Anthony Ralston

Curriculum '78—is computer science really that unmathematical?

In 1979, Ralston was investigating curricula for discrete mathematics [95] and Shaw was participating in evaluations of Curriculum 78 and the role of mathematics in undergraduate computer science. They combined their notes to form a criticism of the mathematical content of Curriculum ’78 that appeared in Communications of the ACM [94]. Some comments on the paper appeared a few months later [69]

A new memoryless algorithm for de Bruijn sequences

is unique with subscripts in (2) taken modulo M. Therefore, the string (1) contains all possible substrings of length n. The original formulation of the de Bruijn sequence problem is connected to the number of Eulerian paths in a directed graph in which the nodes are labeled with substrings of length n 1 of our m + 1 symbols. The graphical interpretation is, however, not useful in finding algorithms for generating de Bruijn sequences. Thus, all published algorithms use either an approach based on the theory of finite fields or, as we do here, a direct combinatorial approach. For the case m = 1 the finite field approach leads to a simple algorithm based on a shift register approach (see, for example, [ 5, pp. 28-291). For m > 1, the finite field approach requires a knowledge of the coefficients of primitive polynomials of the fields of characteristics p for each prime p which divides m. Given these, however, the algorithm to generate de Btijn sequences is reasonably straightforward [8]. With the exception of the algorithm of Fredricksen and Maiorana [4] and the related algorithm for the case m = 1 of Fredricksen and Kessler [3] (see also Fredricksen[ 2]), all combinatorial algorithms require a storage of at least Mbits to record whether or not a given substring of length n has appeared in the sequence as it is generated. This is prodigal not only of

The first course in computer science needs a mathematics corequisite

First-year computer science students need to know and use a considerable amount of mathematics. A corequired course in discrete mathematics is a good solution.

The new (1982) Computing Reviews classification system—final version

I am pleased to announce that after consultation with all the other people involved, I have decided to adopt and implement the new Computing Reviews classification system, incorporating the changes described on page 14 by Anthony Ralston. The final version is on pages 17-25. Despite our having given people very little time, the responses we received were very constructive, as Tony has indicated, and many encouraged us to proceed and make the change. This will be done beginning with the January 1982 issue of CR. However, because we are simultaneously installing an automated system at ACM Headquarters to let the computer handle those things which it can do better than people, the staff is busily coping with these major changes as well as trying to keep CR and the A CM Guide to Computing Literature moving ahead. As a result of all these changes, the January 1982 issue of CR will surely be several months late. However, we hope that when the dust settles we will have the tools required to really make CR the type of review journal that is so needed in the computing field. As part of this activity, we invite new reviewers to help us to upgrade and revitalize CR. Let me urge all of you who have an interest in participating in the reviewing process to fill out the form on pages 15-16 and send it to: (Current reviewers will receive a form in the mail.) If you know any qualified colleagues who may be willing to review for CR, please send the Editor their names and addresses. With our new automated system now being developed, we should be able to make better assignments. I want to take this opportunity to thank Tony Ralston as the Chairman, as well as all the other members of the Computing Reviews Category Revision Committee (whose names were listed in the July 1981 issue of Communications) for their diligent and creative work. We expect CR eventually to become significantly better because of their efforts. Permission to copy without fee all or part of this material is granted provided that the copies are not made or distributed for direct commercial advantage, the ACM copyright notice and the title of the publication and its date appear, and notice is given that copying is by permission of the Association for Computing Machinery. To copy otherwise, or to republish, requires a fee and/or …

