098 more information on casing, see the discussion of Special 1099 Casing in the Unicode Character Database.</p> 1100 <h4>3.1.2 <a name="Restrictions" href="#Restrictions" id= 1101 "Restrictions">Restrictions</a></h4> 1102 <ol> 1103 <li>The main, auxiliary and index sets are normally 1104 restricted to those letters with a specific <a href= 1105 "https://unicode.org/Public/UNIDATA/Scripts.txt">Script</a> 1106 character property (that is, not the values Common or 1107 Inherited) or required <a href= 1108 "https://unicode.org/Public/UNIDATA/DerivedCoreProperties.txt"> 1109 Default_Ignorable_Code_Point</a> characters (such as a 1110 non-joiner), or combining marks, or the <a href= 1111 "https://www.unicode.org/Public/UNIDATA/auxiliary/WordBreakProperty.txt"> 1112 Word_Break</a> properties <a name="Katakana" href="#Katakana" 1113 id="Katakana">Katakana</a>, <a name="ALetter" href="#ALetter" 1114 id="ALetter">ALetter</a>, or <a name="MidLetter" href= 1115 "#MidLetter" id="MidLetter">MidLetter</a>.</li> 1116 <li>The auxiliary set should not overlap with the main set. 1117 There is one exception to this: Hangul Syllables and CJK 1118 Ideographs can overlap between the sets.</li> 1119 <li>Any <a href= 1120 "https://unicode.org/Public/UNIDATA/DerivedCoreProperties.txt"> 1121 Default_Ignorable_Code_Point</a>s should be in the auxiliary 1122 set , or, if they are only needed for currency formatting, in 1123 the currency set. These can include characters such as U+200E 1124 LEFT-TO-RIGHT MARK and U+200F RIGHT-TO-LEFT MARK which may be 1125 needed in bidirectional text in order for date, currency or 1126 other formats to display correctly.</li> 1127 <li>For exemplar characters the <a href= 1128 "tr35.html#Unicode_Sets">Unicode Set</a> format is restricted 1129 so as to not use properties or boolean combinations .</li> 1130 </ol> 1131 <h3>3.2 <a name="Character_Mapping" href="#Character_Mapping" 1132 id="Character_Mapping">Mapping</a></h3> 1133 <p><b>This element has been deprecated.</b> For information on 1134 its structure and how it was intended to specify 1135 locale-specific preferred encodings for various purposes 1136 (e-mail, web), see the <a href= 1137 "https://www.unicode.org/reports/tr35/tr35-39/tr35-general.html#Character_Mapping"> 1138 Mapping</a> section from the CLDR 27 version of the LDML 1139 Specification.</p> 1140 <h3>3.3 <a name="IndexLabels" href="#IndexLabels" id= 1141 "IndexLabels">Index Labels</a></h3> 1142 <p><b>This element and its subelements have been 1143 deprecated.</b> For information on its structure and how it was 1144 intended to provide data for a compressed display of index 1145 exemplar characters where space is limited, see the <a href= 1146 "https://www.unicode.org/reports/tr35/tr35-39/tr35-general.html#IndexLabels"> 1147 Index Labels</a> section from the CLDR 27 version of the LDML 1148 Specification.</p> 1149 <p class="dtd"><!ELEMENT indexLabels (indexSeparator*, 1150 compressedIndexSeparator*, indexRangePattern*, 1151 indexLabelBefore*, indexLabelAfter*, indexLabel*) ></p> 1152 <h3>3.4 <a name="Ellipsis" href="#Ellipsis" id= 1153 "Ellipsis">Ellipsis</a></h3> 1154 <p class="dtd"><!ELEMENT ellipsis ( #PCDATA ) ><br> 1155 <!ATTLIST ellipsis type ( initial | medial | final | 1156 word-initial | word-medial | word-final ) #IMPLIED ></p> 1157 <p>The ellipsis element provides patterns for use when 1158 truncating strings. There are three versions: initial for 1159 removing an initial part of the string (leaving final 1160 characters); medial for removing from the center of the string 1161 (leaving initial and final characters), and final for removing 1162 a final part of the string (leaving initial characters). For 1163 example, the following uses the ellipsis character in all three 1164 cases (although some languages may have different characters 1165 for different positions).</p> 1166 <p><code><ellipsis 1167 type="initial">…{0}</ellipsis><br> 1168 <ellipsis type="medial">{0}…{1}</ellipsis><br> 1169 <ellipsis type="final">{0}…</ellipsis></code></p> 1170 <p>There are alternatives for cases where the breaks are on a 1171 word boundary, where some languages include a space. For 1172 example, such as case would be:</p> 1173 <p><code><ellipsis type="word-initial">… 1174 {0}</ellipsis></code></p> 1175 <h3>3.5 <a name="Character_More_Info" href= 1176 "#Character_More_Info" id="Character_More_Info">More 1177 Information</a></h3> 1178 <p>The moreInformation string is one that can be displayed in 1179 an interface to indicate that more information is available. 1180 For example:</p> 1181 <p><moreInformation>?</moreInformation></p> 1182 <h3>3.6 <a name="Character_Parse_Lenient" href= 1183 "#Character_Parse_Lenient" id="Character_Parse_Lenient">Parse 1184 Lenient</a></h3> 1185 <p class='dtd'><!ELEMENT parseLenients ( alias | ( 1186 parseLenient*, special* ) ) ><br> 1187 <!ATTLIST parseLenients scope (general | number | date) 1188 #REQUIRED ><br> 1189 <!ATTLIST parseLenients level (lenient | stricter) #REQUIRED 1190 ></p> 1191 <p class='dtd'><!ELEMENT parseLenient ( #PCDATA ) ><br> 1192 <!ATTLIST parseLenient sample CDATA #REQUIRED ><br> 1193 <!ATTLIST parseLenient alt NMTOKENS #IMPLIED ><br> 1194 <!ATTLIST parseLenient draft (approved | contributed | 1195 provisional | unconfirmed) #IMPLIED ><br></p> 1196 <p>Example:</p> 1197 <pre><parseLenients scope="date" level="lenient"> 1198 <parseLenient sample="-">[\-./]</parseLenient> 1199 <parseLenient sample=":">[\:∶]</parseLenient> 1200</parseLenients></pre> 1201 <p>The parseLenient elements are used to indicate that 1202 characters within a particular UnicodeSet are normally to be 1203 treated as equivalent when doing a lenient parse. The 1204 <strong>scope</strong> attribute value defines where the 1205 lenient sets are intended for use. The <strong>level</strong> 1206 attribute value is included for future expansion; currently the 1207 only value is "lenient".</p> 1208 <p>The <strong>sample</strong> attribute value is a paradigm 1209 element of that UnicodeSet, but the only reason for pulling it 1210 out separately is so that different classes of characters are 1211 separated, and to enable inheritance overriding. The first 1212 version of this data is populated with the data used for 1213 lenient parsing from ICU.</p> 1214 <h2>4 <a name="Delimiter_Elements" href="#Delimiter_Elements" 1215 id="Delimiter_Elements">Delimiter Elements</a></h2> 1216 <p class="dtd"><!ELEMENT delimiters (alias | 1217 (quotationStart*, quotationEnd*, alternateQuotationStart*, 1218 alternateQuotationEnd*, special*)) ></p> 1219 <p>The delimiters supply common delimiters for bracketing 1220 quotations. The quotation marks are used with simple quoted 1221 text, such as:</p> 1222 <blockquote> 1223 <p>He said, “Don’t be absurd!”</p> 1224 </blockquote> 1225 <p>When quotations are nested, the quotation marks and 1226 alternate marks are used in an alternating fashion:</p> 1227 <blockquote> 1228 <p>He said, “Remember what the Mad Hatter said: ‘Not the same 1229 thing a bit! Why you might just as well say that “I see what 1230 I eat” is the same thing as “I eat what I see”!’”</p> 1231 </blockquote> 1232 <p><code><quotationStart></code> <span style= 1233 "color: blue">“</span> <code></quotationStart></code><br> 1234 <code><quotationEnd></code> <span style= 1235 "color: blue">”</span> <code></quotationEnd></code><br> 1236 <code><alternateQuotationStart></code> <span style= 1237 "color: blue">‘</span> 1238 <code></alternateQuotationStart></code><br> 1239 <code><alternateQuotationEnd></code> <span style= 1240 "color: blue">’</span> 1241 <code></alternateQuotationEnd></code></p> 1242 <h2>5 <a name="Measurement_System_Data" href= 1243 "#Measurement_System_Data" id= 1244 "Measurement_System_Data">Measurement System Data</a></h2> 1245 <p class="dtd"><!ELEMENT measurementData ( 1246 measurementSystem*, paperSize* ) ><br> 1247 <br> 1248 <!ELEMENT measurementSystem EMPTY ><br> 1249 <!ATTLIST measurementSystem type ( metric | US | UK ) 1250 #REQUIRED ><br> 1251 <!ATTLIST measurementSystem category ( temperature ) 1252 #IMPLIED ><br> 1253 <!ATTLIST measurementSystem territories NMTOKENS #REQUIRED 1254 ><br> 1255 <br> 1256 <!ELEMENT paperSize EMPTY ><br> 1257 <!ATTLIST paperSize type ( A4 | US-Letter ) #REQUIRED 1258 ><br> 1259 <!ATTLIST paperSize territories NMTOKENS #REQUIRED ></p> 1260 <p>The measurement system is the normal measurement system in 1261 common everyday use (except for date/time). For example:</p> 1262 <pre><measurementData> 1263 <measurementSystem type="metric" territories="001"/> 1264 <measurementSystem type="US" territories="LR MM US"/> 1265 <measurementSystem type="metric" category="temperature" territories="LR MM"/> 1266 <measurementSystem type="US" category="temperature" territories="BS BZ KY PR PW"/> 1267 <measurementSystem type="UK" territories="GB"/> 1268 <paperSize type="A4" territories="001"/> 1269 <paperSize type="US-Letter" territories="BZ CA CL CO CR GT MX NI PA PH PR SV US VE"/> 1270</measurementData></pre> 1271 <p>The values are "metric", "US", or "UK"; others may be added 1272 over time.</p> 1273 <ul> 1274 <li>The "metric" value indicates the use of SI [<a href= 1275 "tr35.html#ISO1000">ISO1000</a>] base or derived units, or 1276 non-SI units accepted for use with SI: for example, meters, 1277 kilograms, liters, and degrees Celsius.</li> 1278 <li>The "US" value indicates the customary system of 1279 measurement as used in the United States: feet, inches, 1280 pints, quarts, degrees Fahrenheit, and so on.</li> 1281 <li>The "UK" value indicates the mix of metric units and 1282 Imperial units (feet, inches, pints, quarts, and so on) used 1283 in the United Kingdom, in which Imperial volume units such as 1284 pint, quart, and gallon are different sizes than in the "US" 1285 customary system. For more detail about specific units for 1286 various usages, see <strong>Part 6: Supplemental:</strong> 1287 <em>Section 2.4.1 <a href= 1288 "tr35-info.html#Preferred_Units_For_Usage">Preferred Units 1289 for Specific Usages</a></em>.</li> 1290 </ul> 1291 <p>In some cases, it may be common to use different measurement 1292 systems for different categories of measurements. For example, 1293 the following indicates that for the category of temperature, 1294 in the regions LR and MM, it is more common to use metric units 1295 than US units.</p> 1296 <pre> 1297 <measurementSystem type="metric" category="temperature" territories="LR MM"/> 1298 </pre> 1299 <p>The paperSize attribute gives the height and width of paper 1300 used for normal business letters. The values are "A4" and 1301 "US-Letter".</p> 1302 <p>For both measurementSystem entries and paperSize entries, 1303 later entries for specific territories such as "US" will 1304 override the value assigned to that territory by earlier 1305 entries for more inclusive territories such as "001".</p> 1306 <p>The measurement information was formerly in the main LDML 1307 file, and had a somewhat different format.</p> 1308 <p>Again, for finer-grained detail about specific units for 1309 various usages, see <strong>Part 6: Supplemental:</strong> 1310 <em>Section 2.4.1 <a href= 1311 "tr35-info.html#Preferred_Units_For_Usage">Preferred Units for 1312 Specific Usages</a></em>.</p> 1313 <h3>5.1 <a name="Measurement_Elements" href= 1314 "#Measurement_Elements" id="Measurement_Elements">Measurement 1315 Elements (deprecated)</a></h3> 1316 <p class="dtd"><!ELEMENT measurement (alias | 1317 (measurementSystem?, paperSize?, special*)) ></p> 1318 <p>The measurement element is deprecated in the main LDML 1319 files, because the data is more appropriately organized as 1320 connected to territories, not to linguistic data. Instead, the 1321 measurementData element in the supplemental data file should be 1322 used.</p> 1323 <h2>6 <a name="Unit_Elements" href="#Unit_Elements" id= 1324 "Unit_Elements">Unit Elements</a></h2> 1325 <p class="dtd"><!ELEMENT units (alias | (unit*, unitLength*, 1326 durationUnit*, special*) ) ><br> 1327 <br> 1328 <!ELEMENT unitLength (alias | (compoundUnit*, unit*, 1329 coordinateUnit*, special*) ) ><br> 1330 <!ATTLIST unitLength type (long | short | narrow) #REQUIRED 1331 ><br> 1332 <br> 1333 <!ELEMENT compoundUnit (alias | (compoundUnitPattern*, 1334 special*) ) ><br> 1335 <!ATTLIST compoundUnit type NMTOKEN #REQUIRED ><br> 1336 <br> 1337 <!ELEMENT unit ( alias | ( gender*, displayName*, unitPattern*, perUnitPattern*, special* ) ) ><br> 1338 <!ATTLIST unit type NMTOKEN #REQUIRED ><br> 1339 <br> 1340 <!ELEMENT gender ( #PCDATA )><br> 1341 <br> 1342 <!ELEMENT durationUnit (alias | (durationUnitPattern*, 1343 special*) ) ><br> 1344 <!ATTLIST durationUnit type NMTOKEN #REQUIRED ><br> 1345 <br> 1346 <!ELEMENT unitPattern ( #PCDATA ) ><br> 1347 <!ATTLIST unitPattern count (0 | 1 | zero | one | two | few 1348 | many | other) #REQUIRED ><br> 1349 <br> 1350 <!ELEMENT compoundUnitPattern ( #PCDATA ) ><br> 1351 <!ATTLIST compoundUnitPattern case NMTOKENS #IMPLIED ><br> 1352 <br> 1353 <!ELEMENT compoundUnitPattern1 ( #PCDATA ) ><br> 1354 <!ATTLIST compoundUnitPattern1 count (0 | 1 | zero | one | two | few | many | other) #IMPLIED > <br> 1355 <!ATTLIST compoundUnitPattern1 gender NMTOKENS #IMPLIED > <br> 1356 <!ATTLIST compoundUnitPattern1 case NMTOKENS #IMPLIED ><br> 1357 <br> 1358 <!ELEMENT coordinateUnit ( alias | ( displayName*, 1359 coordinateUnitPattern*, special* ) ) ><br> 1360 <!ELEMENT coordinateUnitPattern ( #PCDATA ) ><br> 1361 <!ATTLIST coordinateUnitPattern type (north | east | south | 1362 west) #REQUIRED ><br> 1363 <br> 1364 <!ELEMENT durationUnitPattern ( #PCDATA ) ><br></p> 1365 1366 <p>These elements specify the localized way of formatting 1367 quantities of units such as years, months, days, hours, minutes 1368 and seconds— for example, in English, "1 day" or "3 days". The 1369 English rules that produce this example are as follows ({0} 1370 indicates the position of the formatted numeric value):</p> 1371 1372 <pre><unit type="duration-day"> 1373 <displayName>days</displayName> 1374 <unitPattern count="one"><span style= 1375"color: blue">{0} day</span></unitName> 1376 <unitPattern count="other"><span style= 1377"color: blue">{0} days</span></unitName> 1378</unit></pre> 1379 <p>The german rules are more complicated, because German has both gender and case. They thus have additional information, as illustrated below. Note that if there is no @case attribute, for backwards compatibility the implied case is nominative. The possible values for @case are listed in the <strong>grammaticalFeatures</strong> element. These follow the inheritance specified in Part 1, Section <a href= 1380 "tr35.html#Lateral_Inheritance">4.1.2 Lateral 1381 Inheritance</a>. Note that the additional grammar elements are only present in the <unitLength type='long'> form.</p> 1382 <pre><unit type="duration-day"> 1383 <strong><gender>masculine</gender></strong> 1384 <displayName>Tage</displayName> 1385 <unitPattern count="one">{0} Tag</unitPattern> 1386 <strong><unitPattern count="one" case="accusative">{0} Tag</unitPattern> 1387 <unitPattern count="one" case="dative">{0} Tag</unitPattern> 1388 <unitPattern count="one" case="genitive">{0} Tages</unitPattern> 1389</strong> <unitPattern count="other">{0} Tage</unitPattern> 1390 <strong><unitPattern count="other" case="accusative">{0} Tage</unitPattern> 1391 <unitPattern count="other" case="dative">{0} Tagen</unitPattern> 1392 <unitPattern count="other" case="genitive">{0} Tage</unitPattern> 1393</strong> <perUnitPattern>{0} pro Tag</perUnitPattern> 1394</unit></pre> 1395 1396 <p>These follow the inheritance specified in Part 1, Section <a href= 1397 "tr35.html#Lateral_Inheritance">4.1.2 Lateral 1398 Inheritance</a>.In addition to supporting language-specific plural cases 1399 such as “one” and “other”, unitPatterns support the 1400 language-independent explicit cases “0” and “1” for special 1401 handling of numeric values that are exactly 0 or 1; see 1402 <a href="tr35-numbers.html#Explicit_0_1_rules">Explicit 0 and 1 1403 rules</a>.</p> 1404 1405 <p>The <unitPattern> elements may be used to format 1406 quantities with decimal values; in such cases the choice of plural form will 1407 depend not only on the numeric value, but also on its formatting (see 1408 <a href="tr35-numbers.html#Language_Plural_Rules">Language Plural Rules</a>). 1409 In addition to formatting units for stand-alone use, <unitPattern> 1410 elements are increasingly being used to format units for use in running text; 1411 for such usages, the developing <a href="#Grammatical_Features">Grammatical Features</a> 1412 information will be very useful.</p> 1413 1414 <p>Note that for certain plural cases, the unit pattern may not 1415 provide for inclusion of a numeric value—that is, it may not include “{0}”. This 1416 is especially true for the explicit cases “0” and “1” (which may have patterns like 1417 “zero seconds”). In certain languages such as Arabic and Hebrew, this may also be 1418 true with certain units for the plural cases “zero”, “one”, or “two” (in these 1419 languages, such plural cases are only used for the corresponding exact numeric 1420 values, so there is no concern about loss of precision without the numeric value).</p> 1421 1422 <p>Units, like other values with a <strong>count</strong> 1423 attribute, use a special inheritance. See <strong>Part 1: 1424 Core:</strong> <em>Section 4.1 <a href= 1425 "tr35.html#Multiple_Inheritance">Multiple Inheritance</a></em> 1426 .</p> 1427 1428 <p>The displayName is used for labels, such as in a UI. It is 1429 typically lowercased and as neutral a plural form as possible, 1430 and then uses the casing context for the proper display. For 1431 example, for English in a UI it would appear as titlecase:</p> 1432 <p><strong>Duration:</strong></p> 1433 <table style="margin-left: 5em"> 1434 <tr> 1435 <td>Days</td> 1436 <td style="color: silver">enter the vacation length</td> 1437 </tr> 1438 </table><br> 1439 <h3>6.1 <a name="Unit_Preference_and_Conversion" href="#Unit_Preference_and_Conversion">Unit Preference and Conversion Data</a> </h3> 1440 1441 1442<p> 1443Different locales have different preferences for which unit or combination of units is used for a particular usage, such as measuring a person’s height. This is more fine-grained than merely a preference for metric versus US or UK measurement systems. For example, one locale may use meters alone, while another may use centimeters alone or a combination of meters and centimeters; a third may use inches alone, or (informally) a combination of feet and inches. 1444</p> 1445<p> 1446The unit preference and conversion data allows formatting functions to pick the right measurement units for the locale and usage, and convert input measurement into those units. For example, a program (or database) could use 1.88 meters internally, but then for person-height have that measurement convert to <em>6 foot 2 inches</em> for en-US and to <em>188 centimeters </em>for de-CH. Using the unit display names and list formats, those results can then be displayed according to the desired width (eg <em>2″</em> vs <em>2 in</em> vs 2 <em>inches</em>) and using the locale display names and number formats. 1447</p> 1448<p> 1449The size of the measurement can also be taken into account, so that an infant can have a height as <em>18 inches</em>, and an adult the height as <em>6 foot 2 inches.</em> 1450</p> 1451<p> 1452This data is supplied in <strong>Part 6: <a href= 1453 "tr35-info.html#Contents">Supplemental</a></strong>: <a href= 1454 "tr35-info.html#Unit_Conversion">Section 13 Unit Conversion</a> and <a href= 1455 "tr35-info.html#Unit_Preferences">Section 13 Unit Preferences</a>. 1456</p> 1457<h3>6.2 <a name="Unit_Identifiers" href="#Unit_Identifiers" >Unit Identifiers</a></h3> 1458<p> 1459 Units are identified internally as described in this section. As with other identifiers in CLDR, the American English spelling is used for unit identifiers. These names should not be presented to end users, however. As in other cases, the translated names for different languages (or variants of English) are available in the CLDR localized data. 1460</p> 1461 1462<table> 1463 <tr> 1464 <td><strong>Name</strong> 1465 </td> 1466 <td><strong>Example</strong> 1467 </td> 1468 </tr> 1469 <tr> 1470 <td>long unit identifier 1471 </td> 1472 <td>length-meter, mass-pound, duration-day 1473 </td> 1474 </tr> 1475 <tr> 1476 <td>core unit identifier 1477 </td> 1478 <td>meter, pound, day 1479 </td> 1480 </tr> 1481</table><br> 1482 1483 1484<p> 1485Both the <em>unit identifier</em> and the <em>core unit identifier</em> are guaranteed to be unique, and clients can use either one to identify a unit. The associations between types and core unit identifiers are as prescribed in CLDR data; it is invalid for a client to create any additional associations. Except as specified in <em>Section 6.6 <a href="https://www.unicode.org/reports/tr35/tr35-general.html#Private_Use_Units">Private-Use Units</a></em>, all values are reserved by CLDR. 1486</p> 1487 1488<table> 1489 <tr> 1490 <td><strong>Name</strong> 1491 </td> 1492 <td><strong>Examples</strong> 1493 </td> 1494 </tr> 1495 <tr> 1496 <td>simple unit ID 1497 </td> 1498 <td>meter, foot, inch, pound, pound-force, … 1499 </td> 1500 </tr> 1501 <tr> 1502 <td>prefixed unit ID 1503 </td> 1504 <td>kilometer, centigram, … 1505<p> 1506<em>plus simple unit IDs</em> 1507 </td> 1508 </tr> 1509 <tr> 1510 <td>single unit ID 1511 </td> 1512 <td>square-foot, cubic-centimeter, …<br> 1513<em>plus prefixed unit IDs</em> 1514 </td> 1515 </tr> 1516 <tr> 1517 <td>core unit ID 1518 </td> 1519 <td>kilometer-per-hour, kilogram-meter, kilogram-meter-per-square-second, … 1520<p> 1521<em>plus single unit IDs</em> 1522 </td> 1523 </tr> 1524 <tr> 1525 <td>mixed unit ID 1526 </td> 1527 <td>foot-and-pound 1528 </td> 1529 </tr> 1530</table><br> 1531 1532 1533<p> 1534A core unit that is not a simple unit is called a <em>complex unit</em>. It is valid to construct a complex unit identifier from multiple simple unit identifiers using multiplication (kilogram-meter) and division (kilogram-per-meter). As usual, with division the part before the (first) -per- is called the <em>numerator</em>, and the part after it is called the <em>denominator</em>. 1535</p> 1536<p> 1537The conversion information uses the short unit identifiers, discarding the unitType. Thus “meter” is used instead of “length-meter”. The translation data currently uses the long unit identifiers, for backwards compatibility. However, that is likely to change in a future version. 