An APL Compiler

Computing machines which directly execute the statements of a high level language have been proposed in the past. This report describes the actual implementation of such a machine: it is a computer whose “machine language” is APL. The machine is fully operational and correctly executes almost all of the APL operations on scalars, vectors, and arrays. The machine automatically allocates memory, executes statements, calls functions, converts numbers from one type to another, checks subscripts, and automatically detects many types of programmer errors.

Implementation of a high level language machine

High-level language computers (HLLC) have attracted interest in the architectural and programming community during the last 15 years; proposals have been made for machines directed towards the execution of various languages such as ALGOL,1,2 APL,3,4,5 BASIC,6,7 COBOL,8,9 FORTRAN,10,ll LISP,12,13 PASCAL,14 PL/I,15,16,17 SNOBOL,18,19 and a host of specialized languages. Though numerous designs have been proposed, only a handful of high-level language computers have actually been implemented.4,7,9,20,21 In examining the goals and successes of high-level language computers, the authors have found that most designs suffer from fundamental problems stemming from a misunderstanding of the issues involved in the design, use, and implementation of cost-effective computer systems. It is the intent of this paper to identify and discuss several issues applicable to high-level language computer architecture, to provide a more concrete definition of high-level language computers, and to suggest a direction for high-level language computer architectures of the future.

/pdf/retrospective-on-high-level-language-computer-architecture-57kbgukt6b.pdf

Retrospective on high-level language computer architecture

Data, like electricity and gravity, are part of the world in which we live. Some occur naturally, as in the genetic code, while most occur as a consequence of language and social organization. The search for a theory of data, which begins with the choice of a model, is as important and interesting as the development of theories in physics, economics, and psychology. Most models of data are collections, such as the unnested array of APL, the one-axis nested list of LISP, and the set, which is nested but lacks the properties of order, repetitions, type, and multiple axes inherent in rectangular arrangement. Nested rectangular arrays have all these properties. The existence of simple, universally valid equations in both set theory and linear algebra suggests that equally simple equations may hold for all arrays. The principles of nested collections developed in set theory apply with few changes to the nesting of arrays. A one-sorted theory of arrays, in which type is preserved for empty arrays, provides an algebra of operations interpreted not only for data but also types of data.

The nested rectangular array as a model of data

This paper is a survey of changes to virtual machine interfaces, implementation, architecture, and simulation techniques as they affect IBM System/370 and 303X (3031, 3032, 3033) processors, the system control program to which virtual machines interface, and other virtual machines executing on the same real computing system or elsewhere. The paper seeks to summarize such changes and provide a perspective on the virtual machine environment. New uses of virtual machine subsystems are discussed as they relate to inter-virtual-machine communication.

The changing virtual machine environment: interfaces to real hardware, virtual hardware, and other virtual machines

The APL language allows subscripted expressions such as A[I;J;K], where A is an array and I, J and K may be scalars, vectors or arrays of any size and shape. We describe an efficient method of evaluating these expressions. The method is quite general and yet it does recognize all the special cases in which multiple subscripts can be reduced to a simpler form. We show how the same mechanism can be used to evaluate other selection operations, such as transpose, and how selection operations may be combined.

Efficient evaluation of array subscripts of arrays

Emulation of an APL machine on a System/370 is exemplified by the APL Assist, a microprogram that enables APL. expressions and defined functions to be executed at the hardware level. This paper discusses what the APL Assist does, how it works, and the way it interacts with System/370 software. Execution times for APL programs with and without the Assist are compared.

An APL emulator on system/370

Trenchard More has proposed an array theory which offers a powerful set of operators and operations on nested arrays. We have written a program called Array Theory/370 (AT/370) which implements these array theoretic operations at the assembler-language level on an IBM/370. We have also written a VS APL auxiliary processor which enables the APL user to access AT/370 and to manipulate nested arrays. This paper describes the simple but powerful APL-AT/370 interface and illustrates how the interface was used to build an interactive array-theoretic system.

Array theory in an APL environment

In an earlier paper (1), we described the implementation of an APL machine on an IBM/360 model 25. we required that the APL system should not need a dedicated machine, but should run under a conventional operating system. This paper describes the microprogrammed part of APL/CMS and how it relates to the rest of the system; the APL/CMS User's Manual (2) describes the complete system.

A. Hassitt

Papers

Implementation of a high level language machine

Efficient evaluation of array subscripts of arrays

An APL emulator on system/370

Array theory in an APL environment

Direct execution of APL on an IBM/370