Iverson's Language(ID:421/ive001)Proto-APLAlso Iverson's Notation The original common name for what became APL: coming out of work with Fred Brooks "Automatic Data Processing", developed as a shorthand for presentation of data. In 1960, owing to the size of the manuscript, John Wiley (publisher) proposed dividing the book in two volumes, one of them on data processing, the other on the new notation. Then in 1964 Iverson writes "Formalism in Programming Languages" (ACM, Volume 7, number 2, February 1964) devloping this idea further. With Falkoff at IBM, they write a description for the System 360 in the new format, which is a landmark in concision Related languages
References: Introduction Although the question of equivalences between algorithms expressed in the same or different languages has received some attention in the literature, the more practical question of formal identities among statements in a single language has received virtually none. The importance of such identities in theoretical work is fairly obvious. The present paper will be addressed primarily to the practical implications for a compiler. The formal identities can be incorporated directly into a compiler, or can alternatively be used by a programmer to derive a more efficient equivalent of a program specified by an analyst. The identities cited include (1) dualities which permit the inclusion of only one of a dual pair as a basic operator, (2) partitioning identities which permit the automatic allocation of limited fast-access storage in operations on arrays, (3) permutation identities which permit the adoption of a processing sequence suited to the particular representation used (e.g., row list or column list of a matrix), (4) general associativity and distributivity identities for double operators (determined as a function of the properties of the basic operators) which permit efficient reordering of operations, (5) transposition identities, and (6) the automatic extension of the appropriate identities to any ad hoc operations (i.e., subroutines or procedures) defined by any user of the compiler. The discussion will be based upon a programming language which has been presented in full elsewhere [1]. However, the relevant aspects of the language will first be summarized for reference. The problems of transliteration and syntax which commonly dominate discussions of language will here be subordinated as follows. The symbols employed will permit the immediate determination of the class to which each belongs; thus literals are denoted by roman type, variables are denoted by italics (lowercase, lowercase bold, and uppercase bold for scalar, vector and matrix, respectively), and operators are denoted by distinct (usually nonalphabetic) symbols. The problems of transliteration (i.e., mapping the set of symbols employed onto the smaller set provided in a computer) and of mapping positional information (such as subscripts and superscripts) onto a linear representation therefore can, and will, be subordinated to questions of the structure of an adequate language. in [ACM] CACM 7(02) (Feb 1964) includes Papers presented at a Working Conference on Mechanical Language Structures, Princeton, N. J., August 1963 view details in [ACM] CACM 7(02) (Feb 1964) includes Papers presented at a Working Conference on Mechanical Language Structures, Princeton, N. J., August 1963 view details in [ACM] CACM 7(02) (Feb 1964) includes Papers presented at a Working Conference on Mechanical Language Structures, Princeton, N. J., August 1963 view details |