Lisp { Almost a Whole Truth !

Lisp is well known for its metacircular deenitions. They diier by their intent (what they want to prove), their means (what linguistic features are allowed for the deenition) and by their scope (what linguistic features are described). This paper provides a new metacircular deenition for a complete Lisp system including traditionally neglected features such as cons, read, print and error. The programming style adopted for this interpreter is inspired both by denotational semantics and its continuation passing style (to explain continuation handling) and by the object oriented paradigm as highlighted by type-driven generic functions. The resulting interpreter lessens the number of primitives it uses to only eight: car, cdr, rplaca, rplacd, eq, read-char, write-char and end, while still providing Scheme-like essential capabilities (less arithmetic). The overall size is near 500 lines of fully encapsulated code that, if eeciency is not the main requirement, can be easily turned into a stand alone program. Since the birth of Lisp, the art of the interpreter has always been considered as one of the major Lisp rites. Ranging from very simple McCarthy 78] to more complex Rees & Clinger 86], a Lisp interpreter is a must of many courses, many books Abelson & Sussman 85, Dybvig 87, Queinnec 84, Winston 88, : : : ] and many articles Friedman & Wand 84, Reynolds 72, : : : ]. The success of this cult is due to | (i) the amazing simplicity that can have such interpreters | (ii) the pedagogical beneets of either reading or writing such descriptions | (iii) the overall conndence given to the writer of such a program that s/he has now fully mastered the language. Interpreters also provide an interesting way to design variants of Lisp Steele & Sussman 78, des Rivieres & Smith 84]. They truly are speciications of languages Reynolds 72] from which can be derived compilers Clinger 84] or entire systems Brooks & Gabriel & Steele 82, Saint James 87]. Without quantifying their intentions (pedagogy Kessler 88], proof length Boyer & Moore 82] or even graphics Lakin 80]) a taxonomy of these interpreters can be attempted along two main axis, see gure 1: the speciied language (given to the user), the language for the speciication (allowed for the implementer). These axis are mere abstractions since power of languages is diicult to appreciate and is probably more related to trees than to straight lines. For example, allowing assignment …