The project is currently being supported by research and development money from a National Institute of Standards and Technology (NIST) - Advanced Technology Project Grant, with the expectation that the product will be available in early 1997. If this mission is successful, Stanly claims we will all be doing 'Scientific Computing without Programming.' His main role, along with the computational physicist, is to decide what information about initial boundary-value problems for partial differential equations and methods for discretizing differential equations are to be included into the environment and then to record this information in a form that can be processed by the problem solving environment. The environment also needs rules and heuristics to guide it in the development of solution algorithms and programs, so he and the physicist are codifying as much of this information as they know and will soon be codifying information obtained from other consultants. These rules and heuristics concern questions such as which discretization method to use, linear equation solver to use, or preconditioner to use.
Stanly is currently on leave from the University of New Mexico, where he is a Professor of Mathematics. "Both the Department of Mathematics and the University believe that it is important for their faculty to have industrial experience and contacts," he says, "and thus have strongly supported this activity. I helped with the proposal writing and joined the company when it started. My plan is to work for the company nearly full time through the initial research and development phase and to continue consulting for the company indefinitely."
He has a B.S. and M.A. in Mathematics from Michigan State University, and a Ph.D. in Mathematics from Stanford University. After spending about seven years working on (and enjoying) pure analysis problems related to partial differential equations, he took the job at the University of New Mexico and decided to do research in more applied areas. "In fact," he says, "I made a commitment to spend some years working only on mathematics that clearly had important industrial applications. I studied and did research on computer algebra, analytic solutions of partial differential equations, finite-difference methods for partial differential equations, automatic code generation, and numerical grid generation." It was during this period he started consulting for a local company that developed numerical modeling codes for numerically solving partial differential equations.
In preparation for a job in industry, Stanly believes particular courses are not very important, but being exposed to certain ideas is critically important. His favorite saying is 'the only important tool in applied mathematics is abstraction.' "Of course this is a dreadful exaggeration used to make a point," he admits, "but, the other hand, abstraction is what allows us to clearly organize information, and it is just this skill that is critically important in the design and implementation of a problem solving environment. What is needed is to understand what problems are important, what is difficult about the problems, what information is needed to solve them, and how the solution will be used." He believes it is important to attend lots of colloquia and seminars in pure mathematics, applied mathematics, statistics, computer science, and some engineering discipline. "You will not understand at first, or maybe even later." he admits, "but you will become exposed to the language and mathematics used in other disciplines, and eventually you will understand much of what is going on. This will help prepare you to work with groups that have diverse interests." In addition, writing skills are critically important. He recommends students keep reading work by others, practicing writing up their work clearly, and take some writing courses.
You may find more information about Stanly on his homepage.
Return to Archived Profiles and Forums