1538</p> 1539<p> 1540The identifiers and unit conversion data are built to handle arbitrary combinations of core unit IDs using division (kilometer-per-hour), multiplication (kilogram-meter), powers (square-second), and SI prefixes (kilo-). Thus they support converting generated units such as inch-pound-per-square-week into comparable units, such as newtons. 1541</p><ul> 1542 1543<li>A power (square, cubic, pow4, etc) modifies one prefixed unit ID, and must occur immediately before it in the identifier: square-foot, not foot-square. 1544<li>Multiplication binds more tightly than division, so kilogram-meter-per-second-ampere is interpreted as (kg ⋅ m) / (s ⋅ a). 1545<li>Thus if -per- occurs multiple times, each occurrence after the first is equivalent to a multiplication: <ul> 1546 1547 <li>kilogram-meter-per-second-ampere ⩧ kilogram-meter-per-second-per-ampere. 1548</ul> 1549</li></ul> 1550 1551<h4>Nomenclature</h4> 1552 1553 1554<p> 1555For the US spelling, see the <a href="https://www.nist.gov/pml/special-publication-811">Preface of the Guide for the Use of the International System of Units (SI), NIST special publication 811</a>, which is explicit about the discrepancy with the English-language BIPM spellings: 1556</p> 1557<p> 1558 1559 In keeping with U.S. and International practice (see Sec. C.2), this Guide uses the dot on the line as the decimal marker. In addition this Guide utilizes the American spellings “meter,” “liter,” and “deka” rather than “metre,” “litre,” and “deca,” and the name “metric ton” rather than “tonne.” 1560</p> 1561<h4>Syntax</h4> 1562 1563 1564<p> 1565The formal syntax for identifiers is provided below. 1566</p> 1567 1568<table> 1569 <tr> 1570 <td>unit_identifier 1571 </td> 1572 <td>:= 1573 </td> 1574 <td>long_unit_identifier<br> 1575| mixed_unit_identifier 1576<p> 1577| core_unit_identifier 1578 </td> 1579 </tr> 1580 <tr> 1581 <td>long_unit_identifier 1582 </td> 1583 <td>:= 1584 </td> 1585 <td>type "-" core_unit_identifier 1586 </td> 1587 </tr> 1588 <tr> 1589 <td>core_unit_identifier 1590 </td> 1591 <td>:= 1592 </td> 1593 <td>product_unit ("-per-" product_unit)*<br>| "per-" product_unit ("-per-" product_unit)* 1594 <ul> 1595 <li><em>Examples:</em> 1596 <ul> 1597 <li>foot-per-second-per-second</li> 1598 <li>per-second</li> 1599 </ul> 1600 </li> 1601 <li><em>Note: </em>The normalized form will have only one "per"</li> 1602 </ul> 1603 </td> 1604 </tr> 1605 <tr> 1606 <td>mixed_unit_identifier 1607 </td> 1608 <td>:= 1609 </td> 1610 <td>(single_unit | pu_single_unit) ("-and-" (single_unit | pu_single_unit ))*<ul> 1611 1612<li><em>Example: foot-and-inch</em> 1613 </ul> 1614 1615 </td> 1616 </tr> 1617 <tr> 1618 <td>product_unit 1619 </td> 1620 <td>:= 1621 </td> 1622 <td>single_unit ("-" single_unit)* ("-" pu_single_unit)* 1623<p> 1624| pu_single_unit ("-" pu_single_unit)*<ul> 1625 1626<li><em>Example: </em>foot-pound-force 1627<li><em>Constraint:</em> No pu_single_unit may precede a single unit</li></ul> 1628 1629 </td> 1630 </tr> 1631 <tr> 1632 <td>single_unit 1633 </td> 1634 <td>:= 1635 </td> 1636 <td>(dimensionality_prefix)? prefixed_unit<ul> 1637 1638<li><em>Example: </em>square-meter</li></ul> 1639 1640 </td> 1641 </tr> 1642 <tr> 1643 <td>pu_single_unit 1644 </td> 1645 <td>:= 1646 </td> 1647 <td>“xxx-” single-unit | “x-” single-unit<ul> 1648 1649<li><em>Example: </em>xxx-square-knuts (a Harry Potter unit) 1650<li><em>Note: </em>“x-” is only for backwards compatibility 1651<li>See Section 6.6 <a href="#Private_Use_Units">Private-Use Units</a> 1652 </ul></td> 1653 </tr> 1654 <tr> 1655 <td>dimensionality_prefix 1656 </td> 1657 <td>:= 1658 </td> 1659 <td>"square-" 1660<p> 1661| "cubic-" 1662<p> 1663| "pow" ([2-9]|1[0-5]) "-"<ul> 1664 1665<li><em>Note: </em>"pow2-" and "pow3-" canonicalize to "square-" and "cubic-"</li></ul> 1666 1667 </td> 1668 </tr> 1669 <tr> 1670 <td>prefixed_unit 1671 </td> 1672 <td> 1673 </td> 1674 <td>(si_prefix)? simple_unit<ul> 1675 1676<li><em>Example: </em>kilometer</li></ul> 1677 1678 </td> 1679 </tr> 1680 <tr> 1681 <td>si_prefix 1682 </td> 1683 <td>:= 1684 </td> 1685 <td>"deka" | "hecto" | "kilo", … <ul> 1686 1687<li><em>Note: </em>See full list at <a href="https://www.nist.gov/pml/special-publication-811">NIST special publication 811</a></li></ul> 1688 1689 </td> 1690 </tr> 1691 <tr> 1692 <td>simple_unit 1693 </td> 1694 <td>:= 1695 </td> 1696 <td>unit_component ("-" unit_component)* 1697<p> 1698 | “em” | “g” | “us” | “hg” | “of”<ul> 1699 1700 <li><em>Example: </em>gallon-imperial</li> 1701<li><em>Constraint: </em>At least one unit_component must not itself be a simple_unit 1702<li><em>Note: </em>3 simple units are currently allowed as legacy usage, 1703where a component wouldn’t be a unit_component (eg for “<strong>g</strong>-force”) 1704 1705 <ul> 1706 1707 <li>We will likely deprecate those and add conformant aliases in the future. 1708 <li>“hg” and “of” are already only in deprecated simple_units. 1709 </ul> 1710</li> 1711 </ul> 1712 1713 </td> 1714 </tr> 1715 <tr> 1716 <td>unit_component 1717 </td> 1718 <td>:= 1719 </td> 1720 <td>[a-z]{3,∞} | “1” “0”{2,3}<ul> 1721 1722<li><em>Constraints: </em> 1723 <ul><li>Cannot be "per", "and", "square", "cubic", "xxx", or "x"; or start with an SI prefix.</li> 1724<li>While the syntax allows any number of letters greater than 3, the unit_components need to be distinct if truncated to 8 letters. This allows for possible future support of units in Unicode Locale Identifiers.</li></ul> 1725<li><em>Example: </em>foot</li></ul> 1726 1727 </td> 1728 </tr> 1729</table> 1730 1731 1732 <h3>6.3 <a name="Example_Units" href="#Example_Units">Example Units</a></h3> 1733 <p>The 1734 following table contains examples of types and units currently 1735 defined by CLDR. The units in CLDR are not 1736 comprehensive; it is anticipated that more will be added over 1737 time. The complete list of supported units is in the validity 1738 data: see <em>Section <a href="tr35.html#Validity_Data">3.11 1739 Validity Data</a></em>. The compound 1740 units in the table below either require specialized formatting 1741 or have a numerator and/or demoninator that are not defined as 1742 valid standalone units. Note: as explained in <em>Section 6.4 1743 <a href="#compound-units">Compound Units</a></em>, CLDR 1744 provides data to format any compound unit composed of two 1745 simple units from the following table.</p> 1746 <table> 1747 <tr> 1748 <th>Type</th> 1749 <th>Core Unit Identifier</th> 1750 <th>Compound?</th> 1751 <th>Sample Format</th> 1752 </tr> 1753 <tr> 1754 <td><em>acceleration</em></td> 1755 <td>g-force</td> 1756 <td>simple</td> 1757 <td>{0} G</td> 1758 </tr> 1759 <tr> 1760 <td><em>acceleration</em></td> 1761 <td>meter-per-square-second</td> 1762 <td>compound</td> 1763 <td>{0} m/s²</td> 1764 </tr> 1765 <tr> 1766 <td><em>angle</em></td> 1767 <td>revolution</td> 1768 <td>simple</td> 1769 <td>{0} rev</td> 1770 </tr> 1771 <tr> 1772 <td><em>angle</em></td> 1773 <td>radian</td> 1774 <td>simple</td> 1775 <td>{0} rad</td> 1776 </tr> 1777 <tr> 1778 <td><em>angle</em></td> 1779 <td>degree</td> 1780 <td>simple</td> 1781 <td>{0}°</td> 1782 </tr> 1783 <tr> 1784 <td><em>angle</em></td> 1785 <td>arc-minute</td> 1786 <td>simple</td> 1787 <td>{0}′</td> 1788 </tr> 1789 <tr> 1790 <td><em>angle</em></td> 1791 <td>arc-second</td> 1792 <td>simple</td> 1793 <td>{0}″</td> 1794 </tr> 1795 <tr> 1796 <td><em>area</em></td> 1797 <td>square-kilometer</td> 1798 <td>simple</td> 1799 <td>{0} km²</td> 1800 </tr> 1801 <tr> 1802 <td><em>area</em></td> 1803 <td>hectare</td> 1804 <td>simple</td> 1805 <td>{0} ha</td> 1806 </tr> 1807 <tr> 1808 <td>...</td> 1809 <td>...</td> 1810 <td>...</td> 1811 <td>...</td> 1812 </tr> 1813 <tr> 1814 <td><em>area</em></td> 1815 <td>square-inch</td> 1816 <td>simple</td> 1817 <td>{0} in²</td> 1818 </tr> 1819 <tr> 1820 <td><em>area</em></td> 1821 <td>dunam</td> 1822 <td>simple</td> 1823 <td>{0} dunam</td> 1824 </tr> 1825 <tr> 1826 <td><em>concentr</em></td> 1827 <td>karat</td> 1828 <td>simple</td> 1829 <td>{0} kt</td> 1830 <td>dimensionless</td> 1831 </tr> 1832 <tr> 1833 <td><em>concentr</em></td> 1834 <td>milligram-per-deciliter</td> 1835 <td>compound</td> 1836 <td>{0} mg/dL</td> 1837 </tr> 1838 <tr> 1839 <td><em>concentr</em></td> 1840 <td>millimole-per-liter</td> 1841 <td>compound</td> 1842 <td>{0} mmol/L</td> 1843 </tr> 1844 <tr> 1845 <td><em>concentr</em></td> 1846 <td>permillion</td> 1847 <td>compound</td> 1848 <td>{0} ppm</td> 1849 <td>dimensionless</td> 1850 </tr> 1851 <tr> 1852 <td><em>concentr</em></td> 1853 <td>percent</td> 1854 <td>simple</td> 1855 <td>{0}%</td> 1856 <td>dimensionless</td> 1857 </tr> 1858 <tr> 1859 <td><em>concentr</em></td> 1860 <td>permille</td> 1861 <td>simple</td> 1862 <td>{0}‰</td> 1863 <td>dimensionless</td> 1864 </tr> 1865 <tr> 1866 <td><em>concentr</em></td> 1867 <td>permyriad</td> 1868 <td>simple</td> 1869 <td>{0}‱</td> 1870 <td>dimensionless</td> 1871 </tr> 1872 <tr> 1873 <td><em>concentr</em></td> 1874 <td>mole</td> 1875 <td>simple</td> 1876 <td>{0} mol</td> 1877 <td>dimensionless</td> 1878 </tr> 1879 <tr> 1880 <td><em>consumption</em></td> 1881 <td>liter-per-kilometer</td> 1882 <td>compound</td> 1883 <td>{0} L/km</td> 1884 </tr> 1885 <tr> 1886 <td><em>consumption</em></td> 1887 <td>liter-per-100-kilometer</td> 1888 <td>compound</td> 1889 <td>{0} L/100km</td> 1890 </tr> 1891 <tr> 1892 <td><em>consumption</em></td> 1893 <td>mile-per-gallon (US)</td> 1894 <td>compound</td> 1895 <td>{0} mpg</td> 1896 </tr> 1897 <tr> 1898 <td><em>consumption</em></td> 1899 <td>mile-per-gallon-imperial</td> 1900 <td>compound</td> 1901 <td>{0} mpg Imp.</td> 1902 </tr> 1903 <tr> 1904 <td><em>digital</em></td> 1905 <td>petabyte</td> 1906 <td>simple</td> 1907 <td>{0} PB</td> 1908 </tr> 1909 <tr> 1910 <td>...</td> 1911 <td>...</td> 1912 <td>...</td> 1913 <td>...</td> 1914 </tr> 1915 <tr> 1916 <td><em>digital</em></td> 1917 <td>byte</td> 1918 <td>simple</td> 1919 <td>{0} byte</td> 1920 </tr> 1921 <tr> 1922 <td><em>digital</em></td> 1923 <td>bit</td> 1924 <td>simple</td> 1925 <td>{0} bit</td> 1926 </tr> 1927 <tr> 1928 <td><em>duration</em></td> 1929 <td>century</td> 1930 <td>simple</td> 1931 <td>{0} c</td> 1932 </tr> 1933 <tr> 1934 <td><em>duration</em></td> 1935 <td>year</td> 1936 <td>simple</td> 1937 <td>{0} y</td> 1938 </tr> 1939 <tr> 1940 <td><em>duration</em></td> 1941 <td>year-person</td> 1942 <td>simple</td> 1943 <td>{0} y</td> 1944 <td>for duration or age related to a person</td> 1945 </tr> 1946 <tr> 1947 <td><em>duration</em></td> 1948 <td>month</td> 1949 <td>simple</td> 1950 <td>{0} m</td> 1951 </tr> 1952 <tr> 1953 <td><em>duration</em></td> 1954 <td>month-person</td> 1955 <td>simple</td> 1956 <td>{0} m</td> 1957 <td>for duration or age related to a person</td> 1958 </tr> 1959 <tr> 1960 <td><em>duration</em></td> 1961 <td>week</td> 1962 <td>simple</td> 1963 <td>{0} w</td> 1964 </tr> 1965 <tr> 1966 <td><em>duration</em></td> 1967 <td>week-person</td> 1968 <td>simple</td> 1969 <td>{0} w</td> 1970 <td>for duration or age related to a person</td> 1971 </tr> 1972 <tr> 1973 <td><em>duration</em></td> 1974 <td>day</td> 1975 <td>simple</td> 1976 <td>{0} d</td> 1977 </tr> 1978 <tr> 1979 <td><em>duration</em></td> 1980 <td>day-person</td> 1981 <td>simple</td> 1982 <td>{0} d</td> 1983 <td>for duration or age related to a person</td> 1984 </tr> 1985 <tr> 1986 <td><em>duration</em></td> 1987 <td>hour</td> 1988 <td>simple</td> 1989 <td>{0} h</td> 1990 </tr> 1991 <tr> 1992 <td>...</td> 1993 <td>...</td> 1994 <td>...</td> 1995 <td>...</td> 1996 </tr> 1997 <tr> 1998 <td><em>duration</em></td> 1999 <td>nanosecond</td> 2000 <td>simple</td> 2001 <td>{0} ns</td> 2002 </tr> 2003 <tr> 2004 <td><em>electric</em></td> 2005 <td>ampere</td> 2006 <td>simple</td> 2007 <td>{0} A</td> 2008 </tr> 2009 <tr> 2010 <td><em>electric</em></td> 2011 <td>milliampere</td> 2012 <td>simple</td> 2013 <td>{0} mA</td> 2014 </tr> 2015 <tr> 2016 <td><em>electric</em></td> 2017 <td>ohm</td> 2018 <td>simple</td> 2019 <td>{0} Ω</td> 2020 </tr> 2021 <tr> 2022 <td><em>electric</em></td> 2023 <td>volt</td> 2024 <td>simple</td> 2025 <td>{0} V</td> 2026 </tr> 2027 <tr> 2028 <td><em>energy</em></td> 2029 <td>kilocalorie</td> 2030 <td>simple</td> 2031 <td>{0} kcal</td> 2032 </tr> 2033 <tr> 2034 <td><em>energy</em></td> 2035 <td>calorie</td> 2036 <td>simple</td> 2037 <td>{0} cal</td> 2038 </tr> 2039 <tr> 2040 <td><em>energy</em></td> 2041 <td>foodcalorie</td> 2042 <td>simple</td> 2043 <td>{0} Cal</td> 2044 </tr> 2045 <tr> 2046 <td><em>energy</em></td> 2047 <td>kilojoule</td> 2048 <td>simple</td> 2049 <td>{0} kJ</td> 2050 </tr> 2051 <tr> 2052 <td><em>energy</em></td> 2053 <td>joule</td> 2054 <td>simple</td> 2055 <td>{0} J</td> 2056 </tr> 2057 <tr> 2058 <td><em>energy</em></td> 2059 <td>kilowatt-hour</td> 2060 <td>simple</td> 2061 <td>{0} kWh</td> 2062 </tr> 2063 <tr> 2064 <td><em>energy</em></td> 2065 <td>electronvolt</td> 2066 <td>simple</td> 2067 <td>{0} eV</td> 2068 </tr> 2069 <tr> 2070 <td><em>energy</em></td> 2071 <td>british-thermal-unit</td> 2072 <td>simple</td> 2073 <td>{0} Btu</td> 2074 </tr> 2075 <tr> 2076 <td><em>force</em></td> 2077 <td>pound-force</td> 2078 <td>simple</td> 2079 <td>{0} lbf</td> 2080 </tr> 2081 <tr> 2082 <td><em>force</em></td> 2083 <td>newton</td> 2084 <td>simple</td> 2085 <td>{0} N</td> 2086 </tr> 2087 <tr> 2088 <td><em>frequency</em></td> 2089 <td>gigahertz</td> 2090 <td>simple</td> 2091 <td>{0} GHz</td> 2092 </tr> 2093 <tr> 2094 <td><em>frequency</em></td> 2095 <td>megahertz</td> 2096 <td>simple</td> 2097 <td>{0} MHz</td> 2098 </tr> 2099 <tr> 2100 <td><em>frequency</em></td> 2101 <td>kilohertz</td> 2102 <td>simple</td> 2103 <td>{0} kHz</td> 2104 </tr> 2105 <tr> 2106 <td><em>frequency</em></td> 2107 <td>hertz</td> 2108 <td>simple</td> 2109 <td>{0} Hz</td> 2110 </tr> 2111 <tr> 2112 <td><em>length</em></td> 2113 <td>kilometer</td> 2114 <td>simple</td> 2115 <td>{0} km</td> 2116 </tr> 2117 <tr> 2118 <td>...</td> 2119 <td>...</td> 2120 <td>...</td> 2121 <td>...</td> 2122 </tr> 2123 <tr> 2124 <td><em>length</em></td> 2125 <td>inch</td> 2126 <td>simple</td> 2127 <td>{0} in</td> 2128 </tr> 2129 <tr> 2130 <td><em>length</em></td> 2131 <td>parsec</td> 2132 <td>simple</td> 2133 <td>{0} pc</td> 2134 </tr> 2135 <tr> 2136 <td><em>length</em></td> 2137 <td>light-year</td> 2138 <td>simple</td> 2139 <td>{0} ly</td> 2140 </tr> 2141 <tr> 2142 <td><em>length</em></td> 2143 <td>astronomical-unit</td> 2144 <td>simple</td> 2145 <td>{0} au</td> 2146 </tr> 2147 <tr> 2148 <td><em>length</em></td> 2149 <td>furlong</td> 2150 <td>simple</td> 2151 <td>{0} fur</td> 2152 </tr> 2153 <tr> 2154 <td><em>length</em></td> 2155 <td>fathom</td> 2156 <td>simple</td> 2157 <td>{0} fm</td> 2158 </tr> 2159 <tr> 2160 <td><em>length</em></td> 2161 <td>nautical-mile</td> 2162 <td>simple</td> 2163 <td>{0} nmi</td> 2164 </tr> 2165 <tr> 2166 <td><em>length</em></td> 2167 <td>mile-scandinavian</td> 2168 <td>simple</td> 2169 <td>{0} smi</td> 2170 </tr> 2171 <tr> 2172 <td><em>length</em></td> 2173 <td>point</td> 2174 <td>simple</td> 2175 <td>{0} pt</td> 2176 <td>typographic point, 1/72 inch</td> 2177 </tr> 2178 <tr> 2179 <td><em>length</em></td> 2180 <td>solar-radius</td> 2181 <td>simple</td> 2182 <td>{0} R☉</td> 2183 </tr> 2184 <tr> 2185 <td><em>light</em></td> 2186 <td>lux</td> 2187 <td>simple</td> 2188 <td>{0} lx</td> 2189 </tr> 2190 <tr> 2191 <td><em>light</em></td> 2192 <td>solar-luminosity</td> 2193 <td>simple</td> 2194 <td>{0} L☉</td> 2195 </tr> 2196 <tr> 2197 <td><em>mass</em></td> 2198 <td>metric-ton</td> 2199 <td>simple</td> 2200 <td>{0} t</td> 2201 </tr> 2202 <tr> 2203 <td><em>mass</em></td> 2204 <td>kilogram</td> 2205 <td>simple</td> 2206 <td>{0} kg</td> 2207 </tr> 2208 <tr> 2209 <td>...</td> 2210 <td>...</td> 2211 <td>...</td> 2212 <td>...</td> 2213 </tr> 2214 <tr> 2215 <td><em>mass</em></td> 2216 <td>ounce</td> 2217 <td>simple</td> 2218 <td>{0} oz</td> 2219 </tr> 2220 <tr> 2221 <td><em>mass</em></td> 2222 <td>ounce-troy</td> 2223 <td>simple</td> 2224 <td>{0} oz t</td> 2225 </tr> 2226 <tr> 2227 <td><em>mass</em></td> 2228 <td>carat</td> 2229 <td>simple</td> 2230 <td>{0} CD</td> 2231 </tr> 2232 <tr> 2233 <td><em>mass</em></td> 2234 <td>dalton</td> 2235 <td>simple</td> 2236 <td>{0} Da</td> 2237 </tr> 2238 <tr> 2239 <td><em>mass</em></td> 2240 <td>earth-mass</td> 2241 <td>simple</td> 2242 <td>{0} M⊕</td> 2243 </tr> 2244 <tr> 2245 <td><em>mass</em></td> 2246 <td>solar-mass</td> 2247 <td>simple</td> 2248 <td>{0} M☉</td> 2249 </tr> 2250 <tr> 2251 <td><em>power</em></td> 2252 <td>gigawatt</td> 2253 <td>simple</td> 2254 <td>{0} GW</td> 2255 </tr> 2256 <tr> 2257 <td>...</td> 2258 <td>...</td> 2259 <td>...</td> 2260 <td>...</td> 2261 </tr> 2262 <tr> 2263 <td><em>power</em></td> 2264 <td>milliwatt</td> 2265 <td>simple</td> 2266 <td>{0} mW</td> 2267 </tr> 2268 <tr> 2269 <td><em>power</em></td> 2270 <td>horsepower</td> 2271 <td>simple</td> 2272 <td>{0} hp</td> 2273 </tr> 2274 <tr> 2275 <td><em>pressure</em></td> 2276 <td>hectopascal</td> 2277 <td>simple</td> 2278 <td>{0} hPa</td> 2279 </tr> 2280 <tr> 2281 <td><em>pressure</em></td> 2282 <td>millimeter-ofhg</td> 2283 <td>simple</td> 2284 <td>{0} mm Hg</td> 2285 </tr> 2286 <tr> 2287 <td><em>pressure</em></td> 2288 <td>pound-force-per-square-inch</td> 2289 <td>compound</td> 2290 <td>{0} psi</td> 2291 </tr> 2292 <tr> 2293 <td><em>pressure</em></td> 2294 <td>inch-ofhg</td> 2295 <td>simple</td> 2296 <td>{0} inHg</td> 2297 </tr> 2298 <tr> 2299 <td><em>pressure</em></td> 2300 <td>millibar</td> 2301 <td>simple</td> 2302 <td>{0} mbar</td> 2303 </tr> 2304 <tr> 2305 <td><em>pressure</em></td> 2306 <td>atmosphere</td> 2307 <td>simple</td> 2308 <td>{0} atm</td> 2309 </tr> 2310 <tr> 2311 <td><em>pressure</em></td> 2312 <td>kilopascal</td> 2313 <td>simple</td> 2314 <td>{0} kPa</td> 2315 </tr> 2316 <tr> 2317 <td><em>pressure</em></td> 2318 <td>megapascal</td> 2319 <td>simple</td> 2320 <td>{0} MPa</td> 2321 </tr> 2322 <tr> 2323 <td><em>speed</em></td> 2324 <td>kilometer-per-hour</td> 2325 <td>compound</td> 2326 <td>{0} km/h</td> 2327 </tr> 2328 <tr> 2329 <td><em>speed</em></td> 2330 <td>meter-per-second</td> 2331 <td>compound</td> 2332 <td>{0} m/s</td> 2333 </tr> 2334 <tr> 2335 <td><em>speed</em></td> 2336 <td>mile-per-hour</td> 2337 <td>compound</td> 2338 <td>{0} mi/h</td> 2339 </tr> 2340 <tr> 2341 <td><em>speed</em></td> 2342 <td>knot</td> 2343 <td>simple</td> 2344 <td>{0} kn</td> 2345 </tr> 2346 <tr> 2347 <td><em>temperature</em></td> 2348 <td>generic</td> 2349 <td>simple</td> 2350 <td>{0}°</td> 2351 </tr> 2352 <tr> 2353 <td><em>temperature</em></td> 2354 <td>celsius</td> 2355 <td>simple</td> 2356 <td>{0}°C</td> 2357 </tr> 2358 <tr> 2359 <td><em>temperature</em></td> 2360 <td>fahrenheit</td> 2361 <td>simple</td> 2362 <td>{0}°F</td> 2363 </tr> 2364 <tr> 2365 <td><em>temperature</em></td> 2366 <td>kelvin</td> 2367 <td>simple</td> 2368 <td>{0} K</td> 2369 </tr> 2370 <tr> 2371 <td><em>torque</em></td> 2372 <td>pound-force-foot</td> 2373 <td>simple</td> 2374 <td>{0} lbf⋅ft</td> 2375 </tr> 2376 <tr> 2377 <td><em>torque</em></td> 2378 <td>newton-meter</td> 2379 <td>simple</td> 2380 <td>{0} N⋅m</td> 2381 </tr> 2382 <tr> 2383 <td><em>volume</em></td> 2384 <td>cubic-kilometer</td> 2385 <td>simple</td> 2386 <td>{0} km³</td> 2387 </tr> 2388 <tr> 2389 <td>...</td> 2390 <td>...</td> 2391 <td>...</td> 2392 <td>...</td> 2393 </tr> 2394 <tr> 2395 <td><em>volume</em></td> 2396 <td>cubic-inch</td> 2397 <td>simple</td> 2398 <td>{0} in³</td> 2399 </tr> 2400 <tr> 2401 <td><em>volume</em></td> 2402 <td>megaliter</td> 2403 <td>simple</td> 2404 <td>{0} ML</td> 2405 </tr> 2406 <tr> 2407 <td>...</td> 2408 <td>...</td> 2409 <td>...</td> 2410 <td>...</td> 2411 </tr> 2412 <tr> 2413 <td><em>volume</em></td> 2414 <td>pint</td> 2415 <td>simple</td> 2416 <td>{0} pt</td> 2417 </tr> 2418 <tr> 2419 <td><em>volume</em></td> 2420 <td>cup</td> 2421 <td>simple</td> 2422 <td>{0} c</td> 2423 </tr> 2424 <tr> 2425 <td><em>volume</em></td> 2426 <td>fluid-ounce (US)</td> 2427 <td>simple</td> 2428 <td>{0} fl oz</td> 2429 </tr> 2430 <tr> 2431 <td><em>volume</em></td> 2432 <td>fluid-ounce-imperial</td> 2433 <td>simple</td> 2434 <td>{0} fl oz Imp.</td> 2435 </tr> 2436 <tr> 2437 <td><em>volume</em></td> 2438 <td>tablespoon</td> 2439 <td>simple</td> 2440 <td>{0} tbsp</td> 2441 </tr> 2442 <tr> 2443 <td><em>volume</em></td> 2444 <td>teaspoon</td> 2445 <td>simple</td> 2446 <td>{0} tsp</td> 2447 </tr> 2448 <tr> 2449 <td><em>volume</em></td> 2450 <td>barrel</td> 2451 <td>simple</td> 2452 <td>{0} bbl</td> 2453 </tr> 2454 </table><br> 2455 <p>There are three widths: <strong>long</strong>, 2456 <strong>short</strong>, and <strong>narrow</strong>. As usual, 2457 the narrow forms may not be unique: in English, 1′ could mean 1 2458 minute of arc, or 1 foot. Thus narrow forms should only be used 2459 where the context makes the meaning clear.</p> 2460 <p>Where the unit of measurement is one of the <a href= 2461 "https://physics.nist.gov/cuu/Units/units.html">International 2462 System of Units (SI)</a>, the short and narrow forms will 2463 typically use the international symbols, such as “mm” for 2464 millimeter. They may, however, be different if that is 2465 customary for the language or locale. For example, in Russian 2466 it may be more typical to see the Cyrillic characters “мм”.</p> 2467 <p>Units are included for translation even where they are not 2468 typically used in a particular locale, such as kilometers in 2469 the US, or inches in Germany. This is to account for use by 2470 travelers and specialized domains, such as the German 2471 “̌Fernseher von 32 bis 55 Zoll (80 bis 140 cm)” for TV screen 2472 size in inches and centimeters.</p> 2473 <p>For temperature, there is a special unit <unit 2474 type="temperature-generic">, which is used when it is clear 2475 from context whether Celcius or Fahrenheit is implied.</p> 2476 <p>For duration, there are special units such as <unit 2477 type="duration-year-person"> and <unit 2478 type="duration-year-week"> for indicating the age of a 2479 person, which requires special forms in some languages. For 2480 example, in "zh", references to a person being 3 days old or 30 2481 years old would use the forms “他3天大” and “他30岁” 2482 respectively.</p> 2483 <h3>6.4 <a name="compound-units" href="#compound-units" >Compound Units</a><a name= 2484 "compoundUnitPattern" ></a><a name="perUnitPatterns" ></a></h3> 2485 <p>A common combination of units is X per Y, such as <em>miles 2486 per hour</em> or <em>liters per second</em> or <em>kilowatt-hours</em>. </p> 2487 2488 <p>There are different types of structure used to build the localized name of compound units. All of these follow the inheritance specified in <a href= 2489 "tr35.html#Lateral_Inheritance">Part 1, Section 4.1.2 Lateral 2490 Inheritance</a>.