Follow these links to the web sites for EDSER-1 and EDSER-2.
The purpose of computation and software design (CSD), as for any design activity, is to create added value. Yet, current CSD theories, concepts, tools and methods are not clearly based on modern models of value and value creation. For example, past work on software economics has improved cost and risk modeling and management enormously, but there has not been a symmetric development of benefit and opportunity modeling and management. We also lack a theoretical understanding of the means by which core concepts of our field are linked to value creation: modularity in design (architecture), iterative development methods, testing, etc. Thus, the current conceptual state of the art in software design and engineering is not clearly optimal for value. To the extent that it is not, society is not getting its money’s worth for investments in CSD. The emerging area of strategic software design (SSD) takes the value-based view as a premise and seeks to develop and enhance fundamental CSD theories, tools and methods by basing them on mature models of value and value creation. The EDSER workshops provide a forum for exploring, influencing and advancing the development of strategic software design.
A principle of SSD us that no single scale of value can suffice as a foundation, but not all scales are equally important, either. Business value is one very important scale. Much of society’s investment in CSD is intended to create value on this scale. EDSER thus seeks new insights in CSD based on explicit connections between software design and engineering concepts and models and theories of business value creation. For example, recent work has shown that there are deep connections between modularity in design, real options (which are capital analogs of financial options), and task, contract and industry structures. Other disciplines that promise to support for a better grounded discipline of CSD for business value include utility theory, game theory, financial engineering (e.g., portfolio theory, securitization of risk), Lamarckian evolution, and design-oriented studies of contract and industry structures.
Of course, business value is not the only well-developed concept of value. Philanthropies, universities, militaries and other important institutions do not take market value as a metric. Philanthropies value progress against societal problems. Universities value creation, storage and dissemination of knowledge and culture. Militaries value protection of national security. Risk preferences influence the perceived values of risky assets. Individuals often have highly subjective scales of value. Some important value scales are embodied in mature concepts of ethics, aesthetics, justice, humanism, and morality. Software development can and does occur in contexts in which value scales such as these apply; and such scales have been and can be applied to inform the development of software design methods, in general, and methods for business, in particular.
An important observation is that there are some fundamental methods of reasoning about value that are applicable more or less independently of value scale. Multi-attribute utility theory is one example of such a strategic reasoning framework. Thus, many of the most world's most successful organizations take strategic approaches to resource allocation, whether they happen to be optimizing for business value or for other metrics. A key goal of SSD research is to adapt and leverage such ideas to catalyze significant improvements in the intellectual foundations and in the practice of software design and engineering.
Strategic software design is a new area. The EDSER workshops thus function best not as mini-conferences but as working sessions. EDSER seeks good ideas with some plausibility and some support, preliminary results, well thought out but provocative positions, and excellent introductions to and tutorials on relevant art (e.g., game theory, ethical theories, finance, etc.) This year's two-day workshop will include presentations of background material on technical topics (e.g., in the past workshops: real options and decision theory); presentations of work by participants selected on the basis of submitted positions; and extensive discussions. Papers of 2 to 5 pages in ICSE format, tutorials, and other proposal are solicited and strongly encouraged.
Due date for position papers: March 15, 2001.
Selection of presenters: No later than April 15, 2001.
Papers should be 2 to 5 pages, in PDF (preferred) or plain postscript, and formatted according to the
ICSE guidelines
Submit to Kevin Sullivan at Sullivan@Virginia.EDU.
Kevin Sullivan; University of Virginia, USA; sullivan@virginia.edu
(Chair)
Mary Shaw; Carnegie Mellon University, USA; mary.shaw@cs.cmu.edu
David Notkin; University of Washington, USA; notkin@cs.washington.edu
Barry Boehm; University of Southern California, USA; boehm@sunset.usc.edu
Warren Harrison, Portland State University, USA; warren@cs.pdx.edu
Kevin Sullivan
Associate Professor
University of Virginia Department of Computer Science
151 Engineer's Way P.O. Box 400740
Charlottesville, VA 22903 USA
sullivan@Virginia.EDU
Tel. +1 (804) 982-2200
FAX: +1 (804) 982-2214
