i18npool/source/localedata/data/locale.dtd

   1 <?xml version="1.0" encoding="UTF-8"?>
   2
   3 <!-- ............................................................... -->
   4 <!-- Locale data specification DTD ................................. -->
   5 <!-- ............................................................... -->
   6
   7 <!--
   8     ===========================================================================
   9     ATTENTION! PLEASE! HEADS UP! IMPORTANT!
  10     ===========================================================================
  11
  12     Please validate your locale data contribution using a validating parser. A
  13     validating parser, for example, may be found at
  14     http://unicode.org/cldr/data/tools/java/org/unicode/cldr/util/XMLValidator.java
  15     Compile it into a class-jar and call it in the
  16     i18npool/source/localedata/data/ directory:
  17     java -cp <your_path>/XMLValidator.jar org.unicode.cldr.util.XMLValidator your_data.xml
  18
  19
  20     A second possiblity is:
  21
  22     - temporarily (!) change the DOCTYPE of your file to read (all on one line)
  23       <!DOCTYPE Locale SYSTEM "http://cgit.freedesktop.org/libreoffice/core/plain/i18npool/source/localedata/data/locale.dtd">
  24
  25     - upload it to the form available at http://www.validome.org/xml/
  26
  27     This will validate the file against the HEAD revision of locale.dtd
  28
  29
  30     Please test locale data files either in an enable-dbgutil build, which
  31     implements some checks and pops up assertion message boxes if the tests
  32     fail, or by setting the environment variable OOO_ENABLE_LOCALE_DATA_CHECKS
  33     to 'Y' or 'Yes' (or any other string starting with 'Y') or '1' before
  34     starting the application, which outputs the same messages to stderr and
  35     also works in a product build.
  36
  37     Then follow this procedure:
  38     1. Create a new spreadsheet document.
  39     2. On a cell use context menu -> Format Cells -> Numbers.
  40     3. Select the locale in the Language list box => MUST be assertion free.
  41     3.a. Assertions are only shown at the very first time a number formatter
  42          is created, respectively the first time a specific locale data is
  43          used.  To repeat steps 1.-3. you'd need to create another spreadsheet
  44          document.
  45
  46     ===========================================================================
  47     NOTE the FormatElement comments further down.
  48     ===========================================================================
  49
  50 -->
  51
  52 <!-- ............................................................... -->
  53 <!-- Entities for characters and symbols ........................... -->
  54
  55 <!ENTITY % UNOModule
  56         'unoid                  CDATA           #IMPLIED'>
  57
  58 <!ENTITY % MessageID
  59         'msgid                  CDATA           #REQUIRED'>
  60
  61 <!ENTITY % RefLocale
  62         'ref                    CDATA           #IMPLIED'>
  63 <!-- Where given, an element can be inherited from another locale, e.g.
  64      ref="en_US" -->
  65
  66 <!ENTITY % LIBModule
  67         'module                 CDATA           #IMPLIED'>
  68 <!-- The locale referred to for the sub categories, implementation detail. -->
  69
  70 <!ENTITY % replaceFrom  'replaceFrom    CDATA #IMPLIED'>
  71 <!ENTITY % replaceTo    'replaceTo      CDATA #IMPLIED'>
  72 <!-- See below for the LC_FORMAT element. -->
  73
  74 <!ELEMENT DefaultName  (#PCDATA)>
  75
  76 <!-- Locale is made of different sub categories -->
  77 <!ELEMENT Locale (LC_INFO, LC_CTYPE, LC_FORMAT, LC_FORMAT_1?, LC_COLLATION, LC_SEARCH, LC_INDEX, LC_CALENDAR, LC_CURRENCY,  LC_TRANSLITERATION, LC_MISC, LC_NumberingLevel, LC_OutLineNumberingLevel)>
  78 <!ATTLIST Locale versionDTD CDATA #FIXED "2.0.3">
  79 <!-- Version identifier to prevent mismatching data files being submitted
  80      because of older files being copied as templates. The version should be
  81      less than or equal to the OOo release, or the OOo release number plus some
  82      extension, like "2.0.enhanced", to be able to easily determine the
  83      corresponding OOo version. New versions of the DTD with new required
  84      elements SHOULD REALLY result in a new versionDTD here and
  85      LOCALE_VERSION_DTD in ../LocaleNode.cxx, and *.xml files MUST be adapted
  86      then, otherwise building the data or checking it with a validating parser
  87      will throw an error.
