autoconf TODO:
- Copy more README's to the rebecca/share directory and below.

TODO First:
- Fix the logging mechanism and ensure every method logs at least LOG_BOT_METHOD
- Fix the html generator to have a CustomLearn tag
- Add a method for getting all the predicates from the system.
- Add regression test suite as part of the distribution but keep it 
  it still as implementation specific.
		 * Add it to the autoconf build system under Linux
		
- Make methods on the builder for policies based on what to do when hitting 
    identical "PATTERN <that> THAT <topic> TOPIC"  They would be:
        * skip
        * append
        * overwrite
        * combine
- Make a callback for when an identical "PATTERN <that> THAT <topic> TOPIC" is
    encountered it will call the callback.
- Fix the numeric thatTagNumericConversionError:Second index string is not a valid number, * error

Future:
- Make the xml engine pluggable so that something such as libexpat could be used in place of xerces
- Add api for return name when set
- Make rebecca be able to use the <learn> tag from pandora bots as a layer
- Make Rebecca parse predicates.xml so that she can set predicates as well as the "return name when set".
- Make Rebecca set the predicate top in the predicates.xml to OM and 
     remove the top being set in the Predicates.aiml file.
- Should we perhaps make a utility dll/so?  At the least should we make a
  Prediacates class that encompasses the read/write predicates with a factory
  that can be set in the builder class.  That way the user can define their
  behvaoir of what to do with reading and writing the predicates 
- Support mulituser enviornment.  You can switch users through a "id"
- Add the Javascript engine

Far Future:
- Add multithreading
- Add multi-bot support efficiently.  This can be accomplished by add a "bot id" to the graphmaster.
- Document and release doxygen of the internal api
- Open up and document how others can make their own releases using the batch files.