From: ted.dunning@gmail.com   
      
   raul wrote:   
   > well to be more specific i have just done my schooling with computer   
   > sciences in c++ so my base is zero u see........   
   >   
      
   Peter Norvig's book is a good introduction to a wide range of topics.   
      
   You will need to branch out on languages beyond C++.  I tend to use   
   Java, but I do pretty mathematical/statistically oriented sorts of   
   "AI".  For more symbolic applications, I have used Prolog and Common   
   Lisp extensively, but don't know of any really good options any more.   
   Prolog especially is not very suited for integration into modern   
   delivery systems and I haven't heard much about Common Lisp keeping up   
   with the times either.  Languages like ML, Haskell and Dylan seem to   
   have become even more fractured than ever.   
      
   What you really need is the change notation to fit your problem while   
   still maintaining good debuggability and testability.  This is a very   
   hard problem that has never had a reasonable solution.  For large data   
   mining systems, I tend to write systems that assemble classes on the   
   fly from XML specifications (that may themselves be automagically   
   generated), but without enormous care, you can wind up with systems   
   that only the author can understand.  That can really be bad if you   
   want to ever do anything but maintain legacy code.  Again, these   
   approaches would only be partially useful for other styles of programs.   
      
   [ comp.ai is moderated ... your article may take a while to appear. ]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)