  88 -->
  89 <!ATTLIST Locale allowUpdateFromCLDR (yes|no) #REQUIRED>
  90 <!-- Whether some data elements may be (automatically) updated from the Common
  91      Locale Data Repository, see http://cldr.unicode.org/
  92      Note: This mechanism currently (2010-02-21) is outdated.
  93 -->
  94 <!ATTLIST Locale  version CDATA #REQUIRED>
  95 <!-- Valid number, may designate versioned data -->
  96
  97
  98 <!ELEMENT LC_INFO (Language, Country, Platform?, Variant?)>
  99
 100 <!ELEMENT Language (LangID, DefaultName) >
 101 <!ELEMENT LangID (#PCDATA) >
 102 <!-- LangID must be a valid two or three letter language identifier defined by
 103      ISO 639. Use ISO 639-1 two letter code where available, else ISO 639-2 or
 104      639-3 three letter code.
 105 -->
 106
 107 <!ELEMENT Country (CountryID, DefaultName) >
 108 <!ELEMENT CountryID (#PCDATA) >
 109 <!-- CountryID must be a valid two letter country identifier defined by ISO 3166. -->
 110
 111 <!ELEMENT Platform (PlatformID) >
 112 <!ELEMENT PlatformID (#PCDATA) >
 113 <!-- Unused, deprecated, can be generic|unix|win32|macos, best Platform element
 114      be absent.
 115 -->
 116
 117 <!ELEMENT Variant (#PCDATA) >
 118 <!-- Reserved for future use, currently unused. -->
 119
 120
 121 <!-- The LC_FORMAT element contains number format codes and may actually appear
 122      twice (with the second ocurrence named LC_FORMAT_1). One section is
 123      mandatory and MUST contain at least all required format codes
 124      formatindex="0" to formatindex="47", MUST NOT contain formatindex 48 and
 125      49, and MAY contain other format codes. See below ATTLIST FormatElement
 126      formatindex.
 127
 128      A second LC_FORMAT_1 section may follow containing other format codes. The
 129      difference between the two sections is that they are inherited
 130      independently if a locale uses the RefLocale mechanism (ref="..."). This
 131      may be used to not offer locale dependent format codes to other locales
 132      that otherwise inherit the format codes.
 133
 134 -->
 135 <!ELEMENT LC_FORMAT (DateAcceptancePattern*, FormatElement*) >
 136 <!-- All FormatElement elements must be given if the RefLocale mechanism is not used! -->
 137 <!ATTLIST LC_FORMAT %RefLocale;>
 138 <!ATTLIST LC_FORMAT %replaceFrom;>
 139 <!-- Define placeholder for currency code, usually "[CURRENCY]" -->
 140 <!ATTLIST LC_FORMAT %replaceTo;>
 141 <!-- Currency code to be used to replace the placeholder, e.g. "[$R-1C09]".
 142      Note: The brackets and the leading $ character are mandatory, the
 143      hyphen-minus separates the currency symbol from the hexagesimal MS-LCID,
 144      letters contained in the LCID have to be in upper case, leading zeros are
 145      to be omitted. LCIDs are defined in i18npool/inc/i18npool/lang.h
 146 -->
 147
 148 <!ELEMENT DateAcceptancePattern (#PCDATA)>
 149 <!-- Pattern that defines an input sequence match to be accepted as
 150      (abbreviated) date. For example, in en_US locale "M/D" accepts an input of
 151      11/23 as CurrentYear-November-23 whereas 11/23/ would not be a date. In
 152      de_DE locale "D.M." accepts an input of 23.11. as  CurrentYear-November-23
 153      whereas 23.11 would not be a date.
 154
 155      For each locale one pattern that matches a full date is automatically
 156      generated from FormatElement formatIndex="21" and does not need to be
 157      defined, for example "M/D/Y" or "D.M.Y", so only if you want abbreviations
 158      to be in effect add some patterns.
