My take on write-yourself-a-scheme-in-48-hours from
http://en.wikibooks.org/wiki/Write_Yourself_a_Scheme_in_48_Hours
=======

To build, use cabal:

$ cabal sandbox init
$ cabal install --only-dependencies --enable-tests
$ cabal configure --enable-tests
$ cabal build

======
To run:

$ cabal run
tlisp>>> (load "stdlib.scm")
(lambda ("pred" . lst) ...)
tlisp>>> (map (curry + 2) ’(1 2 3 4))
(3 4 5 6)

=======
Interesting features, beyond those in write-yourself-a-scheme:

*: uses ContT to implement call-with-current-continuation
*: native support for integers, rationals, floating point and complex numbers
*: arithmetic operations perform type promotion between the numerical types
*: uses haskeline for command history and tab-completion in the repl
*: has doctest tests for the parser
*: models the environment using StateT [IORef (Data.Map String LispVal)]

=======
To do:

*: support immutable and mutable lisp values
*: add more tests, especially for the primitives
*: add eval support for quasiquotes and unquotes
*: add macros
*: integrate test.scm into the tests (currently you just run it and look for #t as the output)

=======
To run the tests:

$ cabal test
Building tlisp-0.1.0.0...
Preprocessing executable 'tlisp' for tlisp-0.1.0.0...
Preprocessing test suite 'doctests' for tlisp-0.1.0.0...
Running 1 test suites...
Test suite doctests: RUNNING...
Test suite doctests: PASS
Test suite logged to: dist/test/tlisp-0.1.0.0-doctests.log
1 of 1 test suites (1 of 1 test cases) passed.