Originally^204, Lisp syntax was similar to that of Algol and other higher level programming languages. In the first few months, that syntax gets the uniform shape of meta-expressions. Programmers translated programs into assembler personally, obtaining experiences necessary for writing a compiler. Unexpected consequence of definition and implementation of S-function eval is appereance of useful interpreter for programs represented in the computer memory in form of list structures. Maling and Silver implemented function READ that reads symbolic expressions and converts them into a list structure which made it possible for programmers to start writing programs in form of symbolic expressions.
The Lisp community didn’t accepted easily symbolic expressions as the language syntax. Already McCarthy deemed them as “hard to read”^205, “somewhat readable”^206, and M-expressions as “easier to read”.^207 Later he suggested marking parenthesis with arbitrary number of dots and commas, for example (.. and ..), where closing such parenthesis would close all others that were left open.^208 Also some other members of the Lisp project considered Lisp not practical because of the big number of parenthesis.^209,210 Lisp programmers, as noted by McCarthy, if a problem requires more significant work with symblic expressions gladly choose other data representations and implement their own, ad hoc translators. For that reason he considered that support for expressions in other forms should be
built into Lisp.^211 Maling and Silver also tried, beside never finnished^212 support for meta-expressions, to support “algebraic expressions” such as a*b+c*d.^213 There were thoughts about S-function eval accepting arguments in “usual notation”.^214 Programs in memoes and books were still long written in form of M-expressions. It is also less known that Lisp 1.5 also support expressions in form of
(IF (a THEN b ELSE c))^215
and similar form for FOR expression. After McCarthy’s leave from MIT some public criticism appears^216 as well as attempts to create alternative forms of syntax.^217,218
Despite, the community kept symbolic expressions as the language syntax and gradually stopped using M-expressions. In Lisp 1.5 manual, prog-expression is introduced as meta-expression and equivalent S-expression, but without explicit rules for the translation.^219 The first bigger document in which there are no mentioning of M-expressions is Handbook of LISP functions by a group of students from Baltimore, from August 1961.^220
204 see chapter Imperative Lisp. 205 “This notation is rather formidable for a human to read …” McCarthy, Recursive functions …, AIM-008, 1959., p. 14. 206 “This notation is writable and somewhat readable.” McCarthy, Recursive functions …, CAMC, 1960., p. 189. 207 McCarthy et al., LISP I programmer’s manual, 1960., p. 53. 208 McCarthy, New eval function, AIM-034, 1962., p. 2. 209 “I thought it was a marvelous sort of language, except that it was somehow impractical. You can’t ask people to do that; you need a translator. Such programs for LISP may exist now. I don’t know. A program you could give simple statements to that would be automatically be translated into the multiply-parenthesized code.” Fox, An interview with Phyllis A. Fox, 2005., p. 31. 210 “The task of writing out S-expressions to define programs is a tedious one, especially since the LISP notation seems to run counter to the natural way that people think of mathematical expressions.” Abrahams, Application of LISP to checking mathematical proofs, 1964., p.159. 211 McCarthy, LISP − notes on its past and future, 1980., str. vi. 212 McCarthy, History of LISP, 1981., str. 179. 213 Maling, The Maling-Silver read program, AIM-013, 1959. 214 McCarthy et al., LISP programmer’s manual, 1959., str. 3/1. 215 McCarthy et al., LISP 1.5 programmers manual, 1962., str. 98. 216 “… it has become evident that the general list processing languages (such as LISP) are not appropriate for processing complex mathematical data. This is due not only to the extreme complexity of the coding that necessarily ensues in the analysis of an intricate problem, but also to the fact that the input-output procedures most ideally suited for these languages are totally inadequate for the individual prlmarily interested in the analysis of a problem and not the underlying mathematical and input-output structure involved.” Morris, Models for mathematical systems, 1966., str. 1520. 217 Henneman, An auxiliary language for more natural expressions …, 1964. 218 Abrahams et al., The LISP 2 programming language and system, 1966. 219 McCarthy et al., LISP 1.5 programmer’s manual, 1962., str. 29. 220 Conrad et al., A handbook of LISP functions, 1961.