</p> 2491 <p><strong>Prefixes</strong> are for powers of 10 and powers of 1024 (the latter only used with digital units of measure). These are invariant for case, gender, or plural (though those could be added in the future if needed by a language).</p> 2492 <pre><compoundUnit type="10p9"><br> <unitPrefixPattern>Giga{0}</unitPrefixPattern><br></compoundUnit> 2493 2494<compoundUnit type="1024p3"><br> <unitPrefixPattern>Gibi{0}</unitPrefixPattern><br></compoundUnit> 2495</pre> 2496 <p><strong>compoundUnitPatterns</strong> are used for compounding units by multiplication or division: kilowatt-hours, or meters per second. These are invariant for case, gender, or plural (though those could be added in the future if needed by a language).</p> 2497 <pre> 2498<compoundUnit type="per"><br> <compoundUnitPattern>{0} pro {1}</compoundUnitPattern><br></compoundUnit> 2499 2500<compoundUnit type="times"><br> <compoundUnitPattern>{0}⋅{1}</compoundUnitPattern><br></compoundUnit> 2501</pre><p>There can be at most one "per" pattern used in producing a compound unit, while the "times" pattern can be used multiple times.</p> 2502 <p><strong>compoundUnitPattern1s</strong> are used for expressing powers, such as square meter or cubic foot. These are the most complicated, since they can vary by plural category (count), by case, and by gender. However, these extra attributes are only used if the are present in the <strong>grammaticalFeatures</strong> element for the language in question. See <a href="tr35-general.html#Grammatical_Features">Section 15, Grammatical Features</a>. Note that the additional grammar elements are only present in the <unitLength type='long'> form. </p> 2503 <pre> 2504<compoundUnit type="power2"><br> <compoundUnitPattern1>{0} kw.</compoundUnitPattern1><br> <compoundUnitPattern1 count="one">{0} kwadratowe</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" case="accusative">{0} kwadratowe</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" case="dative">{0} kwadratowemu</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" case="genitive">{0} kwadratowego</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" case="instrumental">{0} kwadratowym</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" case="locative">{0} kwadratowym</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" case="vocative">{0} kwadratowe</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" gender="feminine">{0} kwadratowa</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" gender="feminine" case="accusative">{0} kwadratową</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" gender="feminine" case="dative">{0} kwadratowej</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" gender="feminine" case="genitive">{0} kwadratowej</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" gender="feminine" case="instrumental">{0} kwadratową</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" gender="feminine" case="locative">{0} kwadratowej</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" gender="feminine" case="vocative">{0} kwadratowa</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" gender="inanimate">{0} kwadratowy</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" gender="inanimate" case="accusative">{0} kwadratowy</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" gender="inanimate" case="dative">{0} kwadratowemu</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" gender="inanimate" case="genitive">{0} kwadratowego</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" gender="inanimate" case="instrumental">{0} kwadratowym</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" gender="inanimate" case="locative">{0} kwadratowym</compoundUnitPattern1><br> <compoundUnitPattern1 count="one" gender="inanimate" case="vocative">{0} kwadratowy</compoundUnitPattern1><br> <compoundUnitPattern1 count="few">{0} kwadratowe</compoundUnitPattern1><br> <compoundUnitPattern1 count="few" case="accusative">{0} kwadratowe</compoundUnitPattern1><br> <compoundUnitPattern1 count="few" case="dative">{0} kwadratowym</compoundUnitPattern1> 2505 …</pre> 2506 2507 <p> </p> 2508 <p>Some units already 2509 have 'precomputed' forms, such as <strong>kilometer-per-hour</strong>; where such units exist, 2510 they should be used in preference. </p> 2511 2512 <p>If there is no precomputed form, the following process in pseudocode is used to generate a pattern for the compound unit. </p> 2513 <p> </p> 2514 <p><strong>pattern(unitId, locale, length, pluralCategory, caseVariant)</strong></p> 2515 <ol> 2516 <li>If the unitId is empty or invalid, fail</li> 2517 <li>Put the unitId into normalized order: hour-kilowatt => kilowatt-hour, meter-square-meter-per-second-second => cubic-meter-per-square-second</li> 2518 <li>Set result to be getValue(unitId with length, pluralCategory, caseVariant) 2519 <ol> 2520 <li>If result is not empty, return it</li> 2521 </ol> 2522 </li> 2523 <li>Divide the unitId into numerator (the part before the "-per-") and denominator (the part after the "-per-). If both are empty, fail</li> 2524 <li>Set both globalPlaceholder and globalPlaceholderPosition to be empty</li> 2525 <li>Set numeratorUnitString to patternTimes(numerator, length, per0(pluralCategory), per0(caseVariant))</li> 2526 <li>Set denominatorUnitString to patternTimes(denominator, length, per1(pluralCategory), per1(caseVariant)) </li> 2527 <li>Set perPattern to be getValue(times, locale, length)</li> 2528 <li>If the denominatorString is empty, set result to denominatorString, otherwise set result to format(perPattern, numeratorString, denominatorString) </li> 2529 <li>return format(result, globalPlacholder, globalPlaceholderPosition)</li> 2530 </ol> 2531 <p><strong>patternTimes(product_unit, locale, length, pluralCategory, caseVariant)</strong></p> 2532 <ol> 2533 <li>Set hasMultiple to true iff product_unit has more than one single_unit</li> 2534 <li>Set timesPattern to be getValue(times, locale, length)</li> 2535 <li>Set result to be empty</li> 2536 <li>For each single_unit in product_unit 2537 <ol> 2538 <li>If hasMultiple 2539 <ol> 2540 <li>Set singlePluralCategory to be times0(pluralCategory)</li> 2541 <li>Set singleCaseVariant to be times0(caseVariant)</li> 2542 <li>Set pluralCategory to be times1(pluralCategory)</li> 2543 <li>Set caseVariant to be times1(caseVariant)</li> 2544 </ol> 2545 </li> 2546 <li>Get the gender of that single_unit</li> 2547 <li>If singleUnit starts with a dimensionality_prefix, such as 'square-' 2548 <ol> 2549 <li>set dimensionalityPrefixPattern to be getValue(that dimensionality_prefix, locale, length, singlePluralCategory, singleCaseVariant, gender), such as "{0} kwadratowym"</li> 2550 <li>set singlePluralCategory to be power0(singlePluralCategory)</li> 2551 <li>set singleCaseVariant to be power0(singleCaseVariant)</li> 2552 <li>remove the dimensionality_prefix from singleUnit</li> 2553 </ol> 2554 </li> 2555 <li>if singleUnit starts with an si_prefix, such as 'centi' 2556 <ol> 2557 <li>set siPrefixPattern to be getValue(that si_prefix, locale, length), such as "centy{0}"</li> 2558 <li>set singlePluralCategory to be prefix0(singlePluralCategory)</li> 2559 <li>set singleCaseVariant to be prefix0(singleCaseVariant)</li> 2560 <li>remove the si_prefix from singleUnit</li> 2561 </ol> 2562 </li> 2563 <li>Set corePattern to be the getValue(singleUnit, locale, length, singlePluralCategory, singleCaseVariant), such as "{0} metrem"</li> 2564 <li>Extract(corePattern, coreUnit, placeholder, placeholderPosition) from that pattern.</li> 2565 <li>If the position is <em>middle</em>, then fail</li> 2566 <li>If globalPlaceholder is empty 2567 <ol> 2568 <li>Set globalPlaceholder to placeholder</li> 2569 <li>Set globalPlaceholderPosition to placeholderPosition</li> 2570 </ol> 2571 </li> 2572 <li>If siPrefixPattern is not empty 2573 <ol> 2574 <li>Set coreUnit to be the combineLowercasing(locale, length, siPrefixPattern, coreUnit)</li> 2575 </ol> 2576 </li> 2577 <li>If dimensionalityPrefixPattern is not empty 2578 <ol> 2579 <li>Set coreUnit to be the combineLowercasing(locale, length, dimensionalityPrefixPattern, coreUnit)</li> 2580 </ol> 2581 </li> 2582 <li>If the result is empty, set result to be coreUnit</li> 2583 <li>Otherwise set result to be format(timesPattern, result, coreUnit)</li> 2584 </ol> 2585 </li> 2586 <li>Return result</li> 2587 </ol> 2588 <p><strong>combineLowercasing(locale, length, prefixPattern, coreUnit)</strong></p> 2589 <ol> 2590 <li>If the length is "long" and the prefixPattern contains no spaces, lowercase the coreUnit according to the locale, thus "Quadrat{0}" causes "Zentimeter" to become "zentimeter"</li> 2591 <li>return format(prefixPattern, unitPattern), eg "Quadratzentimeter"</li> 2592 </ol> 2593 <p><strong>format(pattern, arguments…)</strong></p> 2594 <ol> 2595 <li>return the result of substituting the arguments for the placeholders {0}, {1}, etc.</li> 2596 </ol> 2597 <p><strong>getValue(key, locale, length, variants…)</strong></p> 2598 <ol> 2599 <li>return the element value in the locale for the path corresponding to the key, locale, length, and variants — using normal inheritance including <a href="https://unicode-org.github.io/cldr/ldml/tr35.html#Multiple_Inheritance">Lateral Inheritance</a> and <a href="https://unicode-org.github.io/cldr/ldml/tr35.html#Parent_Locales">Parent Locales</a>.</li> 2600 </ol> 2601 <p><strong>Extract(corePattern, coreUnit, placeHolder, placeholderPosition)</strong></p> 2602 <ol> 2603 <li>Find the position of the <strong>placeHolder</strong> in the core pattern</li> 2604 <li>Set <strong>placeHolderPosition</strong> to that postion (start, middle, or end)</li> 2605 <li>Remove the <strong>placeHolder</strong> from the <strong>corePattern</strong> and set <strong>coreUnit</strong> to that result</li> 2606 </ol> 2607 <p><strong>per0(...), times0(...), etc.</strong></p> 2608 <ol> 2609 <li>These represent the <strong>deriveCompound</strong> data values from <strong>Section 16 <a href="#Grammatical_Derivations">Grammatical Derivations</a></strong>, where value0 of the per-structure is given as per0(...), and so on.</li> 2610 <li>"power" corresponds to dimensionality_prefix, while "prefix" corresponds to si_prefix.</li> 2611 </ol> 2612 <p>If the locale does not provide full modern coverage, the process could fall back to root locale for some localized patterns. That may give a "ransom-note" effect for the user. To avoid that, it may be preferable to abort the process at that point, and then localize the unitId for the root locale.</p> 2613 <p>If a unit is not supported by root, then the localization is not supported by CLDR and will fail.</p> 2614 <h4>Precomposed Compound Units</h4> 2615 <p>At each point in the process, if there is a precomposed form for a segment of the unitId, then that precomposed form should be used instead. For example, if there is a pattern in the locale for (square-kilometer, length, singlePluralCategory, singleCaseVariant, gender), then it should be used instead of composing the name from "square" and "kilometer".</p> 2616 <p></p> 2617 2618 <p>There is also a precomposed <strong>perUnitPattern</strong> which is used as the 2619 denominator with another unit name. For example, a form such as 2620 "{0} per second" can be used to form "2 feet <strong>per 2621 second</strong>". The difference between these is that in some inflected 2622 languages, the compoundUnit cannot be used to form grammatical 2623 phrases. This is typically because the "per" + "second" combine 2624 in a non-trivial way. The <strong>perUnitPattern</strong> should be applied if the denominator has only one element, and matches the perUnitPattern type.</p> 2625 2626 <h3>6.5 <a name="Unit_Sequences" href="#Unit_Sequences" id= 2627 "Unit_Sequences">Unit Sequences (Mixed Units)</a></h3> 2628 <p>Units may be used in composed sequences (aka <em>mixed units</em>), such as <strong>5° 2629 30′</strong> for 5 degrees 30 minutes, or <strong>3 ft 2 2630 in.</strong> For that purpose, the appropriate width of the unit 2631 listPattern can be used to compose the units in a sequence.</p> 2632 <pre><listPattern type="unit"> (for the long form) 2633<listPattern type="unit-narrow"> 2634<listPattern type="unit-short"> 2635</pre> 2636 <p>In such a sequence, decimal fractions are typically only displayed for 2637 the last element of the sequence, if at all.</p> 2638 2639 <h3>6.6 <a name="durationUnit" href="#durationUnit" id= 2640 "durationUnit">durationUnit</a></h3> 2641 <p>The durationUnit is a special type of unit used for composed 2642 time unit durations.</p> 2643 <pre><durationUnit type="hms"> 2644 <durationUnitPattern>h:mm:ss</durationUnitPattern> <!-- 33:04:59 --> 2645</durationUnit> </pre> 2646 <p>The type contains a skeleton, where 'h' stands for hours, 2647 'm' for minutes, and 's' for sections. These are the same 2648 symbols used in availableFormats, except that there is no need 2649 to distinguish different forms of the hour.</p> 2650 <h3>6.7 <a name="coordinateUnit" href="#coordinateUnit" id= 2651 "coordinateUnit">coordinateUnit</a></h3> 2652 <p>The <strong>coordinateUnitPattern</strong> is a special type 2653 of pattern used for composing degrees of latitude and 2654 longitude, with an indicator of the quadrant. There are exactly 2655 4 type values, plus a displayName for the items in this 2656 category. An angle is composed using the appropriate 2657 combination of the <strong>angle-degrees</strong>, 2658 <strong>angle-arc-minute</strong> and 2659 <strong>angle-arc-second</strong> values. It is then 2660 substituted for the placeholder field {0} in the appropriate 2661 <strong>coordinateUnit</strong> pattern.</p> 2662 <p class="xmlExample"> 2663 <displayName>direction</displayName><br> 2664 <coordinateUnitPattern 2665 type="east">{0}E</coordinateUnitPattern><br> 2666 <coordinateUnitPattern 2667 type="north">{0}N</coordinateUnitPattern><br> 2668 <coordinateUnitPattern 2669 type="south">{0}S</coordinateUnitPattern><br> 2670 <coordinateUnitPattern 2671 type="west">{0}W</coordinateUnitPattern></p> 2672 <h3>6.8 <a name="Territory_Based_Unit_Preferences" href= 2673 "#Territory_Based_Unit_Preferences" id= 2674 "Territory_Based_Unit_Preferences">Territory-Based Unit 2675 Preferences</a></h3> 2676 <p>Different locales have different preferences for which unit 2677 or combination of units is used for a particular usage, such as 2678 measuring a person’s height. This is more fine-grained than 2679 merely a preference for metric versus US or UK measurement 2680 systems. For example, one locale may use meters alone, while 2681 another may use centimeters alone or a combination of meters 2682 and centimeters; a third may use inches alone, or (informally) 2683 a combination of feet and inches.</p> 2684 <p>The <unitPreferenceData> element, described in 2685 <a href="tr35-info.html#Preferred_Units_For_Usage">Preferred 2686 Units for Specific Usages</a>, provides information on which 2687 unit or combination of units is used for various purposes in 2688 different locales, with options for the level of formality and 2689 the scale of the measurement (e.g measuring the height of an 2690 adult versus that of an infant).</p> 2691 2692 <h3>6.9 <a name="Private_Use_Units" href="#Private_Use_Units" 2693 id="Private_Use_Units">Private-Use Units</a></h3> 2694 <p> 2695 CLDR has reserved the "xxx-" prefix in the simple_unit part of the unit identifier BNF 2696 for private-use units. CLDR will never define a type, simple unit, or compound unit 2697 such that the unit identifier starts with "xxx-", ends with "-xxx", or contains "-xxx-". 2698 </p> 2699 <p> 2700 For example, if you wanted to define your own unit "foo", you could use the simple unit "xxx-foo". 2701 </p> 2702 <p> 2703 It is valid to construct compound units containing one or more private-use simple units. 2704 For example, "xxx-foo-per-second" and "xxx-foo-per-xxx-bar" are both valid core unit 2705 identifiers for compound units. 2706 </p> 2707 <p> 2708 As explained earlier, CLDR defines all associations between types and units. It is 2709 therefore not possible to construct a valid long unit identifier containing a private-use 2710 unit; only core unit identifiers are possible. 2711 </p> 2712 <p>The older syntax used “x-”, which was expanded to “xxx-” to simplify use with BCP47 2713 syntax. That should be converted to “xxx-”.</p> 2714 <h2>7 <a name="POSIX_Elements" href="#POSIX_Elements" id= 2715 "POSIX_Elements">POSIX Elements</a></h2> 2716 <p class="dtd"><!ELEMENT posix (alias | (messages*, 2717 special*)) ><br> 2718 <!ELEMENT messages (alias | ( yesstr*, nostr*)) ></p> 2719 <p>The following are included for compatibility with POSIX.</p> 2720 <p><posix><br> 2721 <posix:messages><br> 2722 <posix:yesstr><span style="color: #0000FF">ja</span></posix:yesstr><br> 2723 2724 <posix:nostr><span style="color: #0000FF">nein</span></posix:nostr><br> 2725 2726 </posix:messages><br> 2727 <posix></p> 2728 <ol> 2729 <li>The values for yesstr and nostr contain a colon-separated 2730 list of strings that would normally be recognized as "yes" 2731 and "no" responses. For cased languages, this shall include 2732 only the lower case version. POSIX locale generation tools 2733 must generate the upper case equivalents, and the abbreviated 2734 versions, and add the English words wherever they do not 2735 conflict. Examples: 2736 <ul> 2737 <li>ja → ja:Ja:j:J:yes:Yes:y:Y</li> 2738 <li>ja → ja:Ja:j:J:yes:Yes // exclude y:Y if it conflicts 2739 with the native "no".</li> 2740 </ul> 2741 </li> 2742 <li>The older elements yesexpr and noexpr are deprecated. 2743 They should instead be generated from yesstr and nostr so 2744 that they match all the responses.</li> 2745 </ol> 2746 <p>So for English, the appropriate strings and expressions 2747 would be as follows:</p> 2748 <p>yesstr "yes:y"<br> 2749 nostr "no:n"</p> 2750 <p>The generated yesexpr and noexpr would be:</p> 2751 <p><code>yesexpr "^([yY]([eE][sS])?)"<br></code> This would 2752 match y,Y,yes,yeS,yEs,yES,Yes,YeS,YEs,YES.<br> 2753 <br> 2754 <code>noexpr "^([nN][oO]?)"</code><br> 2755 This would match n,N,no,nO,No,NO.</p> 2756 <h2>8 <a name="Reference_Elements" href="#Reference_Elements" 2757 id="Reference_Elements">Reference Element</a></h2> 2758 <p>(Use only in supplemental data; deprecated for ldml.dtd and 2759 locale data)</p> 2760 <p class="dtd"><!ELEMENT references ( reference* ) ><br> 2761 <!ELEMENT reference ( #PCDATA ) ><br> 2762 <!ATTLIST reference type NMTOKEN #REQUIRED><br> 2763 <!ATTLIST reference standard ( true | false ) #IMPLIED 2764 ><br> 2765 <!ATTLIST reference uri CDATA #IMPLIED ></p> 2766 <p>The references section supplies a central location for 2767 specifying references and standards. The uri should be supplied 2768 if at all possible. If not online, then a ISBN number should be 2769 supplied, such as in the following example:</p> 2770 <p class="example"><reference type="R2" 2771 uri="https://www.ur.se/nyhetsjournalistik/3lan.html">Landskoder 2772 på Internet</reference><br> 2773 <reference type="R3" uri="URN:ISBN:91-47-04974-X">Svenska 2774 skrivregler</reference></p> 2775 <h2>9 <a name="Segmentations" href="#Segmentations" id= 2776 "Segmentations">Segmentations</a></h2> 2777 <p class="dtd"><!ELEMENT segmentations ( alias | 2778 segmentation*) ></p> 2779 <p class="dtd"><!ELEMENT segmentation ( alias | (variables?, 2780 segmentRules? , exceptions?, suppressions?) | special*) 2781 ><br> 2782 <!ATTLIST segmentation type NMTOKEN #REQUIRED ></p> 2783 <p class="dtd"><!ELEMENT variables ( alias | variable*) 2784 ></p> 2785 <p class="dtd"><!ELEMENT variable ( #PCDATA ) ><br> 2786 <!ATTLIST variable id CDATA #REQUIRED ></p> 2787 <p class="dtd"><!ELEMENT segmentRules ( alias | rule*) 2788 ></p> 2789 <p class="dtd"><!ELEMENT rule ( #PCDATA ) ><br> 2790 <!ATTLIST rule id NMTOKEN #REQUIRED ></p> 2791 <p class="dtd"><!ELEMENT suppressions ( suppression* ) 2792 ></p> 2793 <p class="dtd"><!ATTLIST suppressions type NMTOKEN 2794 "standard" ></p> 2795 <p class="dtd"><!ATTLIST suppressions draft ( approved | 2796 contributed | provisional | unconfirmed ) #IMPLIED ></p> 2797 <p class="dtd"><!ELEMENT suppression ( #PCDATA ) ></p> 2798 <p>The segmentations element provides for segmentation of text 2799 into words, lines, or other segments. The structure is based on 2800 [<a href= 2801 "https://www.unicode.org/reports/tr41/#UAX29">UAX29</a>] 2802 notation, but adapted to be machine-readable. It uses a list of 2803 variables (representing character classes) and a list of rules. 2804 Each must have an id attribute.</p> 2805 <p>The rules in <i>root</i> implement the segmentations found 2806 in [<a href= 2807 "https://www.unicode.org/reports/tr41/#UAX29">UAX29</a>] and 2808 [<a href= 2809 "https://www.unicode.org/reports/tr41/#UAX14">UAX14</a>], for 2810 grapheme clusters, words, sentences, and lines. They can be 2811 overridden by rules in child locales.</p> 2812 <p>Here is an example:</p> 2813 <pre><segmentations> 2814 <segmentation type="GraphemeClusterBreak"> 2815 <variables> 2816 <variable id="$CR">\p{Grapheme_Cluster_Break=CR}</variable> 2817 <variable id="$LF">\p{Grapheme_Cluster_Break=LF}</variable> 2818 <variable id="$Control">\p{Grapheme_Cluster_Break=Control}</variable> 2819 <variable id="$Extend">\p{Grapheme_Cluster_Break=Extend}</variable> 2820 <variable id="$L">\p{Grapheme_Cluster_Break=L}</variable> 2821 <variable id="$V">\p{Grapheme_Cluster_Break=V}</variable> 2822 <variable id="$T">\p{Grapheme_Cluster_Break=T}</variable> 2823 <variable id="$LV">\p{Grapheme_Cluster_Break=LV}</variable> 2824 <variable id="$LVT">\p{Grapheme_Cluster_Break=LVT}</variable> 2825 </variables> 2826 <segmentRules> 2827 <rule id="3"> $CR × $LF </rule> 2828 <rule id="4"> ( $Control | $CR | $LF ) ÷ </rule> 2829 <rule id="5"> ÷ ( $Control | $CR | $LF ) </rule> 2830 <rule id="6"> $L × ( $L | $V | $LV | $LVT ) </rule> 2831 <rule id="7"> ( $LV | $V ) × ( $V | $T ) </rule> 2832 <rule id="8"> ( $LVT | $T) × $T </rule> 2833 <rule id="9"> × $Extend </rule> 2834 </segmentRules> 2835 </segmentation> 2836...</pre> 2837 <p><b>Variables:</b> All variable ids must start with a $, and 2838 otherwise be valid identifiers according to the Unicode 2839 definitions in [<a href= 2840 "https://www.unicode.org/reports/tr41/#UAX31">UAX31</a>]. The 2841 contents of a variable is a regular expression using variables 2842 and <a href="tr35.html#Unicode_Sets">UnicodeSet</a>s. The 2843 ordering of variables is important; they are evaluated in order 2844 from first to last (see <i><a href= 2845 "#Segmentation_Inheritance">Section 9.1 Segmentation 2846 Inheritance</a></i>). It is an error to use a variable before 2847 it is defined.