Initital skeleton of separation of separate parsing scheme This is the first commit in a long series related to separating the parsing of the input code from the execution of the code. It should allow for more efficient, and probably simpler, execution while giving quicker feedback when someone types in syntactically invalid code.

AND,token_and
DATA,token_data
ELSE,token_else
END,token_end
GO,token_go
LET,token_let
LIST,token_list
NEW,token_new
NOT,token_not
OR,token_or
POP,token_pop
PRINT,token_print
REM,token_rem
RETURN,token_return
RUN,token_run
STOP,token_stop
SUB,token_sub
TO,token_to