Search

This article presents information on the execution time performance of the Cobol code versus PL/1 code that was generated by two compilers and tested by a benchmark program implemented in both languages. The goal of the experiments was to probe and study the relative performance characteristics of Cobol and PL/1 in a specific implementation. The results, according to the author should not be considered exhaustive but rather as a guide reflecting one particular set of insights in an area where little information is available, and where these insights are bound to be colored as much by the benchmark program used as by the specific translations from Cobol to PL/1. The article presents Test 1, which was originally designed to measure the speed of execution of the internal subroutine facility in Cobol, namely, the PERFORM statement. In this test, a sequence of nested PERFORMs is executed, the nesting is three levels deep and the first two levels contain TIMES options for looping. In PL/I the simplest way to implement this is via PROCEDURE calls without parameters.

This paper reviews the current state of programming language standards activities with respect to the anomalies which exist between the various published and proposed standards for Fortran, Cobol, PL/I, and Basic. Proposals are made for the inclusion of formalisms within future standards and the extension of the standards to include additional items such as error conditions and documentation. [ABSTRACT FROM AUTHOR]

Presents a technique for the symbol-table lookup of structured variables based on the automata theory. Use of the technique in solving a problem in PL/I and COBOL compilers; Basic functions needed in a compiler; Use of the hierarchical nature of the structured variables.

Comments on the technique for lookup of programming structured variables. Basis of the technique on the construction of a deterministic finite-state machine; Example of a PL/I structure; Difficulties of the PL/I and Cobol algorithm.