</p> 2848 <p><b>Rules:</b> The contents of a rule uses the syntax of 2849 [<a href= 2850 "https://www.unicode.org/reports/tr41/#UAX29">UAX29</a>]. The 2851 rules are evaluated in numeric id order (which may not be the 2852 order in which the appear in the file). The first rule that 2853 matches determines the status of a boundary position, that is, 2854 whether it breaks or not. Thus ÷ means a break is allowed; × 2855 means a break is forbidden. It is an error if the rule does not 2856 contain exactly one of these characters (except where a rule 2857 has no contents at all, or if the rule uses a variable that has 2858 not been defined.</p> 2859 <p>There are some implicit rules:</p> 2860 <ul> 2861 <li>The implicit initial rules are always "start-of-text ÷" 2862 and "÷ end-of-text"; these are not to be included 2863 explicitly.</li> 2864 <li>The implicit final rule is always "Any ÷ Any". This is 2865 not to be included explicitly.</li> 2866 </ul> 2867 <blockquote> 2868 <p><b>Note:</b> A rule like X Format* -> X in [<a href= 2869 "https://www.unicode.org/reports/tr41/#UAX29">UAX29</a>] and 2870 [<a href= 2871 "https://www.unicode.org/reports/tr41/#UAX14">UAX14</a>] is 2872 not supported. Instead, this needs to be expressed as normal 2873 regular expressions. The normal way to support this is to 2874 modify the variables, such as in the following example:</p> 2875 <pre id="line870"> 2876 <variable id="$Format">\p{Word_Break=Format}</variable> 2877<variable id="$Katakana">\p{Word_Break=Katakana}</variable> 2878... 2879<!-- In place of rule 3, add format and extend to everything --> 2880<variable id="$X">[$Format $Extend]*</variable> 2881<variable id="$Katakana">($Katakana $X)</variable> 2882<variable id="$ALetter">($ALetter $X)</variable> 2883...</pre> 2884 </blockquote> 2885 <h3>9.1 <a name="Segmentation_Inheritance" href= 2886 "#Segmentation_Inheritance" id= 2887 "Segmentation_Inheritance">Segmentation Inheritance</a></h3> 2888 <p>Variables and rules both inherit from the parent.</p> 2889 <p><b>Variables:</b> The child's variable list is logically 2890 appended to the parent's, and evaluated in that order. For 2891 example:</p> 2892 <p><font color="#0000FF"><code>// in parent</code></font> 2893 <code><br> 2894 <variable id="$AL">[:linebreak=AL:]</variable><br> 2895 <variable 2896 id="$YY">[[:linebreak=XX:]$AL]</variable></code> 2897 <font color="#0000FF"><code>// adds $AL</code></font></p> 2898 <p><font color="#0000FF"><code>// in child</code></font> 2899 <code><br> 2900 <variable id="$AL">[$AL && 2901 [^a-z]]</variable> <font color="#0000FF">// changes $AL, 2902 does not affect $YY</font><br> 2903 <variable id="$ABC">[abc]</variable></code> 2904 <font color="#0000FF"><code>// adds new rule</code></font></p> 2905 <p><b>Rules:</b> The rules are also logically appended to the 2906 parent's. Because rules are evaluated in numeric id order, to 2907 insert a rule in between others just requires using an 2908 intermediate number. For example, to insert a rule after 2909 id="10.1" and before id="10.2", just use id="10.15". To delete 2910 a rule, use empty contents, such as:</p> 2911 <p><code><rule id="3"/></code> <font color= 2912 "#0000FF"><code>// deletes rule 3</code></font></p> 2913 <h3>9.2 <a name="Segmentation_Exceptions" href= 2914 "#Segmentation_Exceptions" id= 2915 "Segmentation_Exceptions">Segmentation Suppressions</a></h3> 2916 <p><b>Note:</b> As of CLDR 26, the 2917 <code><suppressions></code> data is to be considered a 2918 technology preview. Data currently in CLDR was extracted from 2919 the Unicode Localization Interoperability project, or ULI. See 2920 <a href="http://uli.unicode.org">http://uli.unicode.org</a> for 2921 more information on the ULI project.</p> 2922 <p>The segmentation <b>suppressions</b> list provides a set of 2923 cases which, though otherwise identified as a segment by rules, 2924 should be skipped (suppressed) during segmentation.</p> 2925 <p>For example, in the English phrase "Mr. Smith", CLDR 2926 segmentation rules would normally find a Sentence Break between 2927 "Mr" and "Smith". However, typically, "Mr." is just an 2928 abbreviation for "Mister", and not actually the end of a 2929 sentence.</p> 2930 <p>Each suppression has a separate 2931 <code><suppression></code> element, whose contents are 2932 the break to be skipped.</p> 2933 <p>Example:</p> 2934 <pre> 2935 <segmentation type="SentenceBreak"> 2936 <suppressions type="standard" draft="provisional"> 2937 <suppression>Maj.</suppression> 2938 <suppression>Mr.</suppression> 2939 <suppression>Lt.Cdr.</suppression> 2940 . . . 2941 </suppressions> 2942 </segmentation> 2943 </pre> 2944 <p><b>Note:</b> These elements were called 2945 <code><exceptions></code> and 2946 <code><exception></code> prior to CLDR 26, but those 2947 names are now deprecated.</p> 2948 <h2>10 <a name="Transforms" href="#Transforms" id= 2949 "Transforms">Transforms</a></h2> 2950 <p>Transforms provide a set of rules for transforming text via 2951 a specialized set of context-sensitive matching rules. They are 2952 commonly used for transliterations or transcriptions, but also 2953 other transformations such as full-width to half-width (for 2954 <i>katakana</i> characters). The rules can be simple one-to-one 2955 relationships between characters, or involve more complicated 2956 mappings. Here is an example:</p> 2957 <pre> 2958 <transform source="Greek" target="Latin" variant="UNGEGN" direction="both"> 2959... 2960 <comment>Useful variables</comment> 2961 <tRule>$gammaLike = [ΓΚΞΧγκξχϰ] ;</tRule> 2962 <tRule>$egammaLike = [GKXCgkxc] ;</tRule> 2963... 2964 <comment>Rules are predicated on running NFD first, and NFC afterwards</comment> 2965 <tRule>::NFD (NFC) ;</tRule> 2966... 2967 <tRule>λ ↔ l ;</tRule> 2968 <tRule>Λ ↔ L ;</tRule> 2969... 2970 <tRule>γ } $gammaLike ↔ n } $egammaLike ;</tRule> 2971 <tRule>γ ↔ g ;</tRule> 2972... 2973 <tRule>::NFC (NFD) ;</tRule> 2974... 2975</transform></pre> 2976 <p>The source and target values are valid locale identifiers, 2977 where 'und' means an unspecified language, plus some additional 2978 extensions.</p> 2979 <ul> 2980 <li>The long names of a script according to [<a href= 2981 "https://www.unicode.org/reports/tr41/#UAX24">UAX24</a>] may 2982 be used instead of the short script codes. The script 2983 identifier may also omit und; that is, "und_Latn" may be 2984 written as just "Latn".</li> 2985 <li>The long names of the English languages may also be used 2986 instead of the languages.</li> 2987 <li>The term "Any" may be used instead of a solitary 2988 "und".</li> 2989 <li>Other identifiers may be used for special purposes. In 2990 CLDR, these include: Accents, Digit, Fullwidth, Halfwidth, 2991 Jamo, NumericPinyin, Pinyin, Publishing, Tone. (Other than 2992 these values, valid private use locale identifiers should be 2993 used, such as "x-Special".)</li> 2994 <li>When presenting localizing transform names, the "und_" is 2995 normally omitted. Thus for a transliterator with the ID 2996 "und_Latn-und_Grek" (or the equivalent "Latin-Greek"), the 2997 translated name for Greek would be Λατινικό-Ελληνικό.</li> 2998 </ul> 2999 <p>In version 29.0, BCP47 identifiers were added as aliases 3000 (while retaining the old identifiers). The following table 3001 shows the relationship between the old identifiers and the 3002 BCP47 format identifiers.</p> 3003 <table class='simple'> 3004 <tbody> 3005 <tr> 3006 <th>Old ID</th> 3007 <th>BCP47 ID</th> 3008 <th>Comments</th> 3009 </tr> 3010 <tr> 3011 <td><strong>es_FONIPA</strong>-es_419_FONIPA</td> 3012 <td>es-419-fonipa-t-<strong>es-fonipa</strong></td> 3013 <td rowspan="2">The order reverses with -t-. That is, the 3014 language subtag part is what results.</td> 3015 </tr> 3016 <tr> 3017 <td><strong>hy_AREVMDA</strong>-hy_AREVMDA_FONIPA</td> 3018 <td>hy-arevmda-fonipa-t-<strong>hy-arevmda</strong></td> 3019 </tr> 3020 <tr> 3021 <td><strong>Devanagari</strong>-Latin</td> 3022 <td>und-Latn-t-<strong>und-deva</strong></td> 3023 <td rowspan="2">Scripts add <strong>und-</strong></td> 3024 </tr> 3025 <tr> 3026 <td><strong>Latin</strong>-Devanagari</td> 3027 <td>und-Deva-t-<strong>und-latn</strong></td> 3028 </tr> 3029 <tr> 3030 <td>Greek-Latin/UNGEGN</td> 3031 <td>und-Latn-t-und-grek-<strong>m0-ungegn</strong></td> 3032 <td>Variants use the <strong>-m0-</strong> key.</td> 3033 </tr> 3034 <tr> 3035 <td>Russian-Latin/BGN</td> 3036 <td>ru<strong>-Latn</strong>-t-ru-m0-bgn</td> 3037 <td>Languages will have a script when it isn’t the 3038 default.</td> 3039 </tr> 3040 <tr> 3041 <td>Any-Hex/xml</td> 3042 <td>und-t-<strong>d0-hex</strong>-m0-xml</td> 3043 <td rowspan="2"><strong>Any</strong> becomes 3044 <strong>und</strong>, and keys <strong>d0</strong> 3045 (destination) and <strong>s0</strong> (source) are used 3046 for non-locales.</td> 3047 </tr> 3048 <tr> 3049 <td>Hex-Any/xml</td> 3050 <td>und-t-<strong>s0-hex</strong>-m0-xml</td> 3051 </tr> 3052 <tr> 3053 <td>Any-<strong>Publishing</strong></td> 3054 <td>und-t-d0-<strong>publish</strong></td> 3055 <td rowspan="2">Non-locales are normally the lowercases 3056 of the old ID, but may change because of BCP47 length 3057 restrictions.</td> 3058 </tr> 3059 <tr> 3060 <td><strong>Publishing</strong>-Any</td> 3061 <td>und-t-s0-<strong>publish</strong></td> 3062 </tr> 3063 </tbody> 3064 </table> 3065 <p>Note that the script and region codes are cased iff they are 3066 in the main subtag, but are lowercase in extensions.</p> 3067 <h3>10.1 <a name="Inheritance" href="#Inheritance" id= 3068 "Inheritance">Inheritance</a></h3> 3069 <p>The CLDR transforms are built using the following locale 3070 inheritance. While this inheritance is not required of LDML 3071 implementations, the transforms supplied with CLDR may not 3072 otherwise behave as expected without some changes.</p> 3073 <p>For either the source or the target, the fallback starts 3074 from the maximized locale ID (using the likely-subtags data). 3075 It also uses the country for lookup before the base language is 3076 reached, and root is never accessed: instead the script(s) 3077 associated with the language are used. Where there are multiple 3078 scripts, the maximized script is tried first, and then the 3079 other scripts associated with the language (from supplemental 3080 data).</p> 3081 <p>For example, see the bolded items below in the fallback 3082 chain for <strong>az_IR</strong>.</p> 3083 <table> 3084 <tr> 3085 <th> </th> 3086 <th>Locale ID</th> 3087 <th>Comments</th> 3088 </tr> 3089 <tr> 3090 <td>1</td> 3091 <td><strong>az_Arab_IR</strong></td> 3092 <td>The maximized locale for az_IR</td> 3093 </tr> 3094 <tr> 3095 <td>2</td> 3096 <td>az_Arab</td> 3097 <td>Normal fallback</td> 3098 </tr> 3099 <tr> 3100 <td>3</td> 3101 <td><strong>az_IR</strong></td> 3102 <td>Inserted country locale</td> 3103 </tr> 3104 <tr> 3105 <td>4</td> 3106 <td>az</td> 3107 <td>Normal fallback</td> 3108 </tr> 3109 <tr> 3110 <td>5</td> 3111 <td><strong>Arab</strong></td> 3112 <td>Maximized script</td> 3113 </tr> 3114 <tr> 3115 <td>6</td> 3116 <td><strong>Cyrl</strong></td> 3117 <td>Other associated script</td> 3118 </tr> 3119 </table> 3120 <p>The source, target, and variant use "laddered" fallback, 3121 where the source changes the most quickly (using the above 3122 rules), then the target (using the above rules), then the 3123 variant if any, is discarded. That is, in pseudo code:</p> 3124 <ul> 3125 <li>for variant in {variant, ""} 3126 <ul> 3127 <li>for target in target-chain 3128 <ul> 3129 <li>for source in source-chain 3130 <ul> 3131 <li>transform = lookup source-target/variant</li> 3132 <li>if transform != null return transform</li> 3133 </ul> 3134 </li> 3135 </ul> 3136 </li> 3137 </ul> 3138 </li> 3139 </ul> 3140 <p>For example, here is the fallback chain for 3141 <strong>ru_RU-el_GR/BGN</strong>.</p> 3142 <div align="center"> 3143 <table> 3144 <tr> 3145 <th>source</th> 3146 <th> </th> 3147 <th>target</th> 3148 <th>variant</th> 3149 </tr> 3150 <tr> 3151 <td>ru_RU</td> 3152 <td>-</td> 3153 <td>el_GR</td> 3154 <td>/BGN</td> 3155 </tr> 3156 <tr> 3157 <td>ru</td> 3158 <td>-</td> 3159 <td>el_GR</td> 3160 <td>/BGN</td> 3161 </tr> 3162 <tr> 3163 <td>Cyrl</td> 3164 <td>-</td> 3165 <td>el_GR</td> 3166 <td>/BGN</td> 3167 </tr> 3168 <tr> 3169 <td>ru_RU</td> 3170 <td>-</td> 3171 <td>el</td> 3172 <td>/BGN</td> 3173 </tr> 3174 <tr> 3175 <td>ru</td> 3176 <td>-</td> 3177 <td>el</td> 3178 <td>/BGN</td> 3179 </tr> 3180 <tr> 3181 <td>Cyrl</td> 3182 <td>-</td> 3183 <td>el</td> 3184 <td>/BGN</td> 3185 </tr> 3186 <tr> 3187 <td>ru_RU</td> 3188 <td>-</td> 3189 <td>Grek</td> 3190 <td>/BGN</td> 3191 </tr> 3192 <tr> 3193 <td>ru</td> 3194 <td>-</td> 3195 <td>Grek</td> 3196 <td>/BGN</td> 3197 </tr> 3198 <tr> 3199 <td>Cyrl</td> 3200 <td>-</td> 3201 <td>Grek</td> 3202 <td>/BGN</td> 3203 </tr> 3204 <tr> 3205 <td>ru_RU</td> 3206 <td>-</td> 3207 <td>el_GR</td> 3208 <td></td> 3209 </tr> 3210 <tr> 3211 <td>ru</td> 3212 <td>-</td> 3213 <td>el_GR</td> 3214 <td></td> 3215 </tr> 3216 <tr> 3217 <td>Cyrl</td> 3218 <td>-</td> 3219 <td>el_GR</td> 3220 <td></td> 3221 </tr> 3222 <tr> 3223 <td>ru_RU</td> 3224 <td>-</td> 3225 <td>el</td> 3226 <td></td> 3227 </tr> 3228 <tr> 3229 <td>ru</td> 3230 <td>-</td> 3231 <td>el</td> 3232 <td></td> 3233 </tr> 3234 <tr> 3235 <td>Cyrl</td> 3236 <td>-</td> 3237 <td>el</td> 3238 <td></td> 3239 </tr> 3240 <tr> 3241 <td>ru_RU</td> 3242 <td>-</td> 3243 <td>Grek</td> 3244 <td></td> 3245 </tr> 3246 <tr> 3247 <td>ru</td> 3248 <td>-</td> 3249 <td>Grek</td> 3250 <td></td> 3251 </tr> 3252 <tr> 3253 <td>Cyrl</td> 3254 <td>-</td> 3255 <td>Grek</td> 3256 <td></td> 3257 </tr> 3258 </table> 3259 </div> 3260 <p>Japanese and Korean are special, since they can be 3261 represented by combined script codes, such as ja_Jpan, ja_Hrkt, 3262 ja_Hira, or ja_Kana. These need to be considered in the above 3263 fallback chain as well.</p> 3264 <h4>10.1.1 <a name="Pivots" href="#Pivots" id= 3265 "Pivots">Pivots</a></h4> 3266 <p>Transforms can also use <i>pivots</i>. These are used when 3267 there is no direct transform between a source and target, but 3268 there are transforms X-Y and Y-Z. In such a case, the 3269 transforms can be internally chained to get X-Y = X-Y;Y-Z. This 3270 is done explicitly with the Indic script transforms: to get 3271 Devanagari-Latin, internally it is done by transforming first 3272 from Devanagari to Interindic (an internal superset encoding 3273 for Indic scripts), then from Interindic to Latin. This allows 3274 there to be only N sets of transform rules for the Indic 3275 scripts: each one to and from Interindic. These pivots are 3276 explicitly represented in the CLDR transforms.</p> 3277 <p>Note that the characters currently used by Interindic are 3278 private use characters. To prevent those from “leaking” out 3279 into text, transforms converting from Interindic must ensure 3280 that they convert all the possible values used in 3281 Interindic.</p> 3282 <p>The pivots can also be produced automatically (implicitly), 3283 as a fallback. A particularly useful pivot is IPA, since that 3284 tends to preserve pronunciation. For example, <em>Czech to 3285 IPA</em> can be chained with <em>IPA to Katakana</em> to get 3286 <em>Czech to Katakana</em>.</p> 3287 <p>CLDR often has special forms of IPA: not just "und-FONIPA" 3288 but "cs-FONIPA": specifically IPA that has come from Czech. 3289 These variants typically preserve some features of the source 3290 language — such as double consonants — that are 3291 indistinguishable from single consonants in that language, but 3292 that are often preserved in traditional transliterations. Thus 3293 when matching prospective pivots, FONIPA is treated specially. 3294 If there is an exact match, that match is used (such as 3295 cs-cs_FONIPA + cs_FONIPA-ko). Otherwise, the language is 3296 ignored, as for example in cs-cs_FONIPA + ru_FONIPA-ko.</p> 3297 <p>The interaction of implicit pivots and inheritance may 3298 result in a longer inheritance chain lookup than desired, so 3299 implementers may consider having some sort of caching mechanism 3300 to increase performance.</p> 3301 <h3>10.2 <a name="Variants" href="#Variants" id= 3302 "Variants">Variants</a></h3> 3303 <p>Variants used in CLDR include UNGEGN and BGN, both 3304 indicating sources for transliterations. There is an additional 3305 attribute <code>private="true"</code> which is used to indicate 3306 that the transform is meant for internal use, and should not be 3307 displayed as a separate choice in a UI.</p> 3308 <p>There are many different systems of transliteration. The 3309 goal for the "unqualified" script transliterations are</p> 3310 <ol> 3311 <li>to be lossless when going to Latin and back</li> 3312 <li>to be as lossless as possible when going to other 3313 scripts</li> 3314 <li>to abide by a common standard as much as possible 3315 (possibly supplemented to meet goals 1 and 2).</li> 3316 </ol> 3317 <p>Language-to-language transliterations, and variant 3318 script-to-script transliterations are generally transcriptions, 3319 and not expected to be lossless.</p> 3320 <p>Additional transliterations may also be defined, such as 3321 customized language-specific transliterations (such as between 3322 Russian and French), or those that match a particular 3323 transliteration standard, such as the following:</p> 3324 <ul> 3325 <li>UNGEGN - United Nations Group of Experts on Geographical 3326 Names</li> 3327 <li>BGN - United States Board on Geographic Names</li> 3328 <li>ISO9 - ISO/IEC 9</li> 3329 <li>ISO15915 - ISO/IEC 15915</li> 3330 <li>ISCII91 - ISCII 91</li> 3331 <li>KMOCT - South Korean Ministry of Culture & 3332 Tourism</li> 3333 <li>USLC - US Library of Congress</li> 3334 <li>UKPCGN - Permanent Committee on Geographical Names for 3335 British Official Use</li> 3336 <li>RUGOST - Russian Main Administration of Geodesy and 3337 Cartography</li> 3338 </ul> 3339 <p>The rules for transforms are described in Section 10.3 3340 <a href="#Transform_Rules_Syntax">Transform Rules Syntax</a>. 3341 For more information on Transliteration, see <a href= 3342 "http://cldr.unicode.org/index/cldr-spec/transliteration-guidelines"> 3343 Transliteration Guidelines</a>.</p> 3344 <h3>10.3 <a name="Transform_Rules_Syntax" href= 3345 "#Transform_Rules_Syntax" id="Transform_Rules_Syntax">Transform 3346 Rules Syntax</a></h3> 3347 <p class="dtd"><!ELEMENT transforms ( transform*) ><br> 3348 <!ELEMENT transform ((comment | tRule)*) ><br> 3349 <!ATTLIST transform source CDATA #IMPLIED ><br> 3350 <!ATTLIST transform target CDATA #IMPLIED ><br> 3351 <!ATTLIST transform variant CDATA #IMPLIED ><br> 3352 <!ATTLIST transform direction ( forward | backward | both ) 3353 "both" ><br> 3354 <!ATTLIST transform alias CDATA #IMPLIED ><br> 3355 <!--@VALUE--><br> 3356 <!ATTLIST transform backwardAlias CDATA #IMPLIED ><br> 3357 <!--@VALUE--><br> 3358 <!ATTLIST transform visibility ( internal | external ) 3359 "external" ><br> 3360 <!ELEMENT comment (#PCDATA) ><br> 3361 <!ELEMENT tRule (#PCDATA) ></p> 3362 <p>The transform attributes indicate the 3363 <strong>source</strong>, <strong>target</strong>, 3364 <strong>direction</strong>, and <strong>alias</strong>es. For 3365 example:</p> 3366 <p class='example'><transform<br> 3367 source="ja_Hrkt"<br> 3368 target="ja_Latn"<br> 3369 variant="BGN"<br> 3370 direction="forward"<br> 3371 draft="provisional"<br> 3372 alias="Katakana-Latin/BGN 3373 ja-Latn-t-ja-hrkt-m0-bgn"></p> 3374 <p>The direction is either <strong>forward</strong> or 3375 <strong>both</strong> (<strong>backward</strong> is possible in 3376 theory, but not used). This indicates which directions the 3377 rules support.</p> 3378 <p>If the direction is <strong>forward</strong>, then an ID is 3379 composed from <strong>target + "-" + source + "/" + 3380 variant</strong>. If the direction is <strong>both</strong>, 3381 then the inverse ID is also value: <strong>source + "-" + 3382 target + "/" + variant</strong>. The <strong>alias</strong> 3383 attribute contains a space-delimited list of alternant forward 3384 IDs, while the <strong>backwardAlias</strong> contains a 3385 space-delimited list of alternant backward IDs. The BCP47 3386 versions of the IDs will be in the <strong>alias</strong> 3387 and/or <strong>backwardAlias</strong> attributes.</p> 3388 <p>The <strong>visibility</strong> attribute indicates whether 3389 the IDs should be externally visible, or whether they are only 3390 used internally.</p> 3391 <p>In previous versions, the rules were expressed as 3392 fine-grained XML. That was discarded in CLDR version 29, in 3393 favor of a simpler format where the separate rules are simply 3394 terminated with ";".</p> 3395 <p>The transform rules are similar to regular-expression 3396 substitutions, but adapted to the specific domain of text 3397 transformations. The rules and comments in this discussion will 3398 be intermixed, with # marking the comments. The simplest rule 3399 is a conversion rule, which replaces one string of characters 3400 with another. The conversion rule takes the following form:</p> 3401 <table cellspacing="0" cellpadding="8" border="1"> 3402 <tr> 3403 <td valign="top" bgcolor="#EEEEEE"><code>xy → z 3404 ;</code></td> 3405 </tr> 3406 </table> 3407 <p>This converts any substring "xy" into "z". Rules are 3408 executed in order; consider the following rules:</p> 3409 <table cellspacing="0" cellpadding="8" border="1"> 3410 <tr> 3411 <td valign="top" bgcolor="#EEEEEE"><code>sch → sh ;<br> 3412 ss → z ;</code></td> 3413 </tr> 3414 </table> 3415 <p>This conversion rule transforms "bass school" into "baz 3416 shool". The transform walks through the string from start to 3417 finish. Thus given the rules above "bassch" will convert to 3418 "bazch", because the "ss" rule is found before the "sch" rule 3419 in the string (later, we'll see a way to override this 3420 behavior). If two rules can both apply at a given point in the 3421 string, then the transform applies the first rule in the 3422 list.</p> 3423 <p>All of the ASCII characters except numbers and letters are 3424 reserved for use in the rule syntax, as are the characters →, 3425 ←, ↔. Normally, these characters do not need to be converted. 