This panel focuses on the challenges of collecting and analyzing data relating to retention of students in undergraduate computer science education programs. The panelists will share learnings and recommendations from the final report of the ACM Retention Committee and share their individual perspectives on data collection challenges, promising interventions, and recommendations for actively addressing retention for all students.

The main goal of this panel is to share survey data being collected on booming enrollments in computer science. Survey data are being collected from two source types: institutions and students. The institution data are being collected through both the CRA Taulbee Survey, which represents computer science departments in North America with doctoral degrees, and the ACM NDC Survey, which represents computer science departments at U.S. 4-year and master's universities and colleges. The student data are being collected via the Data Buddies Project, run by the CRA's Center for Evaluating the Research Pipeline. The institution data will allow us to answer questions such as "what 'boom' are institutions seeing and how are different types of institutions responding to this boom?" The student data will allow us to answer questions such as "why are students suddenly eager for our classes?". The survey data will help us understand (1) the intensity of the "boom" at different types of institutions, (2) how institutions are responding to the increasing demand for computing education, and (3) whether the current enrollment "boom" represents the "third bubble" or whether it is here to stay. The data will also allow us to ask questions that attempt to understand the impact of the boom on underrepresented students. Data collection is ongoing; we will present results "hot off the presses".

This panel will discuss the enrollment boom, as well as potential effective practice strategies to respond to the increasing demand for computing education. Does the current enrollment "boom" represent the "third bubble", or is it here to stay? Is there a difference in what's happening at universities that offer Ph.D. degrees or what's happening at 4-year universities and colleges? What 'best practice' strategies exist (if any) to handle the demand? And, importantly, will this "boom" have a positive or negative effect on underrepresented students' Join us for a lively discussion on the recent student interest in computing education at our universities and colleges.

No abstract available.

No abstract available.

No abstract available.

The Computing Accreditation Commission (CAC) of ABET has proposed changes in the structure of the criteria used to accredit computer science and information systems programs. The new structure combines, into so-called "general criteria," components of criteria that apply to any computing program accredited by CAC. Individual program areas will supply criteria that only apply to that specific program area, to supplement the general criteria. The new criteria format allows CAC to accredit additional program areas, such as information technology. In fact, program criteria for information technology also have been proposed. This session will discuss the new criteria structure, including the proposed general and program criteria, changes to existing criteria, and the timetable for implementation of the new structure.

No abstract available.

The all-du-paths data flow testing criterion was designed to be more demanding than the all-uses criterion, which itself was designed to be more demanding than the all-edges criterion. However, formal comparison metrics developed within the testing community have failed to validate these relationships, without requiring restrictive or undecidable assumptions regarding the universe of programs to which the criteria apply. We show that the formal relationships among these criteria can be made consistent with their intended relative strengths, without making restrictive or undecidable assumptions.

Software engineers often espouse the importance of using abstraction and encapsulation in developing software components. They advocate the "layering" of new components on top of existing components, using only information about the functionality and interfaces provided by the existing components. This layering approach is in contrast to a "direct implementation" of new components, utilizing unencapsulated access to the representation data structures and code present in the existing components. By increasing the reuse of existing components, the layering approach intuitively should result in reduced development costs, and in increased quality for the new components. However, there is no empirical evidence that indicates whether the layering approach improves developer productivity or component quality. We discuss three controlled experiments designed to gather such empirical evidence. The results support the contention that layering significantly reduces the effort required to build new components. Furthermore, the quality of the components, in terms of the number of defects introduced during their development, is at least as good using the layered approach. Experiments such as these illustrate a number of interesting and important issues in statistical analysis. We discuss these issues because, in our experience, they are not well known to software engineers.>

No abstract available.

A software test data adequacy criterion is a means for determining whether a test set is sufficient, or adequate, for testing a given program. A set of properties that useful adequacy criteria should satisfy have been previously proposed (E. Weyuker, 1986; 1988). The authors identify some additional properties of useful adequacy criteria that are appropriate under certain realistic models of testing. They discuss modifications to the formal definitions of certain popular adequacy criteria to make the criteria consistent with these additional properties.>