 159
 160      NOTE: use only single letter D,M,Y. Multiple patterns can be defined.
 161 -->
 162
 163 <!ELEMENT LC_FORMAT_1 (FormatElement*) >
 164 <!ATTLIST LC_FORMAT_1 %RefLocale;>
 165 <!ATTLIST LC_FORMAT_1 %replaceFrom;>
 166 <!ATTLIST LC_FORMAT_1 %replaceTo;>
 167
 168 <!ELEMENT FormatElement   ( FormatCode, DefaultName?)>
 169 <!ATTLIST FormatElement    %MessageID;>
 170 <!ATTLIST FormatElement    default  (true|false)        #REQUIRED >
 171 <!ATTLIST FormatElement    type     (short|medium|long) #REQUIRED >
 172 <!--
 173     There may be up to three groups (type="short", type="medium", type="long")
 174     for each usage category defined. Each group, if defined, needs excactly one
 175     default. The type roughly determines the display string length, for example
 176     short, medium, and long date formats.
 177
 178     The number formatter determines an ultimate default format of a specific
 179     usage category by looking at the medium, long, and short default formats,
 180     in that very particular order.
 181 -->
 182 <!ATTLIST FormatElement    usage    (FIXED_NUMBER|FRACTION_NUMBER|PERCENT_NUMBER|SCIENTIFIC_NUMBER|CURRENCY|DATE|TIME|DATE_TIME) #REQUIRED >
 183 <!ATTLIST FormatElement formatindex  CDATA #REQUIRED>
 184 <!--
 185     The following FormatElements must follow specific rules:
 186
 187     All:
 188         The format indices 0..49 are reserved and, for backwards compatibility,
 189         MUST be used as stated in
 190         offapi/com/sun/star/i18n/NumberFormatIndex.idl.
 191         Note that indices 10 ("# ?/?"), 11 ("# ??/??"), 48 (BOOLEAN) and 49 (@
 192         Text) are generated internally, as they aren't locale dependent, and
 193         must not be used in locale data XML files. All other formats have to be
 194         present.
 195
 196         Note also that "must be used as stated" does not mean that the format
 197         codes must be identical, of course the meaning of a format code should
 198         match, for example en_US MM/DD/YY matches de_DE DD.MM.YY
 199         Just imagine the same index being used with another locale shouldn't
 200         change the meaning of representation significantly.
 201
 202         You'll notice differences of non-matching format codes only if
 203         documents use the Default language in number formats and either are
 204         stored in old SO5 binary file format and loaded on another system where
 205         languages/locales aren't the same, or if the default locale is switched
 206         under menu Tools.Options.LanguageSettings.Languages.LocaleSetting
 207         dialog, which exchanges formats on the fly in the spreadsheet
 208         application. Please check it out! So far only very few locale data file
 209         we received got that right, especially not in date formats!
 210
 211         For easier comparison between locales in future please sort the
 212         FormatElements by their formatindex="..." value within a usage group.
 213         This isn't necessary to be technically correct and isn't done in many
 214         locales yet, but will certainly help.
 215
 216         If you want to define yet more formats than the preset and reserved
 217         0..49 range that's fine, only make sure those formatindex="..." values
 218         are each >=50 and all values are unique within one locale.
 219
 220     Of usage="FIXED_NUMBER":
 221         formatindex="0" MUST be the format containing the 'General' keyword.
 222         The keyword itself may be localized, it is good practice though to
 223         stick with a wording known from another spreadsheet application for
 224         better user experience. Like other format codes it may be prepended
 225         with a [NatNum1] modifier if values are to be displayed using native
 226         numbering if no specific format was applied. The format must have the
 227         default="true" and type="medium" attributes.
 228
 229     Of usage="DATE":
 230         formatindex="21" is used to edit already existing date data. In order
 231         to always edit the full century the long year YYYY code must be used.
 232         Furthermore, the format has to be of an editable type, of course, which
 233         means parseable. Therefore it should only contain DD, MM, YYYY and date
 234         separators, and the YMD default order is determined from the order
 235         encountered in this format.
 236         TODO: Future versions should make use of an edit="true" attribute
 237         instead of relying on this special requirement.
 238
 239         formatindex="33" must be ISO 8601 YYYY-MM-DD format code.
 240
 241     Of usage="DATE_TIME":
 242         formatindex="47" is used to edit already existing combined date/time
 243         data. The requirements are the same as for formatindex="21" above.