3426 However, to convert them use either a pair of single quotes or 3427 a slash. The pair of single quotes can be used to surround a 3428 whole string of text. The slash affects only the character 3429 immediately after it. For example, to convert from a 3430 U+2190 ( ← ) LEFTWARDS ARROW to the string 3431 "arrow sign" (with a space), use one of the following 3432 rules:</p> 3433 <table cellspacing="0" cellpadding="8" border="1"> 3434 <tr> 3435 <td valign="top" bgcolor="#EEEEEE"><code>\← 3436 → arrow\ sign ;<br> 3437 '←' → 'arrow sign' ;<br> 3438 '←' → arrow' 'sign ;</code></td> 3439 </tr> 3440 </table> 3441 <p>Spaces may be inserted anywhere without any effect on the 3442 rules. Use extra space to separate items out for clarity 3443 without worrying about the effects. This feature is 3444 particularly useful with combining marks; it is handy to put 3445 some spaces around it to separate it from the surrounding text. 3446 The following is an example:</p> 3447 <table cellspacing="0" cellpadding="8" border="1"> 3448 <tr> 3449 <td valign="top" bgcolor="#EEEEEE"><code> → i ; # an 3450 iota-subscript diacritic turns into an i.</code></td> 3451 </tr> 3452 </table> 3453 <p>For a real space in the rules, place quotes around it. For a 3454 real backslash, either double it \\, or quote it '\'. For a 3455 real single quote, double it '', or place a backslash before it 3456 \'.</p> 3457 <p>Any text that starts with a hash mark and concludes a line 3458 is a comment. Comments help document how the rules work. The 3459 following shows a comment in a rule:</p> 3460 <table cellspacing="0" cellpadding="8" border="1"> 3461 <tr> 3462 <td valign="top" bgcolor="#EEEEEE"><code>x → ks ; # change 3463 every x into ks</code></td> 3464 </tr> 3465 </table> 3466 <p>The “\u” and “\x” hex notations can be used instead of any 3467 letter. For instance, instead of using the Greek π, one could 3468 write either of the following:</p> 3469 <table cellspacing="0" cellpadding="8" border="1"> 3470 <tr> 3471 <td valign="top" bgcolor="#EEEEEE"><code>\u03C0 → p ;<br> 3472 \x{3C0} → p ;</code></td> 3473 </tr> 3474 </table> 3475 <p>One can also define and use variables, such as:</p> 3476 <table cellspacing="0" cellpadding="8" border="1"> 3477 <tr> 3478 <td valign="top" bgcolor="#EEEEEE"><code>$pi = \u03C0 ;<br> 3479 $pi → p ;</code></td> 3480 </tr> 3481 </table> 3482 <h4>10.3.1 <a name="Dual_Rules" href="#Dual_Rules" id= 3483 "Dual_Rules">Dual Rules</a></h4> 3484 <p>Rules can also specify what happens when an inverse 3485 transform is formed. To do this, we reverse the direction of 3486 the "←" sign. Thus the above example becomes:</p> 3487 <table cellspacing="0" cellpadding="8"> 3488 <tr> 3489 <td valign="top" bgcolor="#EEEEEE"><code>$pi ← p 3490 ;</code></td> 3491 </tr> 3492 </table> 3493 <p>With the inverse transform, "p" will convert to the Greek p. 3494 These two directions can be combined together into a dual 3495 conversion rule by using the "↔" operator, yielding:</p> 3496 <table cellspacing="0" cellpadding="8" border="1"> 3497 <tr> 3498 <td valign="top" bgcolor="#EEEEEE"><code>$pi ↔ p 3499 ;</code></td> 3500 </tr> 3501 </table> 3502 <h4>10.3.2 <a name="Context" href="#Context" id= 3503 "Context">Context</a></h4> 3504 <p>Context can be used to have the results of a transformation 3505 be different depending on the characters before or after. The 3506 following rule removes hyphens, but only when they follow 3507 lowercase characters:</p> 3508 <table cellspacing="0" cellpadding="8" border="1"> 3509 <tr> 3510 <td valign="top" bgcolor="#EEEEEE"><code>[:Lowercase:] { 3511 '-' → ;</code></td> 3512 </tr> 3513 </table> 3514 <p>Contexts can be before or after or both, such as in a rule 3515 to remove hyphens between lowercase and uppercase letters:</p> 3516 <table cellspacing="0" cellpadding="8" border="1"> 3517 <tr> 3518 <td valign="top" bgcolor="#EEEEEE"><code>[:Lowercase:] { 3519 '-' } [:Uppercase:] → ;</code></td> 3520 </tr> 3521 </table> 3522 <p>Each context is optional and may be empty; the following two 3523 rules are equivalent:</p> 3524 <table cellspacing="0" cellpadding="8" border="1"> 3525 <tr> 3526 <td valign="top" bgcolor="#EEEEEE"><code>$pi ↔ p ;<br> 3527 {$pi} ↔ {p} ;</code></td> 3528 </tr> 3529 </table> 3530 <p>The context itself ([: <code>Lowercase</code> :]) is 3531 unaffected by the replacement; only the text within braces is 3532 changed.</p> 3533 <p>Character classes (UnicodeSets) in the contexts can contain 3534 the special symbol $, which means “off either end of the 3535 string”. It is roughly similar to $ and ^ in regex. Unlike 3536 normal regex, however, it can occur in character classes. Thus 3537 the following rule removes hyphens that are after lowercase 3538 characters, <em>or</em> are at the start of a string.</p> 3539 <table cellspacing="0" cellpadding="8" border="1"> 3540 <tr> 3541 <td valign="top" bgcolor="#EEEEEE"><code>[[:Lowercase:]$] 3542 {'-' → ;</code></td> 3543 </tr> 3544 </table> 3545 <p>Thus the negation of a UnicodeSet will normally also match 3546 before or after the end of a string. The following will remove 3547 hyphens that are not after lowercase characters<em>, including 3548 hyphens at the start of a string</em>.</p> 3549 <table cellspacing="0" cellpadding="8" border="1"> 3550 <tr> 3551 <td valign="top" bgcolor="#EEEEEE"><code>[^[:Lowercase:]] 3552 {'-' → ;</code></td> 3553 </tr> 3554 </table> 3555 <p>It will thus convert “-B A-B a-b” to “B AB a-b”.</p> 3556 <h4>10.3.3 <a name="Revisiting" href="#Revisiting" id= 3557 "Revisiting">Revisiting</a></h4> 3558 <p>If the resulting text contains a vertical bar "|", then that 3559 means that processing will proceed from that point and that the 3560 transform will revisit part of the resulting text. Thus the | 3561 marks a "cursor" position. For example, if we have the 3562 following, then the string "xa" will convert to "w".</p> 3563 <table cellspacing="0" cellpadding="8" border="1"> 3564 <tr> 3565 <td valign="top" bgcolor="#EEEEEE"><code>x → y | z ;<br> 3566 z a → w;</code></td> 3567 </tr> 3568 </table> 3569 <p>First, "xa" is converted to "yza". Then the processing will 3570 continue from after the character "y", pick up the "za", and 3571 convert it. Had we not had the "|", the result would have been 3572 simply "yza". The '@' character can be used as filler character 3573 to place the revisiting point off the start or end of the 3574 string. Thus the following causes x to be replaced, and the 3575 cursor to be backed up by two characters.</p> 3576 <table cellspacing="0" cellpadding="8" border="1"> 3577 <tr> 3578 <td valign="top" bgcolor="#EEEEEE"><code>x → 3579 |@@y;</code></td> 3580 </tr> 3581 </table> 3582 <h4>10.3.4 <a name="Example" href="#Example" id= 3583 "Example">Example</a></h4> 3584 <p>The following shows how these features are combined together 3585 in the Transliterator "Any-Publishing". This transform converts 3586 the ASCII typewriter conventions into text more suitable for 3587 desktop publishing (in English). It turns straight quotation 3588 marks or UNIX style quotation marks into curly quotation marks, 3589 fixes multiple spaces, and converts double-hyphens into a 3590 dash.</p> 3591 <table cellspacing="0" cellpadding="8" border="1"> 3592 <tr> 3593 <td valign="top" bgcolor="#EEEEEE"><code># Variables<br> 3594 <br> 3595 $single = \' ;<br> 3596 $space = ' ' ;<br> 3597 $double = \" ;<br> 3598 $back = \` ;<br> 3599 $tab = '\u0008' ;<br> 3600 <br> 3601 # the following is for spaces, line ends, (, [, {, ...<br> 3602 $makeRight = [[:separator:][:start punctuation:][:initial 3603 punctuation:]] ;<br> 3604 <br> 3605 # fix UNIX quotes<br> 3606 <br> 3607 $back $back → “ ; # generate right d.q.m. (double quotation 3608 mark)<br> 3609 $back → ‘ ;<br> 3610 <br> 3611 # fix typewriter quotes, by context<br> 3612 <br> 3613 $makeRight { $double ↔ “ ; # convert a double to right 3614 d.q.m. after certain chars<br> 3615 ^ { $double → “ ; # convert a double at the start of the 3616 line.<br> 3617 $double ↔ ” ; # otherwise convert to a left q.m.<br> 3618 <br> 3619 $makeRight {$single} ↔ ‘ ; # do the same for s.q.m.s<br> 3620 ^ {$single} → ‘ ;<br> 3621 $single ↔ ’;<br> 3622 <br> 3623 # fix multiple spaces and hyphens<br> 3624 <br> 3625 $space {$space} → ; # collapse multiple spaces<br> 3626 '--' ↔ — ; # convert fake dash into real one</code></td> 3627 </tr> 3628 </table> 3629 <p>There is an online demo where the rules can be tested, 3630 at:</p> 3631 <p><a target="demo" href= 3632 "https://util.unicode.org/UnicodeJsps/transform.jsp">http://unicode.org/cldr/utility/transform.jsp</a></p> 3633 <h4>10.3.5 <a name="Rule_Syntax" href="#Rule_Syntax" id= 3634 "Rule_Syntax">Rule Syntax</a></h4> 3635 <p>The following describes the full format of the list of rules 3636 used to create a transform. Each rule in the list is terminated 3637 by a semicolon. The list consists of the following:</p> 3638 <ul> 3639 <li>an optional filter rule</li> 3640 <li>zero or more transform rules</li> 3641 <li>zero or more variable-definition rules</li> 3642 <li>zero or more conversion rules</li> 3643 <li>an optional inverse filter rule</li> 3644 </ul> 3645 <p>The filter rule, if present, must appear at the beginning of 3646 the list, before any of the other rules. The inverse 3647 filter rule, if present, must appear at the end of the list, 3648 after all of the other rules. The other rules may occur 3649 in any order and be freely intermixed.</p> 3650 <p>The rule list can also generate the inverse of the 3651 transform. In that case, the inverse of each of the rules is 3652 used, as described below.</p> 3653 <h4>10.3.6 <a name="Transform_Rules" href="#Transform_Rules" 3654 id="Transform_Rules">Transform Rules</a></h4> 3655 <p>Each transform rule consists of two colons followed by a 3656 transform name, which is of the form source-target. For 3657 example:</p> 3658 <table cellspacing="0" cellpadding="8" border="1"> 3659 <tr> 3660 <td valign="top" bgcolor="#EEEEEE"><code>:: NFD ;<br> 3661 :: und_Latn-und_Greek ;<br> 3662 :: Latin-Greek; # alternate form</code></td> 3663 </tr> 3664 </table> 3665 <p>If either the source or target is 'und', it can be omitted, 3666 thus 'und_NFC' is equivalent to 'NFC'. For compatibility, the 3667 English names for scripts can be used instead of the und_Latn 3668 locale name, and "Any" can be used instead of "und". Case is 3669 not significant.</p> 3670 <p>The following transforms are defined not by rules, but by 3671 the operations in the Unicode Standard, and may be used in 3672 building any other transform:</p> 3673 <blockquote> 3674 <b>Any-NFC, Any-NFD, Any-NFKD, Any-NFKC</b> - the 3675 normalization forms defined by [<a href= 3676 "https://www.unicode.org/reports/tr41/#UAX15">UAX15</a>].<br> 3677 <p><b>Any-Lower, Any-Upper, Any-Title</b> - full case 3678 transformations, defined by [<a href= 3679 "tr35.html#Unicode">Unicode</a>] Chapter 3.</p> 3680 </blockquote> 3681 <p>In addition, the following special cases are defined:</p> 3682 <blockquote> 3683 <b>Any-Null</b> - has no effect; that is, each character is 3684 left alone.<br> 3685 <b>Any-Remove</b> - maps each character to the empty string; 3686 this, removes each character. 3687 </blockquote> 3688 <p>The inverse of a transform rule uses parentheses to indicate 3689 what should be done when the inverse transform is used. For 3690 example:</p> 3691 <table cellspacing="0" cellpadding="8" border="1"> 3692 <tr> 3693 <td valign="top" bgcolor="#EEEEEE"><code>:: lower () ; # 3694 only executed for the normal<br> 3695 :: (lower) ; # only executed for the inverse<br> 3696 :: lower ; # executed for both the normal and the 3697 inverse</code></td> 3698 </tr> 3699 </table> 3700 <h4>10.3.7 <a name="Variable_Definition_Rules" href= 3701 "#Variable_Definition_Rules" id= 3702 "Variable_Definition_Rules">Variable Definition Rules</a></h4> 3703 <p>Each variable definition is of the following form:</p> 3704 <table cellspacing="0" cellpadding="8" border="1"> 3705 <tr> 3706 <td valign="top" bgcolor="#EEEEEE"><code>$variableName = 3707 contents ;</code></td> 3708 </tr> 3709 </table> 3710 <p>The variable name can contain letters and digits, but must 3711 start with a letter. More precisely, the variable names use 3712 Unicode identifiers as defined by [<a href= 3713 "https://www.unicode.org/reports/tr41/#UAX31">UAX31</a>]. The 3714 identifier properties allow for the use of foreign letters and 3715 numbers.</p> 3716 <p>The contents of a variable definition is any sequence of 3717 Unicode sets and characters or characters. For example:</p> 3718 <table cellspacing="0" cellpadding="8" border="1"> 3719 <tr> 3720 <td valign="top" bgcolor="#EEEEEE"><code>$mac = M [aA] [cC] 3721 ;</code></td> 3722 </tr> 3723 </table> 3724 <p>Variables are only replaced within other variable definition 3725 rules and within conversion rules. They have no effect on 3726 transliteration rules.</p> 3727 <h4>10.3.8 <a name="Filter_Rules" href="#Filter_Rules" id= 3728 "Filter_Rules">Filter Rules</a></h4> 3729 <p>A filter rule consists of two colons followed by a 3730 UnicodeSet. This filter is global in that only the characters 3731 matching the filter will be affected by any transform rules or 3732 conversion rules. The inverse filter rule consists of two 3733 colons followed by a UnicodeSet in parentheses. This filter is 3734 also global for the inverse transform.</p> 3735 <p>For example, the Hiragana-Latin transform can be implemented 3736 by "pivoting" through the Katakana converter, as follows:</p> 3737 <table cellspacing="0" cellpadding="8" border="1"> 3738 <tr> 3739 <td valign="top" bgcolor="#EEEEEE"><code>:: [:^Katakana:] ; 3740 # do not touch any katakana that was in the text!<br> 3741 :: Hiragana-Katakana;<br> 3742 :: Katakana-Latin;<br> 3743 :: ([:^Katakana:]) ; # do not touch any katakana that was 3744 in the text<br> 3745 3746 # for the inverse either!</code></td> 3747 </tr> 3748 </table> 3749 <p>The filters keep the transform from mistakenly converting 3750 any of the "pivot" characters. Note that this is a case where a 3751 rule list contains no conversion rules at all, just transform 3752 rules and filters.</p> 3753 <h4>10.3.9 <a name="Conversion_Rules" href="#Conversion_Rules" 3754 id="Conversion_Rules">Conversion Rules</a></h4> 3755 <p>Conversion rules can be forward, backward, or double. The 3756 complete conversion rule syntax is described below:</p> 3757 <p><b>Forward</b></p> 3758 <blockquote> 3759 <p>A forward conversion rule is of the following form:</p> 3760 <blockquote> 3761 <pre> 3762 before_context { text_to_replace } after_context → completed_result | result_to_revisit ;</pre> 3763 </blockquote> 3764 <p>If there is no before_context, then the "{" can be 3765 omitted. If there is no after_context, then the "}" can be 3766 omitted. If there is no result_to_revisit, then the "|" can 3767 be omitted. A forward conversion rule is only executed for 3768 the normal transform and is ignored when generating the 3769 inverse transform.</p> 3770 </blockquote> 3771 <p><b>Backward</b></p> 3772 <blockquote> 3773 <p>A backward conversion rule is of the following form:</p> 3774 <blockquote> 3775 <pre> 3776 completed_result | result_to_revisit ← before_context { text_to_replace } after_context ;</pre> 3777 </blockquote> 3778 <p>The same omission rules apply as in the case of forward 3779 conversion rules. A backward conversion rule is only executed 3780 for the inverse transform and is ignored when generating the 3781 normal transform.</p> 3782 </blockquote> 3783 <p><b>Dual</b></p> 3784 <blockquote> 3785 <p>A dual conversion rule combines a forward conversion rule 3786 and a backward conversion rule into one, as discussed above. 3787 It is of the form:</p> 3788 <table cellspacing="0" cellpadding="8" border="1"> 3789 <tr> 3790 <td valign="top" bgcolor="#EEEEEE"><code>a { b | c } d ↔ 3791 e { f | g } h ;</code></td> 3792 </tr> 3793 </table> 3794 <p>When generating the normal transform and the inverse, the 3795 revisit mark "|" and the before and after contexts are 3796 ignored on the sides where they do not belong. Thus, the 3797 above is exactly equivalent to the sequence of the following 3798 two rules:</p> 3799 <table cellspacing="0" cellpadding="8" border="1"> 3800 <tr> 3801 <td valign="top" bgcolor="#EEEEEE"><code>a { b c } 3802 d → f | g ;<br> 3803 b | c ← e { f g } h ; </code></td> 3804 </tr> 3805 </table> 3806 </blockquote> 3807 <h4>10.3.10 <a name= 3808 "Intermixing_Transform_Rules_and_Conversion_Rules" href= 3809 "#Intermixing_Transform_Rules_and_Conversion_Rules" id= 3810 "Intermixing_Transform_Rules_and_Conversion_Rules">Intermixing 3811 Transform Rules and Conversion Rules</a></h4> 3812 <p>Transform rules and conversion rules may be freely 3813 intermixed. Inserting a transform rule into the middle of a set 3814 of conversion rules has an important side effect.</p> 3815 <p>Normally, conversion rules are considered together as a 3816 group. The only time their order in the rule set is 3817 important is when more than one rule matches at the same point 3818 in the string. In that case, the one that occurs earlier 3819 in the rule set wins. In all other situations, when 3820 multiple rules match overlapping parts of the string, the one 3821 that matches earlier wins.</p> 3822 <p>Transform rules apply to the whole string. If you have 3823 several transform rules in a row, the first one is applied to 3824 the whole string, then the second one is applied to the whole 3825 string, and so on. To reconcile this behavior with the 3826 behavior of conversion rules, transform rules have the side 3827 effect of breaking a surrounding set of conversion rules into 3828 two groups: First all of the conversion rules before the 3829 transform rule are applied as a group to the whole string in 3830 the usual way, then the transform rule is applied to the whole 3831 string, and then the conversion rules after the transform rule 3832 are applied as a group to the whole string. For example, 3833 consider the following rules:</p> 3834 <table cellspacing="0" cellpadding="8" border="1"> 3835 <tr> 3836 <td valign="top" bgcolor="#EEEEEE"><code>abc → xyz;<br> 3837 xyz → def;<br> 3838 ::Upper;</code></td> 3839 </tr> 3840 </table> 3841 <p>If you apply these rules to “abcxyz”, you get 3842 “XYZDEF”. If you move the “::Upper;” to the middle of the 3843 rule set and change the cases accordingly, then applying this 3844 to “abcxyz” produces “DEFDEF”.</p> 3845 <table cellspacing="0" cellpadding="8" border="1"> 3846 <tr> 3847 <td valign="top" bgcolor="#EEEEEE"><code>abc → xyz;<br> 3848 ::Upper;<br> 3849 XYZ → DEF;</code></td> 3850 </tr> 3851 </table> 3852 <p>This is because “::Upper;” causes the transliterator to 3853 reset to the beginning of the string. The first rule turns the 3854 string into “xyzxyz”, the second rule upper cases the whole 3855 thing to “XYZXYZ”, and the third rule turns this into 3856 “DEFDEF”.</p> 3857 <p>This can be useful when a transform naturally occurs in 3858 multiple “passes.” Consider this rule set:</p> 3859 <table cellspacing="0" cellpadding="8" border="1"> 3860 <tr> 3861 <td valign="top" bgcolor="#EEEEEE"><code>[:Separator:]* → ' 3862 ';<br> 3863 'high school' → 'H.S.';<br> 3864 'middle school' → 'M.S.';<br> 3865 'elementary school' → 'E.S.';</code></td> 3866 </tr> 3867 </table> 3868 <p>If you apply this rule to “high school”, you get “H.S.”, but 3869 if you apply it to “high school” (with two spaces), you 3870 just get “high school” (with one space). To have “high 3871 school” (with two spaces) turn into “H.S.”, you'd either have 3872 to have the first rule back up some arbitrary distance (far 3873 enough to see “elementary”, if you want all the rules to work), 3874 or you have to include the whole left-hand side of the first 3875 rule in the other rules, which can make them hard to read and 3876 maintain:</p> 3877 <table cellspacing="0" cellpadding="8" border="1"> 3878 <tr> 3879 <td valign="top" bgcolor="#EEEEEE"><code>$space = 3880 [:Separator:]*;<br> 3881 high $space school → 'H.S.';<br> 3882 middle $space school → 'M.S.';<br> 3883 elementary $space school → 'E.S.';</code></td> 3884 </tr> 3885 </table> 3886 <p>Instead, you can simply insert “ <code>::Null;</code> ” in 3887 order to get things to work right:</p> 3888 <table cellspacing="0" cellpadding="8" border="1"> 3889 <tr> 3890 <td valign="top" bgcolor="#EEEEEE"><code>[:Separator:]* → ' 3891 ';<br> 3892 ::Null;<br> 3893 'high school' → 'H.S.';<br> 3894 'middle school' → 'M.S.';<br> 3895 'elementary school' → 'E.S.';</code></td> 3896 </tr> 3897 </table> 3898 <p>The “::Null;” has no effect of its own (the null transform, 3899 by definition, does not do anything), but it splits the other 3900 rules into two “passes”: The first rule is applied to the whole 3901 string, normalizing all runs of white space into single spaces, 3902 and then we start over at the beginning of the string to look 3903 for the phrases. “high school” (with four 3904 spaces) gets correctly converted to “H.S.”.</p> 3905 <p>This can also sometimes be useful with rules that have 3906 overlapping domains. Consider this rule set from 3907 before:</p> 3908 <table cellspacing="0" cellpadding="8" border="1"> 3909 <tr> 3910 <td valign="top" bgcolor="#EEEEEE"><code>sch → sh ;<br> 3911 ss → z ;</code></td> 3912 </tr> 3913 </table> 3914 <p>Apply this rule to “bassch” results in “bazch” because “ss” 3915 matches earlier in the string than “sch”. If you really wanted 3916 “bassh”—that is, if you wanted the first rule to win even when 3917 the second rule matches earlier in the string, you'd either 3918 have to add another rule for this special case...</p> 3919 <table cellspacing="0" cellpadding="8" border="1"> 3920 <tr> 3921 <td valign="top" bgcolor="#EEEEEE"><code>sch → sh ;<br> 3922 ssch → ssh;<br> 3923 ss → z ;</code></td> 3924 </tr> 3925 </table> 3926 <p>...or you could use a transform rule to apply the 3927 conversions in two passes:</p> 3928 <table cellspacing="0" cellpadding="8" border="1"> 3929 <tr> 3930 <td valign="top" bgcolor="#EEEEEE"><code>sch → sh ;<br> 3931 ::Null;<br> 3932 ss → z ;</code></td> 3933 </tr> 3934 </table> 3935 <h4>10.3.11 <a name="Inverse_Summary" href="#Inverse_Summary" 3936 id="Inverse_Summary">Inverse Summary</a></h4> 3937 <p>The following table shows how the same rule list generates 3938 two different transforms, where the inverse is restated in 3939 terms of forward rules (this is a contrived example, simply to 3940 show the reordering):</p> 3941 <table> 3942 <tr bgcolor="#99CCFF"> 3943 <th bgcolor="#CCCCCC">Original Rules</th> 3944 <th bgcolor="#CCCCCC">Forward</th> 3945 <th bgcolor="#CCCCCC">Inverse</th> 3946 </tr> 3947 <tr bgcolor="#99CCFF"> 3948 <td bgcolor="#EEEEEE"><code>:: [:Uppercase Letter:] ;<br> 3949 :: latin-greek ;<br> 3950 :: greek-japanese ;<br> 3951 x ↔ y ;<br> 3952 z → w ;<br> 3953 r ← m ;<br> 3954 :: upper;<br> 3955 a → b ;<br> 3956 c ↔ d ;<br> 3957 :: any-publishing ;<br> 3958 :: ([:Number:]) ;</code></td> 3959 <td bgcolor="#EEEEEE"><code>:: [:Uppercase Letter:] ;<br> 3960 :: latin-greek ;<br> 3961 :: greek-japanese ;<br> 3962 x → y ;<br> 3963 z → w ;<br> 3964 :: upper ;<br> 3965 a → b ;<br> 3966 c → d ;<br> 3967 :: any-publishing ;<br></code></td> 3968 <td bgcolor="#EEEEEE"><code>:: [:Number:] ;<br> 3969 :: publishing-any ;<br> 3970 d → c ;<br> 3971 :: lower ;<br> 3972 y → x ;<br> 3973 m → r ;<br> 3974 :: japanese-greek ;<br> 3975 :: greek-latin ;<br></code></td> 3976 </tr> 3977 </table> 3978 <p>Note how the irrelevant rules (the inverse filter rule and 3979 the rules containing ←) are omitted (ignored, actually) in the 3980 forward direction, and notice how things are reversed: the 3981 transform rules are inverted and happen in the opposite order, 3982 and the groups of conversion rules are also executed in the 3983 opposite relative order (although the rules within each group 3984 are executed in the same order).