Glenn W. Cox

A comparison of cohesion metrics for object-oriented systems

Abstract Cohesion is the degree to which the elements of a class or object belong together. Many different object-oriented cohesion metrics have been developed; many of them are based on the notion of degree of similarity of methods. No consensus has yet arisen as to which of these metrics best measures cohesion; this is a problem for software developers since there are so many suggested metrics, it is difficult to make an informed choice. This research compares various cohesion metrics with ratings of two separate teams of experts over two software packages, to determine which of these metrics best match human-oriented views of cohesion. Additionally, the metrics are compared statistically, to determine which tend to measure the same kinds of cohesion. Differences in results for different object-oriented metrics tools are discussed.

Assessing design instability in iterative (agile) object‐oriented projects

Assessing progress and stability in rapid, highly iterative and agile object-oriented software development processes is difficult because program evolution appears somewhat disorderly. With agile software development methods, the expected time to reach software stability is obscured by the dynamic nature of the agile process itself. The System Design Instability (SDI) metric was developed to help analyze object-oriented software evolution under these conditions. In this paper we examine a new version of SDI, called SDIe, based on the maximum entropy in a system. We believe our new measure is easier to calculate than the original version of SDI since it is based on data that are more commonly available than those used by SDI; also, SDIe gives a more accurate indication of software stability and maturity since it suffers less from data spikes. To validate the new metric, we use maintenance data from a commercial software project, developed using an agile process; these are the same data that were collected and analyzed in the original SDI study. In addition, we apply the refined SDI metric to the evolution of a well-established, open-source, iteratively-developed software product. The results of our case study indicate our new SDIe metric is a viable measure of system design stability. Copyright

COHESION METRIC FOR OBJECT-ORIENTED SYSTEMS BASED ON SEMANTIC CLOSENESS FROM DISAMBIGUITY

Object-oriented semantic metrics address software quality by assessing underlying code meaning. Previous metrics were based on mapping a classs semantic information onto concepts in an application domain knowledge base. Quality measurements were made by operating on the concepts mapped onto. In this work, we consider more complex inter-concept relationships—semantic disambiguities through semantic connections. The idea is that a level of ambiguity is indicated by the connectivity within the knowledge base between two concepts. A cohesion metric based on this idea is shown to perform as well as traditional metrics, and is available much earlier in the development cycle.

A fault-tolerant approach to test control utilizing dual-redundant processors

A simple dual-redundant fault-tolerant test control system architecture has been designed, developed, and demonstrated in a real-time environment using inexpensive personal computers. A survey of existing fault-tolerant control systems was performed to assess the relative cost and capabilities of currently available technology. A cost-benefit analysis was performed comparing the relative benefit of this system to triplex systems and non-fault-tolerant systems for various applications. Functionally identical implementations of a prototype proof-of-concept software design were constructed in two different languages and tested using a unit-under-test model. Bugs (faults) were injected into this model to verify the ability of the system to reliably detect anomalous test hardware operation. Also, simulated bugs (faults) were introduced to verify smooth control transfer between primary and standby, both nominally and in the presence of hardware-under-tests anomalies. Results indicate significant improvement in system reliability, sufficient to justify the additional cost of the proposed duplex system for many potential users.

SEMANTIC METRICS: METRICS BASED ON SEMANTIC ASPECTS OF SOFTWARE

Software practitioners need ways to assess their software, and metrics can provide an automated way to do that, providing valuable feedback with little effort earlier than the testing phase. Semantic metrics were proposed to quantify aspects of software quality based on the meaning of softwares task in the domain. Unlike traditional software metrics, semantic metrics do not rely on code syntax. Instead, semantic metrics are calculated from domain information, using the knowledge base of a program understanding system. Because semantic metrics do not rely on code syntax, they can be calculated before code is fully implemented. This article evaluates the semantic metrics theoretically and empirically. We find that the semantic metrics compare well to existing metrics and show promise as early indicators of software quality.

A conceptual model for data management and distribution in peer-to-peer systems

While Peer-to-Peer (P2P) model gains significant attention in distributed computing, it is also expected to be a powerful model for information sharing. P2P systems are expected to provide exhaustive reliable computational resources and scalable accessibility. The data management and distribution in such systems requires storage, replication, data modeling, indexing, querying, retrieval, streaming, and topology management. While a lot of data management strategies have been proposed through the last years, these strategies have not been investigated with respect to a common model for P2P systems. However, since the services provided by the P2P systems are so diverse, it is very challenging to come up with a common layer-based model for all P2P systems. In this paper, we firstly propose a conceptual model for P2P systems, and then provide a classification and summary of data management and distribution strategies by referring to this model. The horizontal layers of the model correspond to modules of a P2P system whereas the columns are related to the services provided. The modules include base P2P service, storage, indexing, logical, service, and application modules. The services include security, querying, publish, join/leave, collaboration, and streaming. The paper concludes by providing a comprehensive list of data management and distribution strategies used in the existing P2P systems.

Predicting Computer Science Ph.D. Completion: A Case Study

This paper presents the results of an analysis of indicators that can be used to predict whether a student will succeed in a Computer Science Ph.D. program. The analysis was conducted by studying the records of 75 students who have been in the Computer Science Ph.D. program of the University of Alabama in Huntsville. Seventy-seven variables were extracted from each students record, and the variables were correlated with whether the student did or did not successfully graduate from the program. A multivariate model was developed that predicts success with a high degree of accuracy. Importantly, the model relies on variables that can be determined reasonably early in a students Ph.D. class work, enabling its use as a selection metric. Hypotheses about the composition of the model are also presented and discussed.