 244
 245     Of usage="TIME":
 246         * The FormatElement with default="true" type="medium" is used to edit
 247           times and thus must contain all HH and MM and SS codes, e.g. HH:MM:SS
 248         * Formatindices 43, 44, 45 are special in the sense that they are
 249           programmatically used to automatically display values that meet
 250           certain criteria:
 251           * Formatindex="43" contains the [HH] format code that displays hour
 252             values greater than or equal to 24 (as opposed to a simple HH that
 253             displays modulo 24).
 254           * Formatindex="44" uses no hour code but 100th seconds and the
 255             Time100SecSeparator.
 256           * Formatindex="45" uses both [HH] hour code and 100th seconds to be
 257             able to edit such time values without loosing information.
 258
 259     Of usage="CURRENCY":
 260         formatindices 12, 13, 14, 15, 17 with [$xxx-yyy] notation must use the
 261         xxx currency symbol that has the attribute
 262         usedInCompatibleFormatCodes="true".  The hexadecimal yyy LANGID must be
 263         properly set. It can be found in project util module tools file
 264         tools/inc/lang.hxx.  You may verify the proper use of the xxx currency
 265         symbol with the AWK script
 266         i18npool/source/localedata/data/currency-check.awk, it mustn't display
 267         any output. If it does, then there's something wrong.
 268 -->
 269 <!ELEMENT FormatCode      (#PCDATA)>
 270
 271
 272 <!-- The LC_CALENDAR element defines calendars used with a locale. -->
 273 <!ELEMENT LC_CALENDAR (Calendar* ) >
 274 <!-- At least one Calendar element must be given if the RefLocale mechanism is not used! -->
 275 <!ATTLIST LC_CALENDAR %RefLocale;>
 276
 277 <!ELEMENT Calendar (DaysOfWeek, MonthsOfYear, GenitiveMonths*, PartitiveMonths*, Eras, StartDayOfWeek, MinimalDaysInFirstWeek) >
 278 <!ATTLIST Calendar %UNOModule;>
 279 <!-- The unoid of a gregorian calendar MUST be lower case "gregorian",
 280      calendars MUST match the names defined in the OASIS OpenDocument Format
 281      (ODF) 1.2 or later specification. The implementation name registered with
 282      the OOo service resgistry MUST match, e.g.
 283      com.sun.star.i18n.Calendar_gregorian
 284 -->
 285 <!ATTLIST Calendar default (true|false) #REQUIRED >
 286 <!-- Exactly one Calendar element has to be the default calendar. -->
 287
 288 <!ELEMENT DaysOfWeek (Day*)>
 289 <!-- All Day elements of a Calendar must be given if the RefLocale mechanism is not used! -->
 290 <!ATTLIST DaysOfWeek %RefLocale;>
 291 <!-- Sequence of days is important, MUST start with Sunday. -->
 292 <!ELEMENT Day (DayID, DefaultAbbrvName, DefaultFullName, DefaultNarrowName*)>
 293 <!ELEMENT DayID (#PCDATA)>
 294 <!-- Preferably the lower case abbreviated English name like sun for Sunday. -->
 295 <!ELEMENT DefaultAbbrvName (#PCDATA)>
 296 <!-- The abbreviated day name, e.g. Sun for Sunday. -->
 297 <!ELEMENT DefaultFullName (#PCDATA)>
 298 <!-- The full day name, e.g. Sunday for Sunday. -->
 299 <!ELEMENT DefaultNarrowName (#PCDATA)>
 300 <!-- The narrow day name, e.g. S for Sunday.
 301      If not specified, the first letter of the corresponding DefaultFullName is taken.
 302  -->
 303
 304 <!ELEMENT MonthsOfYear (Month*)>
 305 <!-- All Month elements of a Calendar must be given if the RefLocale mechanism is not used! -->
 306 <!ATTLIST MonthsOfYear %RefLocale;>
 307 <!-- Sequence of months is important, MUST start with the first month of a
 308      year, e.g. January in a Gregorian calendar.