Test data adequacy criteria are standards that can be applied to decide if enough testing has been performed. Previous research in software testing has suggested 11 fundamental properties which reasonable criteria should satisfy if the criteria make use of the structure of the program being tested. It is shown that there are several dependencies among the 11 properties making them questionable as a set of fundamental properties, and that the statements of the properties can be generalized so that they can be appropriately analyzed with respect to criteria that do not necessarily make use of the program's structure. An analysis that shows the relationships among the properties with respect to different classes of criteria which utilize the program structure and the specification in different ways is discussed. It is shown how the properties differ under the two models in order to maintain consistency that the dependencies are largely a result of five very weak existential properties, and that by modifying three of the properties, these weaknesses can be eliminated. The result is a reduced set of seven properties, each of which is strong from a mathematical perspective.>

Cloze tests (i.e. fill-in-missing-parts tests) have been a long-standing measure of prose comprehension. Through human-subject experimentation, evidence was gathered to support the practical advantages of using the cloze procedure for measuring software comprehension. Cloze tests were found to be easy to construct, administer, and score and to be capable of discriminating between programs of varying comprehensibility. However, discrepancies between multiple-choice comprehension quiz results and some cloze test results for the same software suggested that certain forms of software cloze tests may not be valid. A model of software cloze tests was developed to identify a software cloze test characteristic that may produce invalid results. The test characteristic was concerned with the relative proportion of `program-dependent' and `program-independent' cloze items within a test. The developed model was shown to be consistent with software cloze test results of another researcher and led to suggestions for improving software cloze testing.

A Consortium of eight Central Ohio colleges and universities is described. The purpose of the Consortium is to provide opportunities for faculty at the participating institutions to be retrained in the field of computer science. These faculty will then be able to return to their home institutions to develop and teach computer science curricula. The program provides flexibility of scheduling the retraining, in terms of the time of year and nature of the retraining undertaken by the individual participants.

In attempting to describe the quality of computer software, one of the more frequently mentioned measurable attributes is complexity of the flow of control. During the past several years, there have been many attempts to quantify this aspect of computer programs, approaching the problem from such diverse points of view as graph theory and software science. Most notable measures in these areas are McCabe's cyclomatic complexity and Halstead's software effort. More recently, Woodward et al. proposed a complexity measure based on the number of crossings, or "knots," of arcs in a linearization of the flowgraph.

In attempting to describe the quality of computer software, one of the more frequently mentioned measurable attributes is complexity of the flow of control. During the past several years, there have been many attempts to quantify this aspect of computer programs, approaching the problem from such diverse points of view as graph theory and software science. Most notable measures in these areas are McCabe's cyclomatic complexity and Halstead's software effort. More recently, Woodward et al. proposed a complexity measure based on the number of crossings, or knots, of arcs in a linearization of the flowgraph.

Previous work in developing metrics of computer software using the methodology of Software Science has concentrated on relatively few program ming languages. More specifically, the experimental work has centered on programs written in Fortran, PL/I and Algol. While the results to date have provided the encouragement of further research in this direction, the apparent generality of the results mandates a study of other languages, par ticularly those whith substantially different characteristics and/or application areas than the similar, scientifically oriented languages of the previous work. This study concentrates on two such languages, namely COBOL and APL. The COBOL study contains programs obtained from an experiment given in an undergraduate pro gramming course, involving the relationship of Software Science metrics with programming time. The APL study contains twenty-nine programs obtained from several programming textbooks. Counting strategies for each language are described, main taining as much consistency with earlier Software Science work as possible. The results generally indicate continued high correlation among observed and predicted quantities. However, some problems remain in applying Software Science formulas to APL and COBOL, and their observed effects on our statis tical measures are discussed.

An investigation is made into the extent to which relationships from software science are useful in analyzing programming methodology principles that are concerned with modularity. Using previously published data from over 500 programs, it is shown that the software science effort measure provides quantitative answers to questions concerning the conditions under which modularization is beneficial. Among the issues discussed are the reduction of similar code sequences by temporary variable and subprogram defmition, and the use of global variables. Using data flow analysis, environmental considerations which affect the applicability of alternative modularity techniques are also discussed.

During the past few years, several investigators have noted definite patterns in the distribution of operators in computer programs. Their proposed models have provided explanations for other observed software phenomena and have suggested possible relationships between programming languages and natural languages. However, these models contain notable deficiencies.

A theory of the structural composition of an algoithm is presented which allows the frequencies of occurrence of the individual operators and operands to be estimated. It provides justification for some recent hypotheses which suggest certain functional relationships between properties of algorithms.