The proposed new Computing Reviews classification scheme

This report presents a proposed new classification system for Computing Reviews (CR) which has been developed over the past two years by a committee of 10 consisting of: Unless the feedback from this report causes us to make major changes in this scheme, every effort will be made to implement the new system on 1 January 1982. The current scheme used to classify reviews in CR has been in effect, with only quite minor changes, since it was installed in 1964 by Aaron Finerman, then Editor-in-Chief of CR. The 1964 scheme was a significant modification of the original CR classification scheme installed by John Carr (the first Editor-in-Chief of CR) in 1960. It hardly needs saying, given the changes in our discipline in the past 17 years, that the current classification scheme no longer reflects the computer field very well. (On the other hand, the fact that it is usable at all is a tribute to the soundness of the 1964 version.) But how much does this matter? Reasons can be adduced for and against changing an imperfect scheme Permission to copy without fee all or part of this material is granted provided that the copies are not made or distributed for direct commercial advantage, the ACM copyright notice and the title of the publication and its date appear, and notice is given that copying is by permission of the Association for Computing Machinery. To copy otherwise, or to republish, requires a fee and/or specific permission. 419 to one which (at least) better reflects the current state of the discipline: For the Change 1. For those CR readers who look for reviews in their specific areas of interest, the accuracy of the classification scheme may have a significant effect on the usefulness of CR. 2. For those individuals or libraries who use the CR classification scheme for their own filing or bibliographic purposes, accuracy is presumably also important. 3. For those who want to do retrospective searches, accuracy is important. Against the Change 4. For those CR readers who browse (or read) from cover to cover, the classification scheme probably does not make much difference. 5. For anyone used to the current scheme or who uses the current scheme (as in 2 above), a major change in the classification scheme could cause significant upset. Even if, as is likely, you believe that the first three are better arguments than the …

Structured Fortran—An Evolution of Standard Fortran

Fortran is 21 years old; many would say that it has not reached adulthood but senility. Yet it remains the language of use in the overwhelming majority of scientific applications of computers and no end to this situation is in sight. Moreover, its evolution through the standards process is very slow, too slow to give any hope that Fortran will ever in this way become a language in which structured programs can be easily and effectively written.

The demographics of candidates for faculty positions in computer science

A n t h o n y R a l s t o n The Demographics of Candidates for Faculty Positions in Computer Science The author describes experiences recruiting and selecting candidates for computer science faculty positions.

Algorithms for De Bruijn sequences—a case study in empirical analysis of algorithms

Most published work about the analysis of algorithms is analytic because, when analytic results can be obtained, they are usually decisive, often elegant and sometimes beautiful. However, such analyses are not always available either because they are mathematically intractable or because the analysis is indecisive. In such cases, empirical analysis may be used to distinguish the effectiveness of one algorithm from another. Indeed, this situation is more common than the computer science literature might lead you to believe. This note is an example of such an empirical analysis. 2. DE BRUIJN SEQUENCES

The mathematics component of the undergraduate curriculum in computer science (Panel Discussion)

Given the simple fact that mathematics forms the foundation of computing, mathematics must play a meaningful role in the computer science program. This meaningful role, however, must be substantially less, in content, than a second major. We must avoid the extreme of little or no mathematics in the curriculum. We must, at the same time, avoid the excess of too much mathematics. The need for restraint becomes most obvious when we recognize that the ideal liberal arts program should be approximately one-third for a major, one-third for a broad “liberal arts” component, and one-third for free electives. A curriculum including all of the mathematics suggested in “Curriculum-68”, all of the computer science now needed by an individual wishing to devote his or her post-baccalaureate life to computing, and a minimal number of necessary cognates from other disciplines, would strain the traditional concept of the conventional four-year liberal arts degree. Thus, the only alternatives to moderation in the mathematics component are moderation in the major itself (unthinkable), extension of the curriculum to five or more years, and/or conversion of the curriculum to a professional degree, i.e., an engineering degree.

The Decline of Calculus—The Rise of Discrete Mathematics

Calculus is one of the great triumphs of the human intellect. For this reason alone no educated person should be without some knowledge of it. When, in addition, you consider the panoply of intellectual and practical conquests of classical analysis, whose foundation is calculus, it is small wonder that calculus has been for so long the basis of all college mathematics study. It may well surprise the reader then that the purpose of this essay is to argue that the position of calculus in the college mathematics curriculum is ripe for change and, to a degree, decline.