Melvin Klerer

Experimental study of a two-dimensional language vs. FORTRAN for first-course progammers

The variability in programming performance for a group of novice Fortran programmers was measured over a set of problems from an introductory programming text. A wide variation was observed despite the elementary nature of each problem and the relatively homogeneous subject group. The implications of these results are examined. Another experiment measured the comparative performance of programming novices using Fortran and the Klerer-May two-dimensional (2-D) language. The results indicated that the 2-D language was much more economically efficient than Fortran for the subject groups in the areas of scientific/engineering application programming.

The effect of language design on time-sharing operational efficiency

The importance of “think time” for operational efficiency of time-sharing systems is re-emphasized. It is pointed out that think time is an experimental dependent parameter of the software-hardware programming system and may be lengthened or shortened as a function of console or programming language design. A simple computational model is used to predict the behavior of response time as a function of think time for different conditions of service loading. The economic implications are considered.

The economics, politics, and sociology of two-dimensional systems

The various factors contributing to the superiority of two-dimensional systems over linear programming systems in terms of their economic advantages is discussed. The failure of 2-D systems to affect current praxis is analyzed in terms of the social and political milieu that characterized computer science research during the previous decade.

Timing results of various compilers using an optimization quality benchmark

According to the criteria used to develop the benchmark test , the ratio of execution times of the program NONOP to OP indicate s the quality of the compiler . The closer the ratio to 1, th e better the quality of optimization . We have defined quality t o be the capacity of a compiler to optimize a set of programmin g structures that are most commonly discussed in compiler textbook s as susceptible to optimization . These eleven structures are : constant, local common subexpression, global commo n subexpression, unnecessary copy, code motion, induction variable , reduction in strength, dead code, loop jamming, redundant code , and unreachable code . The number of iterations of each of thes e structures is equal to the statistical weight observed for eac h structure in an empirical study of programming style [3] among a wide range of users in a research university environment . Other source level structures were not included since their observe d occurrences were not statistically significant .

The Nuclear Winter Hypothesis: Implications for Scientific Credibility

Recent publications raise the possibility that a limited nuclear war could result in global effects that might cause the extinction of most biological species. During the past four decades, this possibility has not been considered in publications which represented a consensus of ``official scientific opinion. The implications for scientific credibility are discussed.

Interactive programming and automated mathematics

The theme of these symposium proceedings is Interactive Systems for Experimental Applied Mathematics. Certainly, there can be little confusion about the basic meaning or broad fundamentals of applied mathematics. However, questions as to the value of the experimental approach toward mathematics are an entirely different matter. The present state-of-the-art is so scant in empirical or theoretical guidelines that this approach must be acknowledged as an expression of faith that a computer, used to explore ill-defined mathematical constructs and problems, might yield powerful insights and a fruitful methodology. The term interactive is difficult to define and I cannot pretend that I fully understand the relevance of the term as applied to some specific systems. However, in some sense, we would suppose that what we mean is close to the definition used by the physicist. When two effects interact, they do so in a nonseparable and usually nonlinear way. Usually, the effect of the interaction is that the whole is not simply equal to the sum of its parts, and that the characteristics of the interacted system can be surprisingly different from the qualities of its constituent parts. Therefore, in the man-machine interaction, we would expect more than in the old process of inputting a well-formulated set of directions with the machine performing in its capacity as an idiot savant. Obviously we can expect that any serious attempt at man-machine interaction will involve on-line response. However, I would not entirely exclude the possibility that significant interaction can occur off-line.

Design philosophy for an interactive keyboard terminal

The basic version of the Klerer--May programming system has been in operation at Columbia Universitys Hudson Laboratories for nearly four years in an off-line mode and for two years in an on-line mode. This system [1--3] permits programming using normal two-dimensional mathematical expressions and flexible language forms. In the area of scientific applications, such a language approach permits faster total throughput, i.e., less time spent in programming or debugging a specific problem compared to conventional FORTRAN-like languages. It also offers a basic framework for extension into other areas such as the manipulation and editing of two-dimensional mathematical input for automatic typesetting of mathematical text [4]. A typical program segment in this language is illustrated in Fig. 1. More visually complex forms, such as multiple integrals, sums, products, and IF conditions are also recognized and compiled.