 309  -->
 310
 311 <!ELEMENT GenitiveMonths (Month*)>
 312 <!-- Possessive genitive case month names, for example in Slavic locales. The
 313      element is optional, but if present all Month elements of a Calendar must
 314      be given if the RefLocale mechanism is not used! If not present,
 315      MonthsOfYear names will be used. -->
 316 <!ATTLIST GenitiveMonths %RefLocale;>
 317 <!-- Sequence of months is important, MUST start with the first month of a
 318      year, e.g. January in a Gregorian calendar.
 319  -->
 320
 321 <!ELEMENT PartitiveMonths (Month*)>
 322 <!-- Partitive case month names, for example in Finnish locales. The
 323      element is optional, but if present all Month elements of a Calendar must
 324      be given if the RefLocale mechanism is not used! If not present,
 325      GenitiveMonths names will be used. -->
 326 <!ATTLIST PartitiveMonths %RefLocale;>
 327 <!-- Sequence of months is important, MUST start with the first month of a
 328      year, e.g. January in a Gregorian calendar.
 329  -->
 330
 331 <!ELEMENT Month (MonthID, DefaultAbbrvName, DefaultFullName, DefaultNarrowName*)>
 332 <!ELEMENT MonthID (#PCDATA)>
 333 <!-- Preferably the lower case abbreviated English name like jan for January. -->
 334
 335 <!ELEMENT Eras (Era*)>
 336 <!-- All Era elements of a Calendar must be given if the RefLocale mechanism is not used! -->
 337 <!ATTLIST Eras %RefLocale;>
 338 <!-- The eras MUST be in chronological order, e.g. first BC then AC. -->
 339 <!ELEMENT Era (EraID, DefaultAbbrvName, DefaultFullName)>
 340 <!ELEMENT EraID (#PCDATA)>
 341 <!-- If a calendar has special eras (like zh_TW ROC or ja_JP Gengou calendar)
 342      and a date before those eras is undefined, a leading (first) dummy era
 343      with EraID="Dummy" has to be defined to enable the number formatter to
 344      fall back to a Gregorian calendar for those date values if the XCalendar
 345      implementation returns an era value of 0.
 346 -->
 347
 348 <!ELEMENT StartDayOfWeek (DayID)>
 349 <!-- MUST exactly match (case significant!) one of the DayID of DaysOfWeek -->
 350
 351 <!ELEMENT MinimalDaysInFirstWeek (#PCDATA)>
 352 <!-- The number of days of a week that must reside in the beginning of a year
 353      to make a week the first week of the year. For example, a value of 4 means
 354      that at least 4 days of a week must be in the new year. So if the week
 355      starts on Monday, the first week of a year will be the week where Thursday
 356      is in the new year.
 357 -->
 358
 359
 360 <!-- The LC_CURRENCY element defines currencies used with a locale. -->
 361 <!ELEMENT LC_CURRENCY (Currency* ) >
 362 <!-- At least one Currency element must be given if the RefLocale mechanism is not used! -->
 363 <!ATTLIST LC_CURRENCY %RefLocale;>
 364 <!ELEMENT Currency (CurrencyID, CurrencySymbol, BankSymbol, CurrencyName, DecimalPlaces)>
 365 <!ATTLIST Currency  default (true|false) #REQUIRED >
 366 <!-- Exactly one Currency element has to be the default currency. -->
 367 <!ATTLIST Currency  usedInCompatibleFormatCodes (true|false) #REQUIRED >
 368 <!-- If this currency is the one used in compatible number format codes with
 369      <member>FormatElement::formatIndex</member> values in the range 12..17.
 370      Those format codes are used to generate some old style currency format
 371      codes for compatibility with StarOffice5 and StarOffice4.
 372      Every locale data file MUST contain exactly one currency having this set to "true",
 373      and that currency MUST be used in format codes 12..17.
 374      For European countries using the EURo it MUST be the old currency, for example, DM.
 375 -->
 376 <!ATTLIST Currency  legacyOnly (true|false) #IMPLIED >
 377 <!-- If this Currency element exists only to be able to correctly load legacy
 378      documents and is not selectable in the UI otherwise. Defaults to "false"
 379      if not specified. If this attribute is "true", 'default' and
 380      'usedInCompatibleFormatCodes' must both be "false".
