On Tue, 19 Feb 2008 16:42:03 +0100, Xavier Roche
<xroche@free.fr.NOSPAM.invalid> wrote:

Euh, vous pouvez éventuellement séparer les katakana, mais pour les
Kanji (idéogrammes), il est totalement impossible de séparer les mots
composés (de kanjis et/ou de kanas) des kanjis "isolés" puisque ce sont
les même caractères (au sens Unicode du terme) sans appliquer des
traitements (très) complexes.

En fait je bosse sur un logiciel. La "complexité" ici dépendra
seulement du temps d'exécution des instructions. Il suffit de
programmer l'algorithme pour le voir.

(Pour prendre un exemple, le "?" de "?"(?) (blanc) et le "?" de "?
?" (riz blanc) ou de "??" ("amusant") sont identiques)

Ca ne m'éclaire que partiellement.

Si je montre ce tableau de la langue des 65536 premiers unicodes, on
peut les localiser ?

  TLanguage =

(Basic_Latin,Latin_1_Supplement,Latin_Extended_A,Latin_Extended_B,IPA_Extensions,Spacing_Modifier_Letters,

Combining_Diacritical_Marks,Greek_and_Coptic,Cyrillic,Cyrillic_Supplement,Armenian,Hebrew,Arabic,Syriac,Arabic_Supplement,

Thaana,Devanagari,Bengali,Gurmukhi,Gujarati,Oriya,Tamil,Telugu,Kannada,Malayalam,Sinhala,Thai,Lao,Tibetan,Myanmar,Georgian,

Hangul_Jamo,Ethiopic,Ethiopic_Supplement,Cherokee,Unified_Canadian_Aboriginal_Syllabics,Ogham,Runic,Tagalog,Hanunoo,Buhid,

Tagbanwa,Khmer,Mongolian,Limbu,Tai_Le,New_Tai_Lue,Khmer_Symbols,Buginese,Phonetic_Extensions,Phonetic_Extensions_Supplement,

Combining_Diacritical_Marks_Supplement,Latin_Extended_Additional,Greek_Extended,General_Punctuation,Superscripts_and_Subscripts,

Currency_Symbols,Combining_Diacritical_Marks_for_Symbols,Letterlike_Symbols,Number_Forms,Arrows,Mathematical_Operators,

Miscellaneous_Technical,Control_Pictures,Optical_Character_Recognition,Enclosed_Alphanumerics,Box_Drawing,Block_Elements,

Geometric_Shapes,Miscellaneous_Symbols,Dingbats,Miscellaneous_Mathematical_Symbols_A,Supplemental_Arrows_A,Braille_Patterns,

Supplemental_Arrows_B,Miscellaneous_Mathematical_Symbols_B,Supplemental_Mathematical_Operators,Miscellaneous_Symbols_and_Arrows,

Glagolitic,Coptic,Georgian_Supplement,Tifinagh,Ethiopic_Extended,Supplemental_Punctuation,CJK_Radicals_Supplement,Kangxi_Radicals,

Ideographic_Description_Characters,CJK_Symbols_and_Punctuation,Hiragana,Katakana,Bopomofo,Hangul_Compatibility_Jamo,Kanbun,

Bopomofo_Extended,CJK_Strokes,Katakana_Phonetic_Extensions,Enclosed_CJK_Letters_and_Months,CJK_Compatibility,CJK_Unified_Ideographs_Extension_A,

Yijing_Hexagram_Symbols,CJK_Unified_Ideographs,Yi_Syllables,Yi_Radicals,Modifier_Tone_Letters,Syloti_Nagri,Hangul_Syllables,

High_Surrogates,High_Private_Use_Surrogates,Low_Surrogates,Private_Use_Area,CJK_Compatibility_Ideographs,Alphabetic_Presentation_Forms,

Arabic_Presentation_Forms_A,Variation_Selectors,Vertical_Forms,Combining_Half_Marks,CJK_Compatibility_Forms,Small_Form_Variants,
 Arabic_Presentation_Forms_B,Halfwidth_and_Fullwidth_Forms,Specials);


Sachant que chaque type de caractère est défini ainsi:

  Lu Letter, Uppercase
  Ll Letter, Lowercase
  Lt Letter, Titlecase
  Lm Letter, Modifier
  Lo Letter, Other
  Mn Mark, Nonspacing
  Mc Mark, Spacing Combining
  Me Mark, Enclosing
  Nd Number, Decimal Digit
  Nl Number, Letter
  No Number, Other
  Pc Punctuation, Connector
  Pd Punctuation, Dash
  Ps Punctuation, Open
  Pe Punctuation, Close
  Pi Punctuation, Initial quote
  Pf Punctuation, Final quote
  Po Punctuation, Other
  Sm Symbol, Math
  Sc Symbol, Currency
  Sk Symbol, Modifier
  So Symbol, Other
  Zs Separator, Space
  Zl Separator, Line
  Zp Separator, Paragraph
  Cc Other, Control
  Cf Other, Format
  Cs Other, Surrogate
  Co Other, Private Use
  Cn Other, Not Assigned


Merci encore pour toute aide.

--
Jean-Phil