PEDLS Nancy Leveson — Panel

Event Date: October 24, 2022
Speaker: Nancy Leveson, Jerome C. Hunsaker Professor in Aeronautics and Astronautics, Massachusetts Institute of Technology
Time: 1:45-2:45 PM EDT
Location: ARMS Atrium
Priority: No
School or Program: Electrical and Computer Engineering
College Calendar: Show
Conflict and complementation between agile methods and systems analysis in developing and maintaining software-intensive systems

Abstract

In this panel session, the PEDLS speaker, Dr. Nancy Leveson from MIT, will join with Purdue researchers in discussing the ways in which systems analysis and agile methods conflict with and complement each another in the development and maintenance of software-intensive systems. Failures in software-intensive systems are in the news every day, from faulty autonomous vehicles to ransomware infecting critical infrastructure. This panel will explore one facet of these failures: the interaction between systems analysis and agile methods in system development. Developing and maintaining holistic system properties, such as safety and security, traditionally involves a top-down system design process prior to system development. This approach allows for system-level analysis for properties of interest. However, this approach has been criticized for their inability to respond to changes in, for example, stakeholders’ requirements, market context, operating environment, or personnel. These concerns commonly appear in software development. To address these concerns, software engineering organizations began to adopt Agile methods beginning in the 1990s, and now over half of Fortune 500 companies use some form of Agile method. In Agile methods, teams evolve a complex system design iteratively through a series of staged enhancements. Agile methods allow teams to get feedback during the development process. However, they tend to eschew rigorous analysis because it is perceived as slowing down development. Agile methods can help teams get a product to market faster, but that product may not be as safe or secure as desired.  As software controls more and more of individual and societal interactions, from smart homes to smart cities, this trade-off becomes less compelling. The panel will discuss the ways in which findings from research can help balance the competing engineering requirements of robustness (improved through analysis) and cost (improved through agility). Particular attention will be paid to the safety and security of cyber-physical systems, whose failures have increasingly profound implications for society. 

Moderator

  • James Davis, Assistant Professor, Elmore Family School of Electrical and Computer Engineering

Panelists

  • Nancy Leveson, Jerome C. Hunsaker Professor in Aeronautics and Astronautics, Massachusetts Institute of Technology
  • Daniel DeLaurentis, Professor of Aeronautics and Astronautics
  • Ryan Newton, Associate Professor, Elmore Family School of Electrical and Computer Engineering and Computer Science
  • Dongyan Xu, Samuel D. Conte Professor of Computer Science