From Scientific Software Libraries to Problem-Solving Environments

  • Authors:

  • John R. Rice;Ronald F. Boisvert

  • Affiliations:

  • -;-

  • Venue:
  • IEEE Computational Science & Engineering
  • Year:
  • 1996

Quantified Score

Hi-index 0.00

Visualization

Abstract

More and more science and engineering is done with computers, but that doesn't mean the people doing it want to learn more and more computer science. Software libraries of mathematical and other routines have helped programmers avoid redundant effort, but practitioners in various application areas will increasingly want true comprehensive problem-solving environments to help them do their work more easily. As this report shows, a lot of computer scientists will keep very busy building these systems for "user-friendly" scientific computing. Software libraries provide encapsulated problem-solving power and problem-solving environments give ordinary users painless access to problem-solving power. Thus the structure of libraries and the design of PSEs are inextricably linked. Experts in these subject areas from academia, government, and industry met at Purdue University on September 25-27, 1995, to explore the state of the art in scientific libraries and PSEs, their interdependence, and their combined future. The workshop, entitled "Scalable Scientific Software Libraries and Problem-Solving Environments," built partly upon an NSF workshop held in 1991. The recent workshop reviewed several years of progress and assessed future steps. For its examples, it focused on applications based on partial differential equations and closely related areas. Afterwards, participants formed four groups, each of which developed a separate white paper. This article summarizes the ideas expressed in the white papers. Seven roadblocks, seven research directions, and six immediate recommendations are described.