Implementation of McCarthy's original Lisp in Python
Python Common Lisp Shell
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
rootlisp
tests
README.md
examples.lisp
lisp
run-tests.sh
stdlib.lisp

README.md

Root Lisp

Root Lisp is a Lisp, as originally described by John McCarthy, implemented in Python. It is based on the description in Recursive Functions of Symbolic Expressions and Their Computation by Machine, Part 1, but also highly inspired and influenced by the explanation in The Roots of Lisp by Paul Graham.

Just like the original this is a small, although neat, language with a lot of bells a whistles missing. It has no side effects (no IO), no types other than atoms (e.g. no numbers, strings, etc), no error handling (programs are expeted to be expressed correctly), and dynamic rather than lexical scoping.

Differences from McCarthys description

This language differs from the original Lisp in some minor regards.

  • Atoms can be written in lower case letters and with special characters (except single quote ' and parentheses), in addition to the originals CAPS ONLY atoms.
  • Atoms cannot contain spaces, thus elminating the need for dotted pair notations or commas within lists.
  • Quoting can be done using single quote tick (e.g. '(a quoted list) instead of (quote a quoted list)) in order to improve readability of programs a bit.

In addition, I could not figure out how to define and refer to functions outside of label expressions, and thus introduced a defun construct like the one described by Graham. If anyone have inputs on how to do this without introducing a new "axiomatic" form I'd be delighted for a message or pull-request.

The implementation itself also differ a lot, obviously, being written in Python rather than for the IBM 704, but I tried to keep the semantics regarding the computational model as close to the original as possible.

What's what?

If you want to have a look at the code, here is an overview of what is where. The most interesting file is probably core.py where eval() is defined.

root-lisp/
├── lisp            # the "executable" for starting a REPL or executing a file
├── README.md
├── rootlisp
│   ├── core.py     # the language core, exposing the eval() function
│   ├── __init__.py
│   ├── lisp.py     # interpret() and REPL
│   └── parser.py   # the parser functions, transforming Lisp source to ASTs and back
├── stdlib.lisp     # implementation of the language within itself
└── tests
    └── ...         # the unit tests

Play

Try out the Lisp using the REPL:

$ ./lisp 
> (cons (cons 'a '()) (cons 'b (cons 'c '(d e f))))                            
((a) b c d e f)
> (atom 'yep)
t
> (defun not (x) (cond (x 'f) ('t 't)))
not
> (not 't)
f
> (not 'f)
t
> (defun null (x) (eq x '()))
null
> (null '())
t
> (not (null '(not empty)))
t
> 

Or interpret a file by providing an argument

$ ./lisp examples.lisp
(a x b y c z d)

Run the tests

In order to run the test suite, you'll need to install nose.

$ pip install nose

You can then run the tests using nosetests from the project's root directory.

$ nosetests
................................................
----------------------------------------------------------------------
Ran 48 tests in 0.033s

OK