ognizant Communication Corporation

FAILURE & LESSONS LEARNED IN INFORMATION AND TECHNOLOGY MANAGEMENT

ABSTRACTS
VOLUME 2, NUMBER 2, 1998

Failure & Lessons Learned in Information Technology Management, Vol. 2, pp. 47-57, 1998
1088-128X/98 $10.00 + .00
Copyright © 1998 Cognizant Comm. Corp.
Printed in the USA. All rights reserved.

Developing an Integrated Curriculum in Science for Higher Education

Richard M. Roberds

Integrated Science and Technology Program, Program Director, James Madison University, Harrisonburg, VA 22807
In the early 1990s James Madison University embarked on the creation of a new college of science. All aspects of the new college were to be constructed from basic beginnings: a provost would be hired to design and lead the development plan, all buildings were to be designed and built as needed, faculty would be hired and student recruitment initiated. The teaching approaches would be consistent with the national calls for teaching reform in science and engineering, and the curriculum would be specifically designed to meet the needs of the 21st century. The college was named the College of Integrated Science and Technology, and the new, innovative program that would be the college's flagship program was called the Integrated Science and Technology Program (ISAT). It was decided early that information and knowledge management would be a unifying glue for the curriculum, which would be aimed at educating science generalists. Because the acquisition of information and creation of knowledge pervades the practice of science and the creation of technology, information and knowledge management skills and tools would be at the heart of the program. After all, information and knowledge, and the digital technologies that manipulate them, are crucial factors in the economy and should be central to the ISAT program. This article focuses on the creation of the curriculum of this new program that was largely done on an ad hoc basis, but with a highly motivated group of interdisciplinary faculty. The process for developing the curriculum was defined, often by hit-and-miss techniques, the goals and objectives were articulated, and the curriculum and courses were produced. But the development process emerged as the basic key to effectively creating the curriculum. It was a process that called for iteration and correction of faults as the program unfolded. Finally, an unexpected by-product was the spontaneous creation of a unique culture among the faculty, staff, and students. This turned out to be a sustaining force and an essential ingredient for success for the enterprise.

Key words: Curriculum development; Information and knowledge management; Integrated curriculum; Integrated science; Science; Teaching reform; Technology

Correspondence and requests for reprints should be addressed to Richard M. Roberds. Tel: (540) 568-2740; Fax: (540) 568-2768; E-mail: roberdrm@jmu.edu




Failure & Lessons Learned in Information Technology Management, Vol. 2, pp. 59-65, 1998
1088-128X/98 $10.00 + .00
Copyright © 1998 Cognizant Comm. Corp.
Printed in the USA. All rights reserved.

Software Cost Models in the Department of Defense Environment: The Accuracy Enigma

Daniel V. Ferens and David S. Christensen

Air Force Research Laboratory, 2241 Avionics Circle, N3F22, Wright-Patterson AFB, OH 45433

The United States Department of Defense (DOD) is challenged with estimating software costs for a wide variety of software programs developed by various contractors. One of the challenges is development cost accuracy. Even after the software cost models are calibrated to DOD databases, many have been shown to be accurate to within only 25% of actual cost or schedule about half the time. In 1995 and 1996, seven popular software cost models were calibrated to a large Air Force database. Results showed that for six of the seven models, calibration did not significantly improve the accuracy from that reported previously. The seventh model, however, did demonstrate superior development effort accuracy for a variety of programs. Although there are some limitations, the model may represent a breakthrough in software cost estimation in the DOD environment.

Key words: Software cost estimation; US Department of Defense; US Air Force; Software engineering

Correspondence and requests for reprints should be addressed to Daniel V. Ferens. Tel: (937) 255-4429, ext. 3558; Fax: (937) 255-4511; E-mail: ferensdv@aa.wpafb.mil




Failure & Lessons Learned in Information Technology Management, Vol. 2, pp. 67-81, 1998
1088-128X/98 $10.00 + .00
Copyright © 1998 Cognizant Comm. Corp.
Printed in the USA. All rights reserved.

Applying COCOMO as a Cost Model for Knowledge-Based Systems

T. M. Shaft,1 R. F. Gamble,2 and R. Keshav2

1MIS/QM Department and 2Department of Mathematical and Computer Sciences, University of Tulsa, 600 South College Avenue, Tulsa, OK 74104

The COCOMO approach to software cost modeling has been improved and refined for 15 years. It is currently being updated to encompass newer approaches to software development (e.g., rapid application development, reuse-driven approaches, object-oriented), though the focus remains on conventional (i.e., algorithm-based) systems. In this article we propose a cost estimation technique for Knowledge-Based Systems (KBSs) based on COCOMO 2.0. Our definitions were developed for a broad range of KBSs. Specifically, our definitions consider rule-based systems, considered the most prevalent form of KBSs, as well as KBSs based upon object-oriented or hybrid knowledge representations. Our definitions provide a starting point for the application of the COCOMO approach to cost modeling to KBSs. Further research will be needed to develop empirically based multipliers to complete the cost estimate and test the effectiveness of this approach.

Key words: Knowledge-based systems; Cost modeling; Complexity

Correspondence and requests for reprints should be addressed to R. F. Gamble. Tel: (918) 631-2988; Fax: (918) 631-3077; E-mail: gamble@utulsa.edu




Failure & Lessons Learned in Information Technology Management, Vol. 2, pp. 83-90, 1998
1088-128X/98 $10.00 + .00
Copyright © 1998 Cognizant Comm. Corp.
Printed in the USA. All rights reserved.

Halstead's Metrics Revisited for Knowledge-Based Systems

H. R. Moreno and R. T. Plant

Department of Computer Information Systems, University of Miami, Coral Gables, FL 33124

As organizations become more virtual in nature and utilize knowledge-based technologies to support this mode of operation, the requirement to create robust industrial-strength systems becomes vital. The project managers and systems developers require tools and metrics to ensure effective planning and monitoring of the systems development. The article focuses upon Halstead's software science effort metric. The article shows the weakness and inapplicability of this method when applied to knowledge-based systems development and indicates the potential hazards associated with the incorrect selection of metrics.

Key words: Software metrics; Knowledge-based systems; Software estimation

Correspondence and requests for reprints should be addressed to R. T. Plant. Tel: (305) 284-1963; Fax: (305) 284-5161; E-mail: rplant@exchange.sba.miami.edu