</p> 3985 <h2>11 <a name="ListPatterns" href="#ListPatterns" id= 3986 "ListPatterns">List Patterns</a></h2> 3987 <p class="dtd"><!ELEMENT listPatterns (alias | 3988 (listPattern*, special*)) ></p> 3989 <p class="dtd"><!ELEMENT listPattern (alias | 3990 (listPatternPart*, special*)) ><br> 3991 <!ATTLIST listPattern type (NMTOKEN) #IMPLIED ></p> 3992 <p class="dtd"><!ELEMENT listPatternPart ( #PCDATA ) 3993 ><br> 3994 <!ATTLIST listPatternPart type (start | middle | end | 2 | 3995 3) #REQUIRED ></p> 3996 <p>List patterns can be used to format variable-length lists of 3997 things in a locale-sensitive manner, such as "Monday, Tuesday, 3998 Friday, and Saturday" (in English) versus "lundi, mardi, 3999 vendredi et samedi" (in French). For example, consider the 4000 following example:</p> 4001 <pre class="example"><listPatterns> 4002 <listPattern> 4003 <listPatternPart type="2">{0} and {1}</listPatternPart> 4004 <listPatternPart type="start">{0}, {1}</listPatternPart> 4005 <listPatternPart type="middle">{0}, {1}</listPatternPart> 4006 <listPatternPart type="end">{0}, and {1}</listPatternPart> 4007 </listPattern> 4008</listPatterns></pre> 4009 <p>The data is used as follows: If there is a type type matches 4010 exactly the number of elements in the desired list (such as "2" 4011 in the above list), then use that pattern. Otherwise,</p> 4012 <ol> 4013 <li>Format the last two elements with the "end" format.</li> 4014 <li>Then use middle format to add on subsequent elements 4015 working towards the front, all but the very first element. 4016 That is, {1} is what you've already done, and {0} is the 4017 previous element.</li> 4018 <li>Then use "start" to add the front element, again with {1} 4019 as what you've done so far, and {0} is the first 4020 element.</li> 4021 </ol> 4022 <p>Thus a list (a,b,c,...m, n) is formatted as: 4023 <code>start(a,middle(b,middle(c,middle(...end(m, n))...)))</code></p> 4024 <p>More sophisticated implementations can customize the process to improve the results for languages where context is important. For example:</p> 4025 <table> 4026 <tr> 4027 <td rowspan="3">Spanish 4028 </td> 4029 <td>AND 4030 </td> 4031 <td>Use ‘e’ instead of ‘y’ in the listPatternPart for "end" and "2" in either of the following cases: 4032 <ol> 4033 4034 <li>The value substituted for {1} starts with ‘i’ 4035 <ol> 4036 <li><em>fuerte <strong>e</strong> indomable, </em>not <em>fuerte <strong>y</strong> indomable</em></li> 4037 </ol></li> 4038 <li>The value substituted for {1} starts with ‘hi’, but not with ‘hie’ or ‘hia’ 4039 <ol> 4040 <li><em>tos <strong>e</strong> hipo,</em> not <em>tos <strong>y</strong> hipo </em></li> 4041 <li><em>gua <strong>y</strong> hielo,</em> not <em>agua <strong>e</strong> hielo </em></li> 4042 </ol> 4043 </li> 4044 </ol></td> 4045 </tr> 4046 <tr> 4047 <td>OR 4048 </td> 4049 <td>Use ‘u’ instead of ‘o’ in the listPatternPart for "end" and "2" in any of the following cases: 4050 <ol> 4051 4052 <li>The value substituted for {1} starts with ‘o’ or ‘ho’ 4053 <ol> 4054 <li><em>delfines <strong>u</strong> orcas,</em> not <em>deflines <strong>o</strong> orcas</em></li> 4055 <li><em>mañana <strong>u</strong> hoy,</em> not <em>mañana <strong>o</strong> hoy</em></li> 4056 </ol> 4057 </li> 4058 <li> The value substituted for {1} starts with ‘8’ 4059 <ol> 4060 <li><em>6 <strong>u</strong> 8,</em> not <em>6 <strong>o</strong> 8</em></li> 4061 </ol> 4062 </li> 4063 <li>The value substituted for {1} starts with ‘11’ where the numeric value is 11 x 10<sup>3×y</sup> 4064 (eg 11 thousand, 11.23 million, ...) 4065 <ol> 4066 <li><em>10 <strong>u</strong> 11,</em> not <em>10 <strong>o</strong> 11</em></li> 4067 <li><em>10 <strong>u</strong> 11.000,</em> not <em>10 <strong>o</strong> 11.000</em></li> 4068 <li><em>10 <strong>o</strong> 111,</em> not <em>10 <strong>u</strong> 111</em></li> 4069 </ol> 4070 </li></ol> 4071 4072 </td> 4073 </tr> 4074 <tr> 4075 <td colspan="2">See <a href='https://www.rae.es/consultas/cambio-de-la-y-copulativa-en-e'>Cambio de la y copulativa en e</a><br> 4076 <strong>Note: </strong>more advanced implementations may also consider the pronunciation, such as foreign words where the ‘h’ is not mute.</td> 4077 </tr> 4078 <tr> 4079 <td rowspan="2">Hebrew 4080 </td> 4081 <td>AND 4082 </td> 4083 <td>Use ‘-ו’ instead of ‘ו’ in the listPatternPart for "end" and "2" in the following case: 4084 <ol> 4085 <li>if the value substituted for {1} starts with something other than a Hebrew letter, such as a digit (0-9) or a Latin-script letter 4086 <ol> 4087 <li><em>one hour and two minutes = "שעה ושתי דקות"</em></li> 4088 <li><em>one hour and 9 minutes = "שעה ו-9 דקות” </em></li> 4089 </ol></li> 4090 4091 4092 </ol></td> 4093 </tr> 4094 <tr> 4095 <td colspan="2">See <a href="https://hebrew-academy.org.il/topic/hahlatot/punctuation/#target-3475">https://hebrew-academy.org.il/topic/hahlatot/punctuation/#target-3475</a></td> 4096 </tr> 4097 </table><br> 4098 <p>The following type attributes are in use:</p> 4099 <table border="1" cellpadding="2" cellspacing="0" class= 4100 'simple'> 4101 <tr> 4102 <th>type attribute value</th> 4103 <th>Description</th> 4104 <th>Examples</th> 4105 </tr> 4106 <tr> 4107 <td nowrap>standard (or no <strong>type</strong>)</td> 4108 <td>A typical 'and' list for arbitrary placeholders</td> 4109 <td nowrap><em>January, February, and March</em></td> 4110 </tr> 4111 <tr> 4112 <td>standard-short</td> 4113 <td>A short version of an 'and' list, suitable for use with 4114 short or abbreviated placeholder values</td> 4115 <td><em>Jan., Feb., and Mar.</em></td> 4116 </tr> 4117 <tr> 4118 <td>standard-narrow</td> 4119 <td>A yet shorter version of a short 'and' list (where possible)</td> 4120 <td><em>Jan., Feb., Mar.</em></td> 4121 </tr> 4122 <tr> 4123 <td>or</td> 4124 <td>A typical 'or' list for arbitrary placeholders</td> 4125 <td><em>January, February, or March</em></td> 4126 </tr> 4127 <tr> 4128 <td>or-short</td> 4129 <td>A short version of an 'or' list</td> 4130 <td><em>Jan., Feb., or Mar.</em></td> 4131 </tr> 4132 <tr> 4133 <td>or-narrow</td> 4134 <td>A yet shorter version of a short 'or' list (where possible)</td> 4135 <td><em>Jan., Feb., or Mar.</em></td> 4136 </tr> 4137 <tr> 4138 <td>unit</td> 4139 <td>A list suitable for wide units</td> 4140 <td><em>3 feet, 7 inches</em></td> 4141 </tr> 4142 <tr> 4143 <td>unit-short</td> 4144 <td>A list suitable for short units</td> 4145 <td><em>3 ft, 7 in</em></td> 4146 </tr> 4147 <tr> 4148 <td>unit-narrow</td> 4149 <td>A list suitable for narrow units, where space on the 4150 screen is very limited.</td> 4151 <td><em>3′ 7″</em></td> 4152 </tr> 4153 </table> 4154 <p>In many languages there may not be a difference among many 4155 of these lists. In others, the spacing, the length or presence 4156 or a conjunction, and the separators may change.</p> 4157 <h3>11.1 <a name="List_Gender" href="#List_Gender" id= 4158 "List_Gender">Gender of Lists</a></h3> 4159 <p class="dtd"><!-- Gender List support --><br> 4160 <!ELEMENT gender ( personList+ ) ><br> 4161 <!ELEMENT personList EMPTY ><br> 4162 <!ATTLIST personList type ( neutral | mixedNeutral | 4163 maleTaints ) #REQUIRED ><br> 4164 <!ATTLIST personList locales NMTOKENS #REQUIRED ><br></p> 4165 <p>This can be used to determine the gender of a list of 2 or 4166 more persons, such as "Tom and Mary", for use with 4167 gender-selection messages. For example,</p> 4168 <pre class="example"> 4169 <supplementalData> 4170 <gender> 4171 <!-- neutral: gender(list) = other --> 4172 <personList type="neutral" locales="af da en..."/> 4173 4174 <!-- mixedNeutral: gender(all male) = male, gender(all female) = female, otherwise gender(list) = other --> 4175 <personList type="mixedNeutral" locales="el"/> 4176 4177 <!-- maleTaints: gender(all female) = female, otherwise gender(list) = male --> 4178 <personList type="maleTaints" locales="ar ca..."/> 4179 </gender> 4180 </supplementalData></pre> 4181 <p>There are three ways the gender of a list can be 4182 formatted:</p> 4183 <ol> 4184 <li><b>neutral:</b> A gender-independent "other" form will be 4185 used for the list.</li> 4186 <li><b>mixedNeutral:</b> If the elements of the list are all 4187 male, "male" form is used for the list. If all the elements 4188 of the lists are female, "female" form is used. If the list 4189 has a mix of male, female and neutral names, the "other" form 4190 is used.</li> 4191 <li><b>maleTaints:</b> If all the elements of the lists are 4192 female, "female" form is used, otherwise the "male" form is 4193 used.</li> 4194 </ol> 4195 <h2>12 <a name="Context_Transform_Elements" href= 4196 "#Context_Transform_Elements" id= 4197 "Context_Transform_Elements">ContextTransform Elements</a></h2> 4198 <p class="dtd"><!ELEMENT contextTransforms ( alias | 4199 (contextTransformUsage*, special*)) ><br> 4200 <!ELEMENT contextTransformUsage ( alias | 4201 (contextTransform*, special*)) ><br> 4202 <!ATTLIST contextTransformUsage type CDATA #REQUIRED 4203 ><br> 4204 <!ELEMENT contextTransform ( #PCDATA ) ><br> 4205 <!ATTLIST contextTransform type ( uiListOrMenu | stand-alone 4206 ) #REQUIRED ></p> 4207 <p>CLDR locale elements provide data for display names or 4208 symbols in many categories. The default capitalization for 4209 these elements is intended to be the form used in the middle of 4210 running text. In many languages, other capitalization may be 4211 required in other contexts, depending on the type of name or 4212 symbol.</p> 4213 <p>Each <contextTransformUsage> element’s type attribute 4214 specifies a category of data from the table below; the element 4215 includes one or more <contextTransform> elements that 4216 specify how to perform capitalization of this category of data 4217 in different contexts. The <contextTransform> elements 4218 are needed primarily for cases in which the capitalization is 4219 other than the default form used in the middle of running text. 4220 However, it is also useful to mark cases in which it is 4221 <em>known</em> that no transformation from this default form is 4222 needed; this may be necessary, for example, to override the 4223 transformation specified by a parent locale. The following 4224 values are currently defined for the <contextTransform> 4225 element:</p> 4226 <ul> 4227 <li>"titlecase-firstword" designates the case in which raw 4228 CLDR text that is in middle-of-sentence form, typically 4229 lowercase, needs to have its first word titlecased.</li> 4230 <li>"no-change" designates the case in which it is known that 4231 no change from the raw CLDR text (middle-of-sentence form) is 4232 needed.</li> 4233 </ul> 4234 <p>Four contexts for capitalization behavior are currently 4235 identified. Two need no data, and hence have no corresponding 4236 <contextTransform> elements:</p> 4237 <ul> 4238 <li>In the middle of running text: This is the default form, 4239 so no additional data is required.</li> 4240 <li>At the beginning of a complete sentence: The initial word 4241 is titlecased, no additional data is required to indicate 4242 this.</li> 4243 </ul> 4244 <p>Two other contexts require <contextTransform> elements 4245 if their capitalization behavior is other than the default for 4246 running text. The context is identified by the type attribute, 4247 as follows:</p> 4248 <ul> 4249 <li>uiListOrMenu: Capitalization appropriate to a 4250 user-interface list or menu.</li> 4251 <li>stand-alone: Capitalization appropriate to an isolated 4252 user-interface element (e.g. an isolated name on a calendar 4253 page)</li> 4254 </ul> 4255 <p>Example:</p> 4256 <pre> <contextTransforms> 4257 <contextTransformUsage type="languages"> 4258 <contextTransform type="uiListOrMenu">titlecase-firstword</contextTransform> 4259 <contextTransform type="stand-alone">titlecase-firstword</contextTransform> 4260 </contextTransformUsage> 4261 <contextTransformUsage type="month-format-except-narrow"> 4262 <contextTransform type="uiListOrMenu">titlecase-firstword</contextTransform> 4263 </contextTransformUsage> 4264 <contextTransformUsage type="month-standalone-except-narrow"> 4265 <contextTransform type="uiListOrMenu">titlecase-firstword</contextTransform> 4266 </contextTransformUsage> 4267 </contextTransforms></pre> 4268 <table cellspacing="0" cellpadding="2" border="1" class= 4269 'simple'> 4270 <caption> 4271 <a name="contextTransformUsage_type_attribute_values" href= 4272 "#contextTransformUsage_type_attribute_values" id= 4273 "contextTransformUsage_type_attribute_values">Element 4274 contextTransformUsage type attribute values</a> 4275 </caption> 4276 <tr> 4277 <th>type attribute value</th> 4278 <th>Description</th> 4279 </tr> 4280 <tr> 4281 <td>all</td> 4282 <td>Special value, indicates that the specified 4283 transformation applies to all of the categories below</td> 4284 </tr> 4285 <tr> 4286 <td>language</td> 4287 <td>localeDisplayNames language names</td> 4288 </tr> 4289 <tr> 4290 <td>script</td> 4291 <td>localeDisplayNames script names</td> 4292 </tr> 4293 <tr> 4294 <td>territory</td> 4295 <td>localeDisplayNames territory names</td> 4296 </tr> 4297 <tr> 4298 <td>variant</td> 4299 <td>localeDisplayNames variant names</td> 4300 </tr> 4301 <tr> 4302 <td>key</td> 4303 <td>localeDisplayNames key names</td> 4304 </tr> 4305 <tr> 4306 <td>keyValue</td> 4307 <td>localeDisplayNames key value type names</td> 4308 </tr> 4309 <tr> 4310 <td>month-format-except-narrow</td> 4311 <td>dates/calendars/calendar[type=*]/months format wide and 4312 abbreviated month names</td> 4313 </tr> 4314 <tr> 4315 <td>month-standalone-except-narrow</td> 4316 <td>dates/calendars/calendar[type=*]/months stand-alone 4317 wide and abbreviated month names</td> 4318 </tr> 4319 <tr> 4320 <td>month-narrow</td> 4321 <td>dates/calendars/calendar[type=*]/months format and 4322 stand-alone narrow month names</td> 4323 </tr> 4324 <tr> 4325 <td>day-format-except-narrow</td> 4326 <td>dates/calendars/calendar[type=*]/days format wide and 4327 abbreviated day names</td> 4328 </tr> 4329 <tr> 4330 <td>day-standalone-except-narrow</td> 4331 <td>dates/calendars/calendar[type=*]/days stand-alone wide 4332 and abbreviated day names</td> 4333 </tr> 4334 <tr> 4335 <td>day-narrow</td> 4336 <td>dates/calendars/calendar[type=*]/days format and 4337 stand-alone narrow day names</td> 4338 </tr> 4339 <tr> 4340 <td>era-name</td> 4341 <td>dates/calendars/calendar[type=*]/eras (wide) era 4342 names</td> 4343 </tr> 4344 <tr> 4345 <td>era-abbr</td> 4346 <td>dates/calendars/calendar[type=*]/eras abbreviated era 4347 names</td> 4348 </tr> 4349 <tr> 4350 <td>era-narrow</td> 4351 <td>dates/calendars/calendar[type=*]/eras narrow era 4352 names</td> 4353 </tr> 4354 <tr> 4355 <td>quarter-format-wide</td> 4356 <td>dates/calendars/calendar[type=*]/quarters format wide 4357 quarter names</td> 4358 </tr> 4359 <tr> 4360 <td>quarter-standalone-wide</td> 4361 <td>dates/calendars/calendar[type=*]/quarters stand-alone 4362 wide quarter names</td> 4363 </tr> 4364 <tr> 4365 <td>quarter-abbreviated</td> 4366 <td>dates/calendars/calendar[type=*]/quarters format and 4367 stand-alone abbreviated quarter names</td> 4368 </tr> 4369 <tr> 4370 <td>quarter-narrow</td> 4371 <td>dates/calendars/calendar[type=*]/quarters format and 4372 stand-alone narrow quarter names</td> 4373 </tr> 4374 <tr> 4375 <td>calendar-field</td> 4376 <td>dates/fields/field[type=*]/displayName field names<br> 4377 (for relative forms see type "tense" below)</td> 4378 </tr> 4379 <tr> 4380 <td>zone-exemplarCity</td> 4381 <td>dates/timeZoneNames/zone[type=*]/exemplarCity city 4382 names</td> 4383 </tr> 4384 <tr> 4385 <td>zone-long</td> 4386 <td>dates/timeZoneNames/zone[type=*]/long zone names</td> 4387 </tr> 4388 <tr> 4389 <td>zone-short</td> 4390 <td>dates/timeZoneNames/zone[type=*]/short zone names</td> 4391 </tr> 4392 <tr> 4393 <td>metazone-long</td> 4394 <td>dates/timeZoneNames/metazone[type=*]/long metazone 4395 names</td> 4396 </tr> 4397 <tr> 4398 <td>metazone-short</td> 4399 <td>dates/timeZoneNames/metazone[type=*]/short metazone 4400 names</td> 4401 </tr> 4402 <tr> 4403 <td>symbol</td> 4404 <td>numbers/currencies/currency[type=*]/symbol symbol 4405 names</td> 4406 </tr> 4407 <tr> 4408 <td>currencyName</td> 4409 <td>numbers/currencies/currency[type=*]/displayName 4410 currency names</td> 4411 </tr> 4412 <tr> 4413 <td>currencyName-count</td> 4414 <td> 4415 numbers/currencies/currency[type=*]/displayName[count=*] 4416 currency names for use with count</td> 4417 </tr> 4418 <tr> 4419 <td>relative</td> 4420 <td>dates/fields/field[type=*]/relative and 4421 dates/fields/field[type=*]/relativeTime relative field 4422 names</td> 4423 </tr> 4424 <tr> 4425 <td>unit-pattern</td> 4426 <td> 4427 units/unitLength[type=*]/unit[type=*]/unitPattern[count=*] 4428 unit names</td> 4429 </tr> 4430 <tr> 4431 <td>number-spellout</td> 4432 <td>rbnf/rulesetGrouping[type=*]/ruleset[type=*]/rbnfrule 4433 number spellout rules</td> 4434 </tr> 4435 </table> 4436 <h2>13 <a name="Choice_Patterns" href="#Choice_Patterns" id= 4437 "Choice_Patterns">Choice Patterns</a></h2> 4438 <p>A choice pattern is a string that chooses among a number of 4439 strings, based on numeric value. It has the following form:</p> 4440 <p><choice_pattern> = <choice> ( '|' <choice> 4441 )*<br> 4442 <choice> = 4443 <number><relation><string><br> 4444 <number> = ('+' | '-')? (<font size="3">'∞' | [0-9]+ ('.' 4445 [0-9]+)?)<br> 4446 <relation> = '<' | '</font> <span style= 4447 "color: blue">≤'</span></p> 4448 <p>The interpretation of a choice pattern is that given a 4449 number N, the pattern is scanned from right to left, for each 4450 choice evaluating <number> <relation> N. The first 4451 choice that matches results in the corresponding string. If no 4452 match is found, then the first string is used. For example:</p> 4453 <table border="1" cellpadding="0" cellspacing="0"> 4454 <tr> 4455 <td width="33%">Pattern</td> 4456 <td width="33%">N</td> 4457 <td width="34%">Result</td> 4458 </tr> 4459 <tr> 4460 <td width="33%" rowspan="4">0≤Rf|1≤Ru|1<Re</td> 4461 <td width="33%">-<font size="3">∞,</font> -3, -1, 4462 -0.000001</td> 4463 <td width="34%">Rf (defaulted to first string)</td> 4464 </tr> 4465 <tr> 4466 <td width="33%">0, 0.01, 0.9999</td> 4467 <td width="34%">Rf</td> 4468 </tr> 4469 <tr> 4470 <td width="33%">1</td> 4471 <td width="34%">Ru</td> 4472 </tr> 4473 <tr> 4474 <td width="33%">1.00001, 5, 99, <font size= 4475 "3">∞</font></td> 4476 <td width="34%">Re</td> 4477 </tr> 4478 </table> 4479 <p>Quoting is done using ' characters, as in date or number 4480 formats.</p> 4481 <h2>14 <a name="Annotations" href="#Annotations" id= 4482 "Annotations">Annotations and Labels</a></h2> 4483 <p>Annotations provide information about characters, typically 4484 used in input. For example, on a mobile keyboard they can be 4485 used to do completion. They are typically used for symbols, 4486 especially emoji characters.</p> 4487 <p>For more information, see version 5.0 or <a href= 4488 "https://unicode.org/reports/tr51/">UTR #51, Unicode Emoji</a>. 4489 (Note that during the period between the publication of CLDR 4490 v31 and that of Emoji 5.0, the “Latest Proposed Update” link 4491 should be used to get to the draft specification for Emoji 4492 5.0.)<br></p> 4493 <p class="dtd"><!ELEMENT annotations ( annotation* ) 4494 ></p> 4495 <p class="dtd"><!ELEMENT annotation ( #PCDATA ) ></p> 4496 <p class="dtd"><!ATTLIST annotation cp CDATA #REQUIRED 4497 ></p> 4498 <p class="dtd"><!ATTLIST annotation type (tts) #IMPLIED 4499 ></p> 4500 <p>There are two kinds of annotations: <strong>short 4501 names</strong>, and <strong>keywords</strong>.</p> 4502 <p>With an attribute <strong>type="tts"</strong>, the value is 4503 a <strong>short name</strong>, such as one that can be used for 4504 text-to-speech. It should be treated as one of the element 4505 values for other purposes.</p> 4506 <p>When there is no <strong>type</strong> attribute, the value 4507 is a set of <strong>keywords</strong>, delimited by |. Spaces 4508 around each element are to be trimmed. The 4509 <strong>keywords</strong> are words associated with the 4510 character(s) that might be used in searching for the character, 4511 or in predictive typing on keyboards. The short name itself can 4512 be used as a keyword.</p> 4513 <p>Here is an example from German:</p> 4514 <pre class="example"> 4515<annotation cp="��">schlecht | Hand | Daumen | nach unten</annotation> 4516<annotation cp="��" type="tts">Daumen runter</annotation> 4517</pre> 4518 <p>The cp attribute value has two formats: either a single 4519 string, or if contained within […] a UnicodeSet. The latter 4520 format can contain multiple code points or strings. A code 4521 point pr string can occur in multiple annotation element 4522 <strong>cp</strong> values, such as the following, which also 4523 contains the "thumbs down" character.</p> 4524 <pre class="example"> 4525 <span><annotation cp='[☝✊-✍��-����-����������������������]'>hand</annotation></span></pre> 4526 <p>Both for short names and keywords, values do not have to 4527 match between different languages. They should be the most 4528 common values that people using <em>that</em> language would 4529 associated with those characters. For example, a "black heart" 4530 might have the association of "wicked" in English, but not in 4531 some other languages.</p> 4532 <p>The cp value may contain sequences, but does not contain any 4533 Emoji or Text Variant (VS15 & VS16) characters. All such 4534 characters should be removed before looking up any short names 4535 and keywords.