I am planning to use indexing.  I agree that indexing is necessary.   
   Indexing also has to be able to work across indexing boundaries as   
   well.  If the algorithm discovers that an indexing choice is incorrect   
   and if it discovers that details of the reasons why it made the wrong   
   choice had been too close to call, the indexing system should contain   
   references to other possible choices.  In other words the sequence of   
   categorization that is produced by initial categorical indexing should   
   not be considered absolute becaus the program can make errors and the   
   errors may be too subtle or infrequent for the program to change the   
   score based indexing system.   
      
   I am using a learning based dictionary that adapts to new words as they   
   are learned.  The dictionary ]then refers to an internally coded system   
   that can represent more complicated conceptual structures. (That is, if   
   my project works at all it will be able to represent more complicated   
   structures.)  Now that I think about it again, a simple initial lookup   
   lexicon might be able to work effectively.  But I do realize that the   
   heart of the program has to be much more sophisticated.  I have been   
   working on my ideas for a long time.   
      
   Jim Bromer   
      
   [ comp.ai is moderated.  To submit, just post and be patient, or if ]   
   [ that fails mail your article to , and ]   
   [ ask your news administrator to fix the problems with your system. ]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)