A key to solving large-scale problems in science and engineering is a unified software environment. The Institute for Complex Scientific Software will develop a new interdisciplinary field of technical research: building complex scientific software systems. The discovery of new software design principles, design patterns, and tools for software integration represents a new discipline. This has the dual benefit of carrying on important computer science research on scalability, and of providing enabling tools to solve grand challenge scale problems.

The current state of the art is to build a working system by cobbling together general purpose application programs with specialized tools to solve the problem at hand. Without proper design principles, such systems are neither extensible nor maintainable. Yet, good design principles do not exist for retrofitting existing systems with new functionality. This limits the class of problems that can be attacked by the research community and greatly impedes progress on a number of important problem domains.

As software size and complexity increase, the need for this new discipline grows yearly. Examples of common needs include addition of parallelism, the addition of a scripting language, component integration, distributed operation, performance monitoring, and an expert system interface. Other examples include the gluing together of multiple existing software packages, such as a front-end data generation package, a simulator, and a back-end analysis tool. Academic research in scalability of software has historically been limited. Academic departments have not had broad access to large industrial-scale applications.

As a proving ground for this new interdisciplinary area, the Institute will concentrate on software environments for the following large scale applications:

• Genomics and Proteomics,
• Subsurface Sensing, Imaging and Modeling. and
• Geant4 (GEometry ANd Tracking) for particle-matter interaction.

The foundation of this institute will be a rare confluence of talent available at Northeastern University. That talent is spread across many departments in the computing, scientific and engineering disciplines. But they have a history of working collaboratively. Many outside institutions are interested in collaboration and have provided letters of support for the Institute. The proposed new discipline currently has many opportunities for research activity that will significantly enhance the reputation of the University. These applications will also provide an important opportunity to integrate practice-oriented, interdisciplinary education into the curriculum.