</p> 4536 <h3>14.1 <a name="SynthesizingNames" href="#SynthesizingNames" 4537 id="SynthesizingNames">Synthesizing Sequence Names</a></h3> 4538 <p>Many emoji are represented by sequences of characters. When 4539 there are no annotation elements for that string, the short 4540 name can be synthesized as follows. <strong>Note:</strong> The 4541 process details may change after the release of this 4542 specification, and may further change in the future if other 4543 sequences are added. Please see the <a href= 4544 'https://sites.google.com/site/cldr/index/downloads/cldr-30#TOC-Known-Issues'> 4545 Known Issues</a> section of the CLDR download page for any 4546 updates.</p> 4547 <ol> 4548 <li>If <strong>sequence</strong> is an <strong>emoji flag 4549 sequence</strong>, look up the territory name in CLDR for the 4550 corresponding ASCII characters and return as the short name. 4551 For example, the regional indicator symbols P+F would map to 4552 “Französisch-Polynesien” in German.</li> 4553 <li>If <strong>sequence</strong> is an <strong>emoji tag 4554 sequence</strong>, look up the subdivision name in CLDR for 4555 the corresponding ASCII characters and return as the short 4556 name. For example, the TAG characters gbsct would map to 4557 “Schottland” in German.</li> 4558 <li>If <strong>sequence</strong> is a keycap sequence or ��, 4559 use the characterLabel for "keycap" as the 4560 <strong>prefixName</strong> and set the 4561 <strong>suffix</strong> to be the sequence (or "10" in the 4562 case of ��), then go to step 8.</li> 4563 <li>Let <strong>suffix</strong> and 4564 <strong>prefixName</strong> be "".</li> 4565 <li>If <strong>sequence</strong> contains any emoji 4566 modifiers, move them (in order) into <strong>suffix</strong>, 4567 removing them from <strong>sequence</strong>.</li> 4568 <li>If <strong>sequence</strong> is a "KISS", "HEART", "FAMILY", or "HOLDING HANDS" emoji ZWJ sequence, move the characters in <strong> 4569 sequence</strong> to the front of <strong>suffix</strong>, 4570 and set the <strong>sequence</strong> to be "��", "��", or 4571 "��" respectively, and go to step 7. 4572 <ol> 4573 <li>A KISS sequence contains ZWJ, "��", and "❤", which are 4574 skipped in moving to <strong>suffix</strong>.</li> 4575 <li>A HEART sequence contains ZWJ and "❤", which are 4576 skipped in moving to <strong>suffix</strong>.</li> 4577 <li>A HOLDING HANDS sequence contains ZWJ+��+ZWJ, which are skipped in moving to <strong>suffix</strong>.</li> 4578 <li>A FAMILY sequence contains only characters from the 4579 set {��, ��, ��, ��, ��, ��, ��}. Nothing is skipped in moving 4580 to <strong>suffix</strong>, except ZWJ.</li> 4581 </ol> 4582 </li> 4583 <li>If <strong>sequence</strong> ends with ♂ or ♀, and does 4584 not have a name, remove the ♂ or ♀ and move the name for "��" 4585 or "��" respectively to the start of 4586 <strong>prefixName</strong>.</li> 4587 <li>Transform <strong>sequence</strong> and append to 4588 <strong>prefixName</strong>, by successively getting names 4589 for the longest subsequences, skipping any singleton ZWJ 4590 characters. If there is more than one name, use the 4591 listPattern for unit-short, type=2 to link them.</li> 4592 <li>Transform <strong>suffix</strong> into 4593 <strong>suffixName</strong> in the same manner.</li> 4594 <li>If both the <strong>prefixName</strong> and 4595 <strong>suffixName</strong> are non-empty, form the name by 4596 joining them with the "category-list" characterLabelPattern 4597 and return it. Otherwise return whichever of them is 4598 non-empty.</li> 4599 </ol> 4600 <p>The synthesized keywords can follow a similar process.</p> 4601 <ol> 4602 <li>For an <strong>emoji flag sequence</strong> or 4603 <strong>emoji tag sequence</strong> representing a 4604 subdivision, use "flag".</li> 4605 <li>For keycap sequences, use "keycap".</li> 4606 <li>For other sequences, add the keywords for the 4607 subsequences used to get the short names for 4608 <strong>prefixName</strong>, and the short names used for 4609 <strong>suffixName</strong>.</li> 4610 </ol> 4611 <p>Some examples for English data (v30) are given in the 4612 following table.</p> 4613 <table cellspacing="0" cellpadding="2" border="1"> 4614 <caption> 4615 Synthesized Emoji Sequence Names 4616 </caption> 4617 <tbody> 4618 <tr> 4619 <th>Sequence</th> 4620 <th>Short Name</th> 4621 <th>Keywords</th> 4622 </tr> 4623 <tr> 4624 <td>����</td> 4625 <td>European Union</td> 4626 <td>flag</td> 4627 </tr> 4628 <tr> 4629 <td>#️⃣</td> 4630 <td>keycap: #</td> 4631 <td>keycap</td> 4632 </tr> 4633 <tr> 4634 <td>9️⃣</td> 4635 <td>keycap: 9</td> 4636 <td>keycap</td> 4637 </tr> 4638 <tr> 4639 <td>��</td> 4640 <td>kiss</td> 4641 <td>couple</td> 4642 </tr> 4643 <tr> 4644 <td>��❤️����</td> 4645 <td>kiss: woman, woman</td> 4646 <td>couple, woman</td> 4647 </tr> 4648 <tr> 4649 <td>��</td> 4650 <td>couple with heart</td> 4651 <td>love, couple</td> 4652 </tr> 4653 <tr> 4654 <td>��❤️��</td> 4655 <td>couple with heart: woman, woman</td> 4656 <td>love, couple, woman</td> 4657 </tr> 4658 <tr> 4659 <td>��</td> 4660 <td>family</td> 4661 <td>family</td> 4662 </tr> 4663 <tr> 4664 <td>������</td> 4665 <td>family: woman, woman, girl</td> 4666 <td>woman, family, girl</td> 4667 </tr> 4668 <tr> 4669 <td>����</td> 4670 <td>boy: light skin tone</td> 4671 <td>young, light skin tone, boy</td> 4672 </tr> 4673 <tr> 4674 <td>����</td> 4675 <td>woman: dark skin tone</td> 4676 <td>woman, dark skin tone</td> 4677 </tr> 4678 <tr> 4679 <td>��⚖</td> 4680 <td>man judge</td> 4681 <td>scales, justice, man</td> 4682 </tr> 4683 <tr> 4684 <td>����⚖</td> 4685 <td>man judge: dark skin tone</td> 4686 <td>scales, justice, dark skin tone, man</td> 4687 </tr> 4688 <tr> 4689 <td>��⚖</td> 4690 <td>woman judge</td> 4691 <td>woman, scales, judge</td> 4692 </tr> 4693 <tr> 4694 <td>����⚖</td> 4695 <td>woman judge: medium-light skin tone</td> 4696 <td>woman, scales, medium-light skin tone, judge</td> 4697 </tr> 4698 <tr> 4699 <td>��</td> 4700 <td>police officer</td> 4701 <td>police, cop, officer</td> 4702 </tr> 4703 <tr> 4704 <td>����</td> 4705 <td>police officer: dark skin tone</td> 4706 <td>police, cop, officer, dark skin tone</td> 4707 </tr> 4708 <tr> 4709 <td>��♂️</td> 4710 <td>man police officer</td> 4711 <td>police, cop, officer, man</td> 4712 </tr> 4713 <tr> 4714 <td>����♂️</td> 4715 <td>man police officer: medium-light skin tone</td> 4716 <td>police, cop, officer, medium-light skin tone, 4717 man</td> 4718 </tr> 4719 <tr> 4720 <td>��♀️</td> 4721 <td>woman police officer</td> 4722 <td>police, woman, cop, officer</td> 4723 </tr> 4724 <tr> 4725 <td>����♀️</td> 4726 <td>woman police officer: dark skin tone</td> 4727 <td>police, woman, cop, officer, dark skin tone</td> 4728 </tr> 4729 <tr> 4730 <td>��</td> 4731 <td>person biking</td> 4732 <td>cyclist, bicycle, biking</td> 4733 </tr> 4734 <tr> 4735 <td>����</td> 4736 <td>person biking: dark skin tone</td> 4737 <td>cyclist, bicycle, biking, dark skin tone</td> 4738 </tr> 4739 <tr> 4740 <td>��♂️</td> 4741 <td>man biking</td> 4742 <td>cyclist, bicycle, biking, man</td> 4743 </tr> 4744 <tr> 4745 <td>����♂️</td> 4746 <td>man biking: dark skin tone</td> 4747 <td>cyclist, bicycle, biking, dark skin tone, man</td> 4748 </tr> 4749 <tr> 4750 <td>��♀️</td> 4751 <td>woman biking</td> 4752 <td>cyclist, woman, bicycle, biking</td> 4753 </tr> 4754 <tr> 4755 <td>����♀️</td> 4756 <td>woman biking: dark skin tone</td> 4757 <td>cyclist, woman, bicycle, biking, dark skin tone</td> 4758 </tr> 4759 </tbody> 4760 </table> 4761 <p>For more information, see <a href= 4762 'https://unicode.org/reports/tr51'>Unicode Emoji</a>.</p> 4763 <h3>14.2 <a name="Character_Labels" href="#Character_Labels" 4764 id="Character_Labels">Annotations Character Labels</a></h3> 4765 <p class="dtd"><!ELEMENT characterLabels ( alias | ( 4766 characterLabelPattern*, characterLabel*, special* ) ) ></p> 4767 <p class="dtd"><!ELEMENT characterLabelPattern ( #PCDATA ) 4768 ></p> 4769 <p class="dtd"><!ATTLIST characterLabelPattern type NMTOKEN 4770 #REQUIRED ></p> 4771 <p class="dtd"><!ATTLIST characterLabelPattern count (0 | 1 4772 | zero | one | two | few | many | other) #IMPLIED > <!-- 4773 count only used for certain patterns" --></p> 4774 <p class="dtd"><!ELEMENT characterLabel ( #PCDATA ) ></p> 4775 <p class="dtd"><!ATTLIST characterLabel type NMTOKEN 4776 #REQUIRED ></p> 4777 <p>The character labels can be used for categories or groups of 4778 characters in a character picker or keyboard palette. They have 4779 the above structure. Items with special meanings are explained 4780 below. Many of the categories are based on terms used in 4781 Unicode. Consult the <a href= 4782 'https://www.unicode.org/glossary/'>Unicode Glossary</a> where 4783 the meaning is not clear.</p> 4784 <p>The following are special patterns used in composing 4785 labels.</p> 4786 <table> 4787 <caption> 4788 characterLabelPattern 4789 </caption> 4790 <tr> 4791 <th>Type</th> 4792 <th>English</th> 4793 <th>Description of the group specified.</th> 4794 </tr> 4795 <tr> 4796 <th>all</th> 4797 <td>{0} — all</td> 4798 <td>Used where the title {0} is just a subset. For example, 4799 {0} might be "Latin", and contain the most common Latin 4800 characters. Then "Latin — all" would be all of them.</td> 4801 </tr> 4802 <tr> 4803 <th>category-list</th> 4804 <td>{0}: {1}</td> 4805 <td>Use for a name, where {0} is the main item like 4806 "Family", and {1} is a list of one or more components or 4807 subcategories. The list is formatted using a list 4808 pattern.</td> 4809 </tr> 4810 <tr> 4811 <th>compatibility</th> 4812 <td>{0} — compatibility</td> 4813 <td>For grouping Unicode compatibility characters 4814 separately, such as "Arabic — compatibility".</td> 4815 </tr> 4816 <tr> 4817 <th>enclosed</th> 4818 <td>{0} — enclosed</td> 4819 <td>For indicating enclosed forms, such as "digits — 4820 enclosed"</td> 4821 </tr> 4822 <tr> 4823 <th>extended</th> 4824 <td>{0} — extended</td> 4825 <td>For indicating a group of "extended" characters 4826 (special use, technical, etc.)</td> 4827 </tr> 4828 <tr> 4829 <th>historic</th> 4830 <td>{0} — historic</td> 4831 <td>For indicating a group of "historic" characters (no 4832 longer in common use).</td> 4833 </tr> 4834 <tr> 4835 <th>miscellaneous</th> 4836 <td>{0} — miscellaneous</td> 4837 <td>For indicating a group of "miscellaneous" characters 4838 (typically that don't fall into a broader class).</td> 4839 </tr> 4840 <tr> 4841 <th>other</th> 4842 <td>{0} — other</td> 4843 <td>Used where the title {0} is just a subset. For example, 4844 {0} might be "Latin", and contain the most common Latin 4845 characters. Then "Latin — other" would be the rest of 4846 them.</td> 4847 </tr> 4848 <tr> 4849 <th>scripts</th> 4850 <td>scripts — {0}</td> 4851 <td>For indicating a group of "scripts" characters matching 4852 {0}. The value for {0} may be a geographic indicator, like 4853 "Africa" (although there are specific combinations listed 4854 below), or some other designation, like "other" (from 4855 below).</td> 4856 </tr> 4857 <tr> 4858 <th>strokes</th> 4859 <td>{0} strokes</td> 4860 <td>Used as an index title for CJK characters. It takes a 4861 "count" value, which allows the right plural form to be 4862 specified for the language.</td> 4863 </tr> 4864 <tr> 4865 <th>subscript</th> 4866 <td>subscript {0}</td> 4867 <td>For indicating subscript forms, such as "subscript digits".</td> 4868 </tr> 4869 <tr> 4870 <th>superscript</th> 4871 <td>superscript {0}</td> 4872 <td>For indicating superscript forms, such as "superscript digits".</td> 4873 </tr> 4874 </table> 4875 <p>The following are character labels. Where the meaning of the 4876 label is fairly clear (like "animal") or is in the Unicode 4877 glossary, it is omitted.</p> 4878 <table> 4879 <caption> 4880 characterLabel 4881 </caption> 4882 <tr> 4883 <th>activities</th> 4884 <td>activity</td> 4885 <td>Human activities, such as running.</td> 4886 </tr> 4887 <tr> 4888 <th>african_scripts</th> 4889 <td>African script</td> 4890 <td>Scripts associated with the continent of Africa.</td> 4891 </tr> 4892 <tr> 4893 <th>american_scripts</th> 4894 <td>American script</td> 4895 <td>Scripts associated with the continents of North and 4896 South America.</td> 4897 </tr> 4898 <tr> 4899 <th>animals_nature</th> 4900 <td>animal or nature</td> 4901 <td>A broad category uses for</td> 4902 </tr> 4903 <tr> 4904 <th>arrows</th> 4905 <td>arrow</td> 4906 <td>Arrow symbols</td> 4907 </tr> 4908 <tr> 4909 <th>body</th> 4910 <td>body</td> 4911 <td>Symbols for body parts, such as an arm.</td> 4912 </tr> 4913 <tr> 4914 <th>box_drawing</th> 4915 <td>box drawing</td> 4916 <td>Unicode box-drawing characters (geometric shapes)</td> 4917 </tr> 4918 <tr> 4919 <th>bullets_stars</th> 4920 <td>bullet or star</td> 4921 <td>Unicode bullets (such as • or ‣ or ⁍) or stars 4922 (★✩✪✵...)</td> 4923 </tr> 4924 <tr> 4925 <th>consonantal_jamo</th> 4926 <td>consonantal jamo</td> 4927 <td>Korean Jamo consonants.</td> 4928 </tr> 4929 <tr> 4930 <th>currency_symbols</th> 4931 <td>currency symbol</td> 4932 <td>Symbols such as $, ¥, £</td> 4933 </tr> 4934 <tr> 4935 <th>dash_connector</th> 4936 <td>dash or connector</td> 4937 <td>Characters like _ or ⁓</td> 4938 </tr> 4939 <tr> 4940 <th>dingbats</th> 4941 <td>dingbat</td> 4942 <td>Font dingbat characters, such as ❿ or ♜.</td> 4943 </tr> 4944 <tr> 4945 <th>downwards_upwards_arrows</th> 4946 <td>downwards upwards arrow</td> 4947 <td>⇕,...</td> 4948 </tr> 4949 <tr> 4950 <th>female</th> 4951 <td>female</td> 4952 <td>Indicates that a character is female or feminine in 4953 appearance.</td> 4954 </tr> 4955 <tr> 4956 <th>format</th> 4957 <td>format</td> 4958 <td>A Unicode format character.</td> 4959 </tr> 4960 <tr> 4961 <th>format_whitespace</th> 4962 <td>format & whitespace</td> 4963 <td>A Unicode format character or whitespace.</td> 4964 </tr> 4965 <tr> 4966 <th>full_width_form_variant</th> 4967 <td>full-width variant</td> 4968 <td>Full width variant, such as a wide A.</td> 4969 </tr> 4970 <tr> 4971 <th>half_width_form_variant</th> 4972 <td>half-width variant</td> 4973 <td>Narrow width variant, such as a half-width katakana 4974 character.</td> 4975 </tr> 4976 <tr> 4977 <th>han_characters</th> 4978 <td>Han character</td> 4979 <td>Han (aka CJK: Chinese, Japanese, or Korean) 4980 ideograph</td> 4981 </tr> 4982 <tr> 4983 <th>han_radicals</th> 4984 <td>Han radical</td> 4985 <td>Radical (component) used in Han characters.</td> 4986 </tr> 4987 <tr> 4988 <th>hanja</th> 4989 <td>hanja</td> 4990 <td>Korean name for Han character.</td> 4991 </tr> 4992 <tr> 4993 <th>hanzi_simplified</th> 4994 <td>Hanzi (simplified)</td> 4995 <td>Simplified Chinese ideograph</td> 4996 </tr> 4997 <tr> 4998 <th>hanzi_traditional</th> 4999 <td>Hanzi (traditional)</td> 5000 <td>Traditional Chinese ideograph</td> 5001 </tr> 5002 <tr> 5003 <th>historic_scripts</th> 5004 <td>historic script</td> 5005 <td>Script no longer in common modern usage, such as Runes 5006 or Hieroglyphs.</td> 5007 </tr> 5008 <tr> 5009 <th>ideographic_desc_characters</th> 5010 <td>ideographic desc. character</td> 5011 <td>Special Unicode characters (see the glossary).</td> 5012 </tr> 5013 <tr> 5014 <th>kanji</th> 5015 <td>kanji</td> 5016 <td>Japanese Han ideograph</td> 5017 </tr> 5018 <tr> 5019 <th>keycap</th> 5020 <td>keycap</td> 5021 <td>A key on a computer keyboard or phone. For example, the 5022 "3" key on a phone or laptop would be "keycap: 3"</td> 5023 </tr> 5024 <tr> 5025 <th>limited_use</th> 5026 <td>limited-use</td> 5027 <td>Not in common modern use.</td> 5028 </tr> 5029 <tr> 5030 <th>male</th> 5031 <td>male</td> 5032 <td>Indicates that a character is male or masculine in 5033 appearance.</td> 5034 </tr> 5035 <tr> 5036 <th>modifier</th> 5037 <td>modifier</td> 5038 <td>A Unicode modifier letter or symbol.</td> 5039 </tr> 5040 <tr> 5041 <th>nonspacing</th> 5042 <td>nonspacing</td> 5043 <td>Uses for characters that occupy no width by themselves, 5044 such as the ¨ over the a in ä.</td> 5045 </tr> 5046 </table> 5047 <h3>14.3 <a name="Typographic_Names" href="#Typographic_Names" 5048 id="Typographic_Names">Typographic Names</a></h3> 5049 <p class='dtd'><!ELEMENT typographicNames ( alias | ( 5050 axisName*, styleName*, featureName*, special* ) ) ></p> 5051 <p class='dtd'><!ELEMENT axisName ( #PCDATA ) ><br> 5052 <!ATTLIST axisName type (ital | opsz | slnt | wdth | wght) 5053 #REQUIRED ><br> 5054 <!ATTLIST axisName alt NMTOKENS #IMPLIED ></p> 5055 <p class='dtd'><!ELEMENT styleName ( #PCDATA ) ><br> 5056 <!ATTLIST styleName type (ital | opsz | slnt | wdth | wght) 5057 #REQUIRED ><br> 5058 <!ATTLIST styleName subtype NMTOKEN #REQUIRED ><br> 5059 <!ATTLIST styleName alt NMTOKENS #IMPLIED ></p> 5060 <p class='dtd'><!ELEMENT featureName ( #PCDATA ) ><br> 5061 <!ATTLIST featureName type (afrc | cpsp | dlig | frac | lnum 5062 | onum | ordn | pnum | smcp | tnum | zero) #REQUIRED ><br> 5063 <!ATTLIST featureName alt NMTOKENS #IMPLIED ></p> 5064 <p>The typographic names provide for names of font features for 5065 use in a UI. This is useful for apps that show the name of font 5066 styles and design axes according to the user’s languages. It 5067 would also be useful for system-level libraries.</p> 5068 <p>The identifers (types) use the tags from the OpenType 5069 Feature Tag Registry. Given their large number, only the names 5070 of frequently-used OpenType feature names are available CLDR. 5071 (Many features are not user-visible settings, but instead serve 5072 as a data channel for sofware to pass information to the font). 5073 The example below shows an approach for using the CLDR data. Of 5074 course, applications are free to implement their own algorithms 5075 depending on their specific needs.</p> 5076 <p>To find a localized subfamily name such as “Extraleicht 5077 Schmal” for a font called “Extralight Condensed”, a system or 5078 application library might do the following:</p> 5079 <ol> 5080 <li> 5081 <p>Determine the set of languages in which the subfamily 5082 name can potentially be returned.This is the union of the 5083 languages for which the font contains ‘name’ table entries 5084 with ID 2 or 17, plus the languages for which CLDR supplies 5085 typographic names.</p> 5086 </li> 5087 <li> 5088 <p>Use a language matching algorithm such as in ICU to find 5089 the best available language given the user preferences. The 5090 resulting subfamily name will be localized to this 5091 language.</p> 5092 </li> 5093 <li> 5094 <p>If the font’s ‘name’ table contains a typographic 5095 subfamily name (ID17) in this language and all font 5096 variation axes are set to their defaults, return this 5097 name.</p> 5098 </li> 5099 <li> 5100 <p>If the font’s ‘name’ table contains a font subfamilyname 5101 (‘name’ID2) in this language and all font variation axes 5102 are set to their defaults, return this name.</p> 5103 </li> 5104 <li> 5105 <p>If the font has a style attributes (STAT) table, lookup 5106 the design axis tags and their ordering. If the font has no 5107 STAT table, assume [Width, Weight, Slant] as axis ordering, 5108 and infer the font’s style atributes from other available 5109 data in the font (eg. the OS/2 table).</p> 5110 </li> 5111 <li>For each design axis, find a localized style name for its 5112 value. 5113 <ol> 5114 <li>If the font’s style attributes point to a ‘name’ 5115 table entry that is available the result language, use 5116 this name.</li> 5117 <li>Otherwise, generate a fallback name from CLDR style 5118 Name data. 5119 <ol> 5120 <li>The type key is the OpenType axis tag ( ‘wght’). 5121 The subtype and alt keys are taken from the entry in 5122 English CLDR where the string is equal to the English 5123 name in the font. For example, when the font uses a 5124 weight whose English style name is “Extralight”, this 5125 will lead to subtype = “200” and alt = “variant”. If 5126 there is no match, take the axis value (“200”) for 5127 subtype and the empty string for alt.</li> 5128 <li>Look up (type, subtype) in a data table derived 5129 from CLDR’s style names. If CLDR supplies multiple 5130 alternate names for this (type, subtype), use the one 5131 whose “alt” key is matching; otherwise, use the 5132 default alternate (which has no “alt” atribute in 5133 CLDR).</li></ol></li> 5134 </ol> 5135 </li> 5136 <li>Concatenate the strings, with a separator between 5137 them.</li> 5138 </ol> 5139 <h2>15 <a name="Grammatical_Features" href="#Grammatical_Features" 5140 id="Grammatical_Features">Grammatical Features</a></h2> 5141 5142 5143<p> 5144LDML supplies grammatical information that can be used to distinguish localized forms on a per-locale basis. The current data is part of an initial phase; the longer term plan is to add structure to permit localized forms based on these features, starting with measurement units such as the dative form in Serbian of “kilometer”. That will allow unit values to be inserted as placeholders into messages and adopt the right forms for grammatical agreement. 5145</p> 5146<p> 5147The current data includes the following: 5148</p><ul> 5149 5150<li>There are currently 3 grammatical features found in the <a href="https://github.com/unicode-org/cldr/blob/master/common/dtd/ldmlSupplemental.dtd#1229">DTD</a>: Gender, Case, Definiteness 5151<li>There are mappings from supported locales to grammatical features they exhibit in the file <a href="https://github.com/unicode-org/cldr/blob/master/common/supplemental/grammaticalFeatures.xml">grammaticalFeatures.xml</a>. Note that this is supplemental data, so the inheritance to the available locales needs to be done by the client.</li></ul> 5152 5153<p> 5154Note that the CLDR plural categories overlap some of these features, since some languages use case and other devices to change words based on the numeric values. 