 381 -->
 382 <!ELEMENT CurrencyID (#PCDATA)>
 383 <!-- The ISO 4217 three letter currency code, e.g. USD or EUR. -->
 384 <!ELEMENT CurrencySymbol (#PCDATA)>
 385 <!-- The currency symbol, e.g. $ or €. -->
 386 <!ELEMENT BankSymbol (#PCDATA)>
 387 <!-- The ISO 4217 three letter currency code, e.g. USD or EUR. -->
 388 <!ELEMENT CurrencyName (#PCDATA)>
 389 <!-- The native currency name, e.g. Dollar or Euro. -->
 390 <!ELEMENT DecimalPlaces (#PCDATA)>
 391 <!-- Number of decimal places used with the currency, usually 2 or 0, e.g. 2
 392      for cents.
 393 -->
 394
 395
 396 <!ELEMENT LC_CTYPE (Separators?, Markers?, TimeAM?, TimePM?,  MeasurementSystem?)>
 397 <!-- All elements must be given if the RefLocale mechanism is not used! -->
 398 <!ATTLIST LC_CTYPE  %RefLocale;>
 399 <!ATTLIST LC_CTYPE  %UNOModule;>
 400
 401 <!ELEMENT Separators (DateSeparator, ThousandSeparator, DecimalSeparator, TimeSeparator, Time100SecSeparator, ListSeparator, LongDateDayOfWeekSeparator, LongDateDaySeparator, LongDateMonthSeparator, LongDateYearSeparator)>
 402 <!ELEMENT DateSeparator (#PCDATA)>
 403 <!ELEMENT ThousandSeparator (#PCDATA)>
 404 <!ELEMENT DecimalSeparator (#PCDATA)>
 405 <!ELEMENT TimeSeparator (#PCDATA)>
 406 <!ELEMENT Time100SecSeparator (#PCDATA)>
 407 <!ELEMENT ListSeparator (#PCDATA)>
 408 <!ELEMENT LongDateDayOfWeekSeparator (#PCDATA)>
 409 <!ELEMENT LongDateDaySeparator (#PCDATA)>
 410 <!ELEMENT LongDateMonthSeparator (#PCDATA)>
 411 <!ELEMENT LongDateYearSeparator (#PCDATA)>
 412
 413 <!ELEMENT Markers (QuotationStart, QuotationEnd, DoubleQuotationStart, DoubleQuotationEnd)>
 414 <!ELEMENT QuotationStart (#PCDATA)>
 415 <!ELEMENT QuotationEnd  (#PCDATA)>
 416 <!ELEMENT DoubleQuotationStart (#PCDATA)>
 417 <!ELEMENT DoubleQuotationEnd (#PCDATA)>
 418
 419 <!ELEMENT TimeAM (#PCDATA)>
 420 <!ELEMENT TimePM (#PCDATA)>
 421 <!ELEMENT MeasurementSystem  (#PCDATA)>
 422
 423
 424 <!ELEMENT LC_COLLATION (Collator*, CollationOptions?)>
 425 <!-- All elements must be given if the RefLocale mechanism is not used! -->
 426 <!ATTLIST LC_COLLATION %RefLocale;>
 427 <!--
 428         Optional ICU tailoring.
 429
 430         See Collation Customization in ICU User Guide for syntax,
 431         http://www.icu-project.org/userguide/Collate_Customization.html
 432
 433         There are two ways to add language specific tailoring in OpenOffice.org.
 434         For small tailoring, you can directly add it in locale data here. For
 435         large tailoring, it is suggested to put the data under collator/data, so
 436         it will be compiled to a binary format in build time and improve performance
 437         in run time.
 438
 439 -->
 440 <!ELEMENT Collator (#PCDATA)>
 441 <!ATTLIST Collator  %UNOModule;>
 442 <!ATTLIST Collator  default (true|false) #REQUIRED >
 443 <!ELEMENT CollationOptions (TransliterationModules+)>
 444 <!ELEMENT TransliterationModules (#PCDATA)>
 445
 446
 447 <!ELEMENT LC_SEARCH (SearchOptions?)>
 448 <!-- All elements must be given if the RefLocale mechanism is not used! -->
 449 <!ATTLIST LC_SEARCH %RefLocale;>
 450 <!ELEMENT SearchOptions (TransliterationModules+)>
 451
 452
 453 <!ELEMENT LC_INDEX (IndexKey*, UnicodeScript*, FollowPageWord*)>
 454 <!ATTLIST LC_INDEX %RefLocale;>
 455
 456 <!--
 457     The IndexKey element is optional, but should be given if the locale
 458     requires a specific sort order in Writer's index tables or entries are to
 459     be combined under keys.
