There was a problem when proofreading this page.
Steele and Sussman
17
The Art of the Interpreter

interpreter, thanks to the merging of the procedural and variable environments, we may have not only bound procedure names, but also free variable names, for after all the two kinds of names are now one.

This interpreter differs in only small details from the one in LISP 1.5 [LISP 1.5M]. Both have dynamically scoped free variables (we will elaborate on this point later). We might note that the reference to VALUE in EVAL when computing the first argument for APPLY can be replaced by a reference to EVAL; this does the same thing if a variable appears in the operator position of a combination, and allows the additional general ability to use any expression to compute the procedure. This difference in fact appears in the LISP 1.5 interpreter. There are other slight differences, such as the representation of primitive operators and the handling of procedures which are not primitive or user-defined. Aside from these, the greatest difference between our interpreter and LISP 1.5's is the use of lambda notation. This we will meet in the next section.