#------------------------------------------------------------------------------
# NCHLT Pronunciation Predictor README
#------------------------------------------------------------------------------

This predictor generates language-specific pronunciations for a given word list, 
using the NCHLT rule sets. 

Requirements:
-------------
- Installed Perl (tested with version >= 4)
- Set the DR_RULE_DIR environment variable to the directory where the NCHLT
  rules can be found
- Current code assumes a Linux-based implementation but as only standard Perl is
  used, other platforms may also support this code if properly configured.

Usage:
------

perl pron_predict.pl <in:words> <language> <out:dict>
     <words>     word list, one word per line, UTF8-encoded.
     <language>  specified using ISO code:
                 afr | eng | nbl | nso | sot | ssw | tsn | tso | ven | xho | zul
     <dict>      name of new output dictionary to be created
                 as a tab-delimited, UTF8-encoded file.

#------------------------------------------------------------------------------

