Engineers and scientists rely heavily on the use of scientific computation to aid in their design decisions. The decisions that are made have far reaching consequences in terms of financial resources, environmental impacts and possibly the health and welfare of the community. Given the importance of the decisions, it is necessary that the computational tools used by engineers perform efficiently and reliably. Some ideas for reaching these goals are as follows:
1) Application of Software Engineering Principles to Scientific Computation
– Requirements documentation for modelling of physical phenomena
– Module interface specification for numerical methods, such as the finite element method
– Use of the principles of information hiding and encapsulation in numerical software
– An open framework for finite element code
2) Improvement in the Application of Numerical Algorithms to Practical Engineering Problems
– Implementation of the finite element method using different descriptions of motion
– Investigation of new finite elements, such as the average strain element
– Investigation of computational problems in polymer processing and metal forming