5155</p> 5156<h2>Features</h2> 5157 5158 5159<p class='dtd'> <!ELEMENT grammaticalData ( grammaticalFeatures*, grammaticalDerivations*) ><br> <!ELEMENT grammaticalFeatures ( grammaticalCase*, grammaticalGender*, grammaticalDefiniteness* ) ><br> 5160<!ATTLIST grammaticalFeatures targets NMTOKENS #REQUIRED ><br> 5161<!ATTLIST grammaticalFeatures locales NMTOKENS #REQUIRED > 5162</p> 5163<p> </p> 5164<p class='dtd'> 5165<!ELEMENT grammaticalCase EMPTY> 5166</p> 5167<p class='dtd'> 5168<!ATTLIST grammaticalCase values NMTOKENS #REQUIRED > 5169</p> 5170<p> </p> 5171<p class='dtd'> 5172<!ELEMENT grammaticalGender EMPTY> 5173</p> 5174<p class='dtd'> 5175<!ATTLIST grammaticalGender values NMTOKENS #REQUIRED > 5176</p> 5177<p> </p> 5178<p class='dtd'> 5179<!ELEMENT grammaticalDefiniteness EMPTY> 5180</p> 5181<p class='dtd'> 5182<!ATTLIST grammaticalDefiniteness values NMTOKENS #REQUIRED > 5183</p> 5184<h3>15.1<a name="Gender" href="#Gender" >Gender</a></h3> 5185<p> 5186 Feature that classifies nouns in classes. This is grammatical gender, which may be assigned on the basis of sex in some languages, but may be completely separate in others. Also used to tag elements in CLDR that should agree with a particular gender of an associated noun. (adapted from: <a href="http://linguistics-ontology.org/gold/2010/GenderProperty">linguistics-ontology.org/gold/2010/GenderProperty</a>) 5187</p> 5188<h4>Example</h4> 5189 5190 5191 5192 5193 5194<pre class="prettyprint"><grammaticalFeatures targets="nominal" locales="es fr it pt"> 5195 <grammaticalGender values="masculine feminine"/> 5196</pre> 5197 5198 5199<h4>Values</h4> 5200 5201 5202 5203<table> 5204 <tr> 5205 <td>animate 5206 </td> 5207 <td>In an animate/inanimate gender system, gender that denotes human or animate entities 5208 </td> 5209 <td>description adapted from: <a href="https://en.wikipedia.org/wiki/Grammatical_gender">wikipedia.org/wiki/Grammatical_gender</a> <a href="http://linguistics-ontology.org/gold/2010/AnimateGender">linguistics-ontology.org/gold/2010/AnimateGender</a> 5210 </td> 5211 </tr> 5212 <tr> 5213 <td>inanimate 5214 </td> 5215 <td>In an animate/inanimate gender system, gender that denotes object or inanimate entities 5216 </td> 5217 <td>adapted from: <a href="https://en.wikipedia.org/wiki/Grammatical_gender">wikipedia.org/wiki/Grammatical_gender</a> <a href="http://linguistics-ontology.org/gold/2010/InanimateGender">linguistics-ontology.org/gold/2010/InanimateGender</a> 5218 </td> 5219 </tr> 5220 <tr> 5221 <td>personal 5222 </td> 5223 <td>In an animate/inanimate gender system in some languages, gender that specifies the masculine gender of animate entities 5224 </td> 5225 <td>adapted from: <a href="https://en.wikipedia.org/wiki/Grammatical_gender">wikipedia.org/wiki/Grammatical_gender</a> <a href="http://linguistics-ontology.org/gold/2010/HumanGender">linguistics-ontology.org/gold/2010/HumanGender</a> 5226 </td> 5227 </tr> 5228 <tr> 5229 <td>common 5230 </td> 5231 <td>In a common–neuter gender system, gender that denotes human entities. 5232 </td> 5233 <td>adapted from: <a href="https://en.wikipedia.org/wiki/Grammatical_gender">wikipedia.org/wiki/Grammatical_gender</a> 5234 </td> 5235 </tr> 5236 <tr> 5237 <td>feminine 5238 </td> 5239 <td>In a masculine/feminine or in a masculine/feminine/neuter gender system, gender that denotes specifically female persons (or animals) or that is assigned arbitrarily to object. 5240 </td> 5241 <td>adapted from: wikipedia.org/wiki/Grammatical_gender <a href="http://linguistics-ontology.org/gold/2010/FeminineGender">linguistics-ontology.org/gold/2010/FeminineGender</a> 5242 </td> 5243 </tr> 5244 <tr> 5245 <td>masculine 5246 </td> 5247 <td>In a masculine/feminine or in a masculine/feminine/neuter gender system, gender that denotes specifically male persons (or animals) or that is assigned arbitrarily to object. 5248 </td> 5249 <td>adapted from: <a href="https://en.wikipedia.org/wiki/Grammatical_gender">wikipedia.org/wiki/Grammatical_gender</a> <a href="http://linguistics-ontology.org/gold/2010/MasculineGender">linguistics-ontology.org/gold/2010/MasculineGender</a> 5250 </td> 5251 </tr> 5252 <tr> 5253 <td>neuter 5254 </td> 5255 <td>In a masculine/feminine/neuter or common-neuter gender system, gender that generally denotes an object. 5256 </td> 5257 <td>adapted from: <a href="https://en.wikipedia.org/wiki/Grammatical_gender">wikipedia.org/wiki/Grammatical_gender</a> <a href="http://linguistics-ontology.org/gold/2010/NeuterGender">linguistics-ontology.org/gold/2010/NeuterGender</a> 5258 </td> 5259 </tr> 5260</table><br> 5261 <h3>15.2<a name="Case" href="#Case">Case</a></h3> 5262 5263<h3>Case</h3> 5264 5265 5266<p> 5267Feature that encodes the syntactic (and sometimes semantic) relationship of a noun with the other constituents of the sentence. (adapted from <a href="http://linguistics-ontology.org/gold/2010/CaseProperty">linguistics-ontology.org/gold/2010/CaseProperty</a>) 5268</p> 5269<h4>Example</h4> 5270 5271 5272 5273 5274 5275<pre class="prettyprint"><grammaticalFeatures targets="nominal" locales="es fr it pt"> 5276 <grammaticalGender values="masculine feminine"/> 5277</pre> 5278 5279 5280<h4>Values</h4> 5281 5282 5283 5284<table> 5285 <tr> 5286 <td><strong>Value</strong> 5287 </td> 5288 <td><strong>Definition</strong> 5289 </td> 5290 <td><strong>References</strong> 5291 </td> 5292 </tr> 5293 <tr> 5294 <td>ablative 5295 </td> 5296 <td>Ablative case expresses that the referent of the noun it marks is the location from which another referent is moving. It has the meaning 'from'. 5297 </td> 5298 <td><a href="https://purl.org/olia/olia.owl#AblativeCase">purl.org/olia/olia.owl#AblativeCase</a> <a href="http://linguistics-ontology.org/gold/2010/AblativeCase">linguistics-ontology.org/gold/2010/AblativeCase</a> 5299 </td> 5300 </tr> 5301 <tr> 5302 <td>accusative 5303 </td> 5304 <td>Accusative case marks certain syntactic functions, usually direct objects. 5305 </td> 5306 <td><a href="https://purl.org/olia/olia.owl#Accusative">purl.org/olia/olia.owl#Accusative</a> <a href="http://linguistics-ontology.org/gold/2010/AccusativeCase">linguistics-ontology.org/gold/2010/AccusativeCase</a> 5307 </td> 5308 </tr> 5309 <tr> 5310 <td>comitative 5311 </td> 5312 <td>Comitative Case expresses accompaniment. It carries the meaning 'with' or 'accompanied by'. 5313 </td> 5314 <td><a href="https://purl.org/olia/olia.owl#ComitativeCase">purl.org/olia/olia.owl#ComitativeCase</a> <a href="http://linguistics-ontology.org/gold/2010/ComitativeCase">linguistics-ontology.org/gold/2010/ComitativeCase</a> 5315 </td> 5316 </tr> 5317 <tr> 5318 <td>dative 5319 </td> 5320 <td>Dative case marks indirect objects (for languages in which they are held to exist), or nouns having the role of a recipient (as of things given), a beneficiary of an action, or a possessor of an item. 5321 </td> 5322 <td><a href="https://purl.org/olia/olia.owl#DativeCase">purl.org/olia/olia.owl#DativeCase</a> <a href="http://linguistics-ontology.org/gold/2010/DativeCase">linguistics-ontology.org/gold/2010/DativeCase</a> 5323 </td> 5324 </tr> 5325 <tr> 5326 <td>ergative 5327 </td> 5328 <td>In ergative-absolutive languages, the ergative case identifies the subject of a transitive verb. 5329 </td> 5330 <td><a href="https://purl.org/olia/olia.owl#ErgativeCase">purl.org/olia/olia.owl#ErgativeCase</a> <a href="http://linguistics-ontology.org/gold/2010/ErgativeCase">linguistics-ontology.org/gold/2010/ErgativeCase</a> 5331 </td> 5332 </tr> 5333 <tr> 5334 <td>genitive 5335 </td> 5336 <td>Genitive case signals that the referent of the marked noun is the possessor of the referent of another noun, e.g. "the man's foot". In some languages, genitive case may express an associative relation between the marked noun and another noun. 5337 </td> 5338 <td><a href="https://purl.org/olia/olia.owl#GenitiveCase">purl.org/olia/olia.owl#GenitiveCase</a> <a href="http://linguistics-ontology.org/gold/2010/GenitiveCase">linguistics-ontology.org/gold/2010/GenitiveCase</a> 5339 </td> 5340 </tr> 5341 <tr> 5342 <td>instrumental 5343 </td> 5344 <td>InstrumentalCase indicates that the referent of the noun it marks is the means of the accomplishment of the action expressed by the clause. 5345 </td> 5346 <td><a href="https://purl.org/olia/olia.owl#InstrumentalCase">purl.org/olia/olia.owl#InstrumentalCase</a> <a href="http://linguistics-ontology.org/gold/2010/InstrumentalCase">linguistics-ontology.org/gold/2010/InstrumentalCase</a> 5347 </td> 5348 </tr> 5349 <tr> 5350 <td>locative 5351 </td> 5352 <td>Case that indicates a final location of action or a time of the action. 5353 </td> 5354 <td><a href="https://purl.org/olia/olia.owl#LocativeCase">purl.org/olia/olia.owl#LocativeCase</a> <a href="http://linguistics-ontology.org/gold/2010/LocativeCase">linguistics-ontology.org/gold/2010/LocativeCase</a> 5355 </td> 5356 </tr> 5357 <tr> 5358 <td>locativecopulative 5359 </td> 5360 <td>Copulative Case marker that indicates a location. 5361 </td> 5362 <td>TBD Add reference, example 5363 </td> 5364 </tr> 5365 <tr> 5366 <td>nominative 5367 </td> 5368 <td>In nominative-accusative languages, nominative case marks clausal subjects and is applied to nouns in isolation 5369 </td> 5370 <td><a href="https://purl.org/olia/olia.owl#Nominative">purl.org/olia/olia.owl#Nominative</a> <a href="http://linguistics-ontology.org/gold/2010/NominativeCase">linguistics-ontology.org/gold/2010/NominativeCase</a> 5371 </td> 5372 </tr> 5373 <tr> 5374 <td>oblique 5375 </td> 5376 <td>Case that is used when a noun is the object of a verb or a proposition, except for nominative and vocative case. 5377 </td> 5378 <td><a href="https://purl.org/olia/olia.owl#ObliqueCase">purl.org/olia/olia.owl#ObliqueCase</a> 5379 </td> 5380 </tr> 5381 <tr> 5382 <td>partitive 5383 </td> 5384 <td>The partitive case is a grammatical case which denotes "partialness", "without result", or "without specific identity". 5385 </td> 5386 <td><a href="https://purl.org/olia/olia.owl#PartitiveCase">purl.org/olia/olia.owl#PartitiveCase</a> <a href="http://linguistics-ontology.org/gold/2010/PartitiveCase">linguistics-ontology.org/gold/2010/PartitiveCase</a> 5387 </td> 5388 </tr> 5389 <tr> 5390 <td>prepositional 5391 </td> 5392 <td>Prepositional case refers to case marking that only occurs in combination with prepositions. 5393 </td> 5394 <td><a href="https://purl.org/olia/olia.owl#PrepositionalCase">purl.org/olia/olia.owl#PrepositionalCase</a> 5395 </td> 5396 </tr> 5397 <tr> 5398 <td>sociative 5399 </td> 5400 <td>Case related to the person in whose company the action is carried out, or to any belongings of people which take part in the action. 5401 </td> 5402 <td><a href="https://purl.org/olia/olia.owl#SociativeCase">purl.org/olia/olia.owl#SociativeCase</a> 5403 </td> 5404 </tr> 5405 <tr> 5406 <td>vocative 5407 </td> 5408 <td>Vocative case marks a noun whose referent is being addressed. 5409 </td> 5410 <td><a href="https://purl.org/olia/olia.owl#VocativeCase">purl.org/olia/olia.owl#VocativeCase</a> <a href="http://linguistics-ontology.org/gold/2010/VocativeCase">linguistics-ontology.org/gold/2010/VocativeCase</a> 5411 </td> 5412 </tr> 5413</table> 5414 5415 5416<h3>Definiteness</h3> 5417 5418 5419<p> 5420Feature that encodes the fact that a noun has been already mentioned, or is familiar in the discourse. (adapted from <a href="https://glossary.sil.org/term/definiteness">https://glossary.sil.org/term/definiteness</a> ) 5421</p> 5422<h4>Values</h4> 5423 5424 5425 5426<table class='simple'> 5427 <tr> 5428 <td>definite 5429 </td> 5430 <td>Value referring to the capacity of identification of an entity. 5431 </td> 5432 <td><a href="https://purl.org/olia/olia.owl#Definite">purl.org/olia/olia.owl#Definite</a> 5433 </td> 5434 </tr> 5435 <tr> 5436 <td>indefinite 5437 </td> 5438 <td>An entity is specified as indefinite when it refers to a non-particularized individual of the species denoted by the noun. 5439 </td> 5440 <td><a href="https://purl.org/olia/olia.owl#Indefinite">purl.org/olia/olia.owl#Indefinite</a> 5441 </td> 5442 </tr> 5443 <tr> 5444 <td>construct 5445 </td> 5446 <td>State of the first noun in a genitive phrase of a possessed noun followed by a possessor noun. 5447 </td> 5448 <td>no direct linked, but explained under: <a href="https://purl.org/olia/olia-top.owl#DefinitenessFeature">purl.org/olia/olia-top.owl#DefinitenessFeature</a> 5449 </td> 5450 </tr> 5451 <tr> 5452 <td>unspecified 5453 </td> 5454 <td>Noun without any definiteness marking in some specific construction (specific to Danish) 5455 </td> 5456 <td> </td> 5457 </tr> 5458</table> 5459 5460<h2>16 <a name="Grammatical_Derivations" href="#Grammatical_Derivations">Grammatical Derivations</a></h2> 5461<pre class='dtd'><!ELEMENT grammaticalData ( grammaticalFeatures*, grammaticalDerivations*) > 5462<!ELEMENT grammaticalDerivations (deriveCompound*, deriveComponent*) > 5463<!ATTLIST grammaticalDerivations locales NMTOKENS #REQUIRED > 5464 5465<!ELEMENT deriveCompound EMPTY > 5466<!ATTLIST deriveCompound feature NMTOKENS #REQUIRED > 5467<!ATTLIST deriveCompound structure NMTOKENS #REQUIRED > 5468<!ATTLIST deriveCompound value NMTOKEN #REQUIRED > 5469 5470<!ATTLIST deriveComponent feature NMTOKENS #REQUIRED > 5471<!ATTLIST deriveComponent structure NMTOKENS #REQUIRED > 5472<!ATTLIST deriveComponent value0 NMTOKEN #REQUIRED > 5473<!ATTLIST deriveComponent value1 NMTOKEN #REQUIRED > 5474</pre> 5475 <p>The grammatical derivation data contains information about the case, gender, and plural categories of compound units. This is supplemental data, so the inheritance by locale needs to be done by the client.</p> 5476 <p><em>Note: In CLDR v38, the data for two locales is provided so that implemenations can ready their code for when more locale data is available. In subsequent releases structure may be further extended as more locales are added, to deal with additional locale requirements.</em></p> 5477 <p>A compound unit can use 4 mechanisms, illustrated here in formatted strings:</p> 5478 <ul> 5479 <li><strong>Prefix</strong>: 1 <strong>kilo</strong>gram</li> 5480 <li><strong>Power</strong>: 3 <strong>square</strong> kilometers</li> 5481 <li><strong>Per</strong>: 3 kilograms <strong>per</strong> meter 5482 <ul> 5483 <li>An edge case is where there is no numerator, such as “1 per-second”</li> 5484 </ul> 5485 </li> 5486 <li><strong>Times</strong>: 3 kilowatt<strong>-</strong>hours</li> 5487 </ul> 5488 <p>For the purposes of grammatical derivation (and name construction), a compound unit ID can be represented as a tree structure where the leaves are the atomic units, and the higher level node are one of the above. Here is an extreme example of that: <em>kilogram-square-kilometer-ampere-candela-per-square-second-mole</em></p> 5489 <table class='simple' style=" margin-left: auto;margin-right: auto;"> 5490 <tr> 5491 <td colspan="6" style="text-align:center"><strong>per</strong></td> 5492 </tr> 5493 <tr> 5494 <td colspan="4" style="text-align:center"><strong>times</strong></td> 5495 <td colspan="2" style="text-align:center"><strong>times</strong></td> 5496 </tr> 5497 <tr> 5498 <td style="text-align:center"><strong>kilo</strong></td> 5499 <td style="text-align:center"><strong>square</strong></td> 5500 <td style="text-align:center">ampere</td> 5501 <td style="text-align:center">candela</td> 5502 <td style="text-align:center"><strong>square</strong></td> 5503 <td style="text-align:center">mole</td> 5504 </tr> 5505 <tr> 5506 <td style="text-align:center">gram</td> 5507 <td style="text-align:center"><strong>kilo</strong></td> 5508 <td style="text-align:center">-</td> 5509 <td style="text-align:center">-</td> 5510 <td style="text-align:center">second</td> 5511 <td style="text-align:center"> </td> 5512 </tr> 5513 <tr> 5514 <td style="text-align:center">-</td> 5515 <td style="text-align:center">meter</td> 5516 <td style="text-align:center">-</td> 5517 <td style="text-align:center">-</td> 5518 <td colspan="2" style="text-align:center">-</td> 5519 </tr> 5520 </table> 5521 <p>Note that the prefix and power nodes are unary (exactly 1 child), the per pattern is unary or binary (1 or 2 children), 5522 and the times pattern is n-ary (where n > 1).</p> 5523 <p> </p> 5524 <p>Each section below is only applicable if the language has more than one value <em>for units</em>: 5525 for example, for plural categories the language has to have more than just "other". 5526 When that information is available for a language, it is found in 5527 <strong>Section 15 <a href="#Grammatical_Features" id="Grammatical_Features2">Grammatical Features</a></strong>.</p> 5528 <p dir="ltr">The gender derivation would be appropriate for an API call like <code>String genderValue = getGrammaticalGender(locale, "kilogram-meter-per-square-second")</code>. This can be used where the choice of word forms in the rest of a phrase can depend on the gender of the unit.</p> 5529 <p dir="ltr">On the other hand, the derivation of plural category and case are used in building up the name of a compound unit, where the desired plural category is available from the number to be formatted with the unit, and the case value is known from the position in a message. For example, the case could be accusative if the formatted unit is to be the direct object in a sentence or phrase. This could be expressed in an API call such as <code>String inflectedName = getUnitName(locale, "kilogram-meter-per-square-second", pluralCategory, caseValue)</code>. </p> 5530 <p dir="ltr">When deriving an inflected compound unit pattern, as the tree-stucture is processed by getting the appropriate localized patterns for the structural components and names for the atomic components. The computation of the plural category and the case of the subtrees can be computed from the <strong>deriveComponent</strong> data. The <strong>times</strong> data is treated as binary, and applied from left to right: with the example from above, the plural categories for the components of <em>kilogram-square-kilometer-ampere-candela</em> are computed by applying </p> 5531 <p align="center"><strong>times</strong>(<em>kilogram, <strong>times</strong>(square-kilometer, <strong>times</strong>(ampere, candela)))</em></p> 5532 <p dir="ltr">For a description of how to use these fields to construct a localized name, see <strong>Section 6.4 <a href="#compound-units">Compound Units</a></strong>.</p> 5533 <p dir="ltr"> </p> 5534 <h3 dir="ltr">16.1 <a name="gender_compound_units" href="#gender_compound_units">Deriving the Gender of Compound Units</a></h3> 5535 <p dir="ltr">The <strong>deriveCompound[@feature="gender"]</strong> data provides information for how to derive the gender of the whole compound from the gender of its atomic units and structure. The attributeValues of value are: 0 (=gender of the first element), 1 (=gender of second element), or one of the valid gender values for the language: </p> 5536 <p dir="ltr">Example:</p> 5537 <pre><deriveCompound feature="gender" structure="per" value="0"/> <!-- gender(gram-per-meter) ← gender(gram) --> <br><deriveCompound feature="gender" structure="times" value="1"/> <!-- gender(newton-meter) ← gender(meter) --> <br><deriveCompound feature="gender" structure="power" value="0"/> <!-- gender(square-meter) ← gender(meter) --> <br><deriveCompound feature="gender" structure="prefix" value="0"/> <!-- gender(kilometer) ← gender(meter)--> </pre> 5538 <p>For example, for gram-per-meter, the first line above means:</p> 5539 <ul> 5540 <li dir="ltr"> 5541 <p>The gender of the compound is the gender of the first component of the 'per', that is, of the "gram". So if gram is feminine in that language, the gender of the compound is feminine. </p> 5542 </li> 5543 </ul> 5544 <h3>16.2 <a name="plural_compound_units" href="#plural_compound_units">Deriving the Plural Category of Unit Components</a></h3> 5545 <p>The <strong>deriveComponent[@feature="plural"]</strong> data provides information for how to derive the plural category for each of the atomic units, from the plural category of the whole compound and the structure of the compound. The attributeValues of value0 and value1 are: "compound" (=the pluralCategory of the compound), or one of the valid plural category values for the language.</p> 5546 <p>Example:</p> 5547 <pre> 5548<deriveComponent feature="plural" structure="per" value0="compound" value1="one"/> <!-- compound(gram-per-meter) ⇒ compound(gram) “per" singular(meter) --> 5549<deriveComponent feature="plural" structure="times" value0="one" value1="compound"/> <!-- compound(newton-meter) ⇒ singular(newton) “-" compound(meter) --> 5550<deriveComponent feature="plural" structure="power" value0="one" value1="compound"/> <!-- compound(square-meter) ⇒ singular(square) compound(meter) --> 5551<deriveComponent feature="plural" structure="prefix" value0="one" value1="compound"/> <!-- compound(kilometer) ⇒ singular(kilo) compound(meter) --></pre> 5552 <p>For example, for gram-per-meter, the first line above means:</p> 5553 5554 <ul> 5555 <li dir="ltr"> 5556 <p>When the plural form of gram-per-meter is needed (rather than singular), then the gram part of the translation has to have a plural form like “grams”, while the meter part of the translation has to have a singular form like “metre”. This would be composed with the pattern for "per" (say "{0} pro {1}") to get "grams pro metre".</p> 5557 </li> 5558 </ul> 5559 <h3>16.3 <a name="case_compound_units" href="#case_compound_units">Deriving the Case of Unit Components</a></h3> 5560 <p>The <strong>deriveComponent[@feature="plural"]</strong> data provides information for how to derive the plural category for each of the atomic units, from the plural category of the whole compound and the structure of the compound.The attributeValues of value0 and value1 are: compound (=the grammatical case of the compound), or one of the valid grammatical case values for the language.</p> 5561 <p>Example:</p> 5562 <pre><deriveComponent feature="case" structure="per" value0="compound" value1="nominative"/> <!-- compound(gram-per-meter) ⇒ compound(gram) “per" accusative(meter) --> 5563<deriveComponent feature="case" structure="times" value0="nominative" value1="compound"/> <!-- compound(newton-meter) ⇒ nominative(newton) “-" compound(meter) --> 5564<deriveComponent feature="case" structure="power" value0="nominative" value1="compound"/> <!-- compound(square-meter) ⇒ nominative(square) compound(meter) --> 5565<deriveComponent feature="case" structure="prefix" value0="nominative" value1="compound"/><!--compound(kilometer) ⇒ nominative(kilo) compound(meter) --></pre> 5566 <p dir="ltr">For example, for gram-per-meter, the first line above means:</p> 5567 <ul> 5568 <li dir="ltr"> 5569 <p>When the accusative form of gram-per-meter is needed, then the gram part of the translation has take the accusative case (eg, “gramu”, in a language that marks the accusative case with 'u'), while the meter part of the translation has a nominative form like “metre”. This would be composed with the pattern for "per" (say "{0} pro {1}") to get "gramu pro metre".</p> 5570 </li> 5571 </ul> 5572 <p> </p> 5573 5574<hr> 5575 <p class="copyright">Copyright © 2001–2020 Unicode, Inc. All 5576 Rights Reserved. The Unicode Consortium makes no expressed or 5577 implied warranty of any kind, and assumes no liability for 5578 errors or omissions. No liability is assumed for incidental and 5579 consequential damages in connection with or arising out of the 5580 use of the information or programs contained or accompanying 5581 this technical report. The Unicode <a href= 5582 "https://unicode.org/copyright.html">Terms of Use</a> apply.</p> 5583 <p class="copyright">Unicode and the Unicode logo are 5584 trademarks of Unicode, Inc., and are registered in some 5585 jurisdictions.</p> 5586 </div> 5587</body> 5588</html> 5589
