a) lose the 2 types of comment lexical scanning regex. remove ? at end of comment. Add a type for '@'
  b) use words() function for builtins
  c) remove whitespace lexer class
  d) streamline punctuation class
  e) unable to modify keywords lexer class to use words() function, so keeping it
  f) streamline identifier use with _taletters variable
  g) remove hex lexer class
  h) update float and integer lexer class. N.B. Ezhil does not use Tamil-number formats - instead we use standard Arabic/Hindu system
2) Update tests for fallout from #1

2) add example Ezhil script 'ezhil_primefactors.n'