 460
 461     Index key for the algorithm and language, like >A-Z< for English => A, B,
 462     C, ..., Y, Z. The letters specify under which key an entry goes and the
 463     order the keys are sorted. Keys may be reordered or letters inserted to
 464     form a specific order, for example (ve_ZA) >A-D Ḓ E-L Ḽ M N Ṋ Ṅ O-T Ṱ U-Z<
 465     Entries that don't match a defined key are appended to the index list. Used
 466     in Writer textprocessor.
 467
 468     The initial data was setup according to the ICU collation chart at
 469     http://oss.software.ibm.com/icu/charts/collation/
 470     Note: ICU site was moved to http://www.icu-project.org/ and as ICU per
 471     default uses CLDR, collation charts are available at
 472     http://unicode.org/cldr/comparison_charts.html#Collation
 473
 474     Possible notations in the syntax of the IndexKey element are:
 475
 476     '-' (dash): Ellipsis, all letters elided by the ellipsis are included as
 477     index keys in alphabetic order. For example, 'A-Z' includes all ASCII
 478     letters A to Z.
 479
 480     '[]' (square brackets): all letters included in square brackets are
 481     skipping letters. It is used for CTL languages, for example in Thai
 482     (th_TH), to skip prefix vowels or signs. For example, if ["] double quote
 483     is defined as skipping letter, index item '"Index"' will be under 'I', not
 484     '"'.
 485
 486     '{}' (curly brackets): define multiple letters index key, for example
 487     '{Cs}' is one of the index keys for Hungarian (hu_HU).
 488
 489     '()' (parentheses): define optional description for index key. If defined,
 490     the description will be shown as index key title, instead of the index key
 491     itself. For example, 'E(E, É)' in Hungarian adds entries with 'E' to the
 492     description 'E, É'.
 493
 494 -->
 495 <!ELEMENT IndexKey (#PCDATA)>
 496 <!ATTLIST IndexKey  %UNOModule;>
 497 <!ATTLIST IndexKey  %LIBModule;>
 498 <!ATTLIST IndexKey  default (true|false) #REQUIRED >
 499 <!ATTLIST IndexKey  phonetic (true|false) #REQUIRED >
 500
 501 <!--
 502     The Unicode script types are those of
 503     offapi/com/sun/star/i18n/UnicodeScript.idl, they define the code range for
 504     the language.
 505 -->
 506 <!ELEMENT UnicodeScript (#PCDATA)>
 507
 508 <!--
 509     The FollowPageWord entries were originally hard-coded in
 510     ../../indexentry/indexentrysupplier.cxx, most locales used the English
 511     ``p.'' and ``pp.'', valid data should be provided by native speakers.
 512     These words or abbreviations are used in the Writer's index table. The
 513     first FollowPageWord element is the abbreviation for "page" (p.), the
 514     second FollowPageWord element the abbreviation for "page and following
 515     pages" (pp.).
 516 -->
 517 <!ELEMENT FollowPageWord (#PCDATA)>
 518
 519
 520 <!ELEMENT LC_TRANSLITERATION (Transliteration*)>
 521 <!ATTLIST LC_TRANSLITERATION %RefLocale;>
 522 <!ELEMENT Transliteration EMPTY>
 523 <!ATTLIST Transliteration  %UNOModule;>
 524
 525 <!ELEMENT LC_MISC (ForbiddenCharacters?, BreakIteratorRules?, ReservedWords?)>
 526 <!ATTLIST LC_MISC %RefLocale;>
 527 <!ELEMENT ForbiddenCharacters (ForbiddenLineBeginCharacters, ForbiddenLineEndCharacters, LineBreakHangingCharacters)>
 528 <!ELEMENT ForbiddenLineBeginCharacters (#PCDATA)>
 529 <!ELEMENT ForbiddenLineEndCharacters (#PCDATA)>
 530
 531 <!-- optional break iterator rules for the languages.
 532      if defined, 5 rules should be all listed, each of them can be blank and default will be used.
 533      order of the rules is significant!
