By the end of the next decade, computers will not be just information gatherers who return reams of keyword-related information. Instead computers will become mobile intelligent helpers who will truly understand entire problems and provide just the right information to solve them. What capabilities will be needed to realize such advanced systems? This paper projects technology advances and assesses the impact such systems might have on future decision-makers. The scenarios and ideas presented by the panelists will provide a reasonable starting point for discussion and solicitation of alternative views.

A panel session organized to project what changes might occur in the near future to make parallel computers easier to program and use and to explore how such computers could benefit many application areas is reported. The following questions are discussed: (1) what type of applications will benefit on a widespread basis from improved performance; (1) whether serial and parallel programming will be integrated for greater reusability and portability; and (3) whether parallel computers will replace serial computers and, if so, what is needed so that concurrency can be handled more easily.>

Issues that need to be resolved to facilitate advances in software development for parallel architectures are examined. These include the establishment of a proper software development paradigm, visualization and animation techniques, language issues, problem decomposition, algorithm design and evaluation, techniques for software design partitioning and allocation, software testing, software reusability, and fault-tolerance models.>

Moving toward high confidence software that can meet ever increasing demands for critical DOD applications will require planning, specifying, selecting, and managing the necessary development and testing activities that will ensure the success of the software project. In order to trust the decisions being made, there must be evidence (i.e., an information base of data and facts) that techniques and tools being chosen for application on critical projects will perform as expected. Today, these expectations are mostly intuitive; there is little hard evidence available to guide acquisition managers and software developers in making necessary decisions.