-----------
  VERSION 1.2
  -----------
  
  
   - VERSION 1.2  -  date: April 2003
  
  	 * bug fixed (April 2004):
  	 	 - the phonetization of a single '0' was doubled (zero zero). This is now fixed.
  
       * new function
           - script/lia_lex2phon_variante: phonetization of a lexicon according to all the
             Part-Of-Speech found in the lexicon $LIA_PHON_LEX or lex10k if not
             defined AND generation of pronunciation variant according to a set of rules
             contained in the file: data/rule_variante_lex.pro
  
   - VERSION 1.1  -  date: January 2003
  
       * bug fixed:
           - all 'main' function end with an 'exit(0)'
           - the script 'lia_check' is done on the 'pho' format and
             not the 'ola' one
      
       * modif:
           - now, the lexicon used can be specified in a envt variable:
      	   LIA_PHON_LEX
             for example: set LIA_PHON_LEX = lex10k
             If this variable is not defined, the default lexicon, lex10k,
             is used
           - all the scripts starts now with "#!/bin/csh"
      
       * new function
           - script/lia_lex2phon: phonetization of a lexicon according to all the
             Part-Of-Speech found in the lexicon $LIA_PHON_LEX or lex10k if not
             defined
      
       * new data
           - a lexicon with more than 80k words is part of the distribution.
             This lexicon is useful for the POS tagging programs and for the
             phonetization of a lexicon. But just for text-to-speech purpose,
             it's better to use only the default lex10k lexicon.
             To use this lexicon, compile it first with the command:
             make lex80k
             Then, modify the envt variable 'LIA_PHON_LEX' as:
             setenv LIA_PHON_LEX lex80k