 534 -->
 535 <!ELEMENT BreakIteratorRules (EditMode, DictionaryMode, WordCountMode, CharacterMode, LineMode)>
 536 <!ELEMENT EditMode (#PCDATA)>
 537 <!ELEMENT DictionaryMode (#PCDATA)>
 538 <!ELEMENT WordCountMode (#PCDATA)>
 539 <!ELEMENT CharacterMode (#PCDATA)>
 540 <!ELEMENT LineMode (#PCDATA)>
 541
 542 <!ELEMENT ReservedWords  (trueWord, falseWord, quarter1Word, quarter2Word, quarter3Word, quarter4Word, aboveWord, belowWord, quarter1Abbreviation, quarter2Abbreviation, quarter3Abbreviation, quarter4Abbreviation)>
 543 <!-- order is significant! -->
 544 <!ELEMENT trueWord (#PCDATA)>
 545 <!ELEMENT falseWord (#PCDATA)>
 546 <!ELEMENT quarter1Word (#PCDATA)>
 547 <!ELEMENT quarter2Word (#PCDATA)>
 548 <!ELEMENT quarter3Word (#PCDATA)>
 549 <!ELEMENT quarter4Word (#PCDATA)>
 550 <!ELEMENT aboveWord (#PCDATA)>
 551 <!ELEMENT belowWord (#PCDATA)>
 552 <!-- aboveWord and belowWord are used as reference field content in Writer, it
 553      is a "physical" object relative position, not "numerical" relative.
 554 -->
 555 <!ELEMENT quarter1Abbreviation (#PCDATA)>    <!-- "Q1" (quarter), "T1" (trimestre), ... -->
 556 <!ELEMENT quarter2Abbreviation (#PCDATA)>
 557 <!ELEMENT quarter3Abbreviation (#PCDATA)>
 558 <!ELEMENT quarter4Abbreviation (#PCDATA)>
 559
 560
 561 <!--
 562     In numbering levels, the NumType attribute is a value of the constants
 563     defined in offapi/com/sun/star/style/NumberingType.idl
 564 -->
 565
 566 <!ELEMENT LC_NumberingLevel (NumberingLevel* )>
 567 <!ATTLIST LC_NumberingLevel %RefLocale;>
 568 <!ELEMENT NumberingLevel EMPTY>
 569 <!ATTLIST NumberingLevel Prefix CDATA #REQUIRED>
 570 <!ATTLIST NumberingLevel NumType CDATA #REQUIRED>
 571 <!ATTLIST NumberingLevel Suffix CDATA  #REQUIRED>
 572 <!ATTLIST NumberingLevel Transliteration CDATA #IMPLIED>
 573 <!ATTLIST NumberingLevel NatNum CDATA #IMPLIED>
 574
 575
 576 <!ELEMENT LC_OutLineNumberingLevel (OutlineStyle* )>
 577 <!ATTLIST LC_OutLineNumberingLevel %RefLocale;>
 578 <!ELEMENT OutlineStyle (OutLineNumberingLevel+)>
 579 <!ELEMENT OutLineNumberingLevel EMPTY>
 580 <!ATTLIST OutLineNumberingLevel Prefix CDATA  #REQUIRED >
 581 <!ATTLIST OutLineNumberingLevel NumType CDATA  #REQUIRED >
 582 <!ATTLIST OutLineNumberingLevel Suffix CDATA  #REQUIRED >
 583 <!ATTLIST OutLineNumberingLevel BulletChar CDATA  #REQUIRED >
 584 <!ATTLIST OutLineNumberingLevel BulletFontName CDATA   #REQUIRED>
 585 <!ATTLIST OutLineNumberingLevel ParentNumbering CDATA  #REQUIRED >
 586 <!ATTLIST OutLineNumberingLevel LeftMargin CDATA   #REQUIRED>
 587 <!ATTLIST OutLineNumberingLevel SymbolTextDistance CDATA  #REQUIRED >
 588 <!ATTLIST OutLineNumberingLevel FirstLineOffset CDATA  #REQUIRED >
 589 <!ATTLIST OutLineNumberingLevel Transliteration CDATA #IMPLIED >
 590 <!ATTLIST OutLineNumberingLevel NatNum CDATA #IMPLIED>