We need to add floating point support for this milestone.

We'll need to support several symbol tables in order to properly implement scoped symbols. The simplest way to do this is to add a link to the next higher scoped symbol table with each symbol table and store a pointer to the top symbol table for processing.

1- is currently parsed as two tokens, 1 and -. This needs to be parsed as the single symbol token 1-.

We need some tests for the code generator. These should verify that the output of the compiler matches the expected gforth program.

The parser does not yet have any tests.

We need some sort of design documentation for the parsing algorithm. This will start with a grammar for the language, and may need more. Research will need to be done for the recursive descent parser.

The lexer does not yet have any tests.

We need a syntax tree structure before we can create the parser.

Need a code generator to translate the syntax tree to gforth.

The syntax tree does not yet have any tests.

The lexer seems to be doing two jobs: parsing the file, and recognizing and parsing tokens. We should probably split the token recognition and parsing off into a separate class that is called by the lexer.

We'll need to improve our functional tests to include verification of our parsing phase.

For this milestone, we're required to implement a recursive descent parser. The interface should be abstract enough that we can easily replace it with a faster parsing algorithm if we choose to optimize.

We need to develop some built in symbols (specifically arithmetic operations). The list of symbols we need to support can be found at the bottom of http://classes.engr.oregonstate.edu/eecs/winter2013/cs480/assignments/M4-ConstantOnly.htm

The symbol table isn't being filled. This needs to be done while parsing.