Toward Application Driven Software Technology.
Teodor Rus, Donald Curtis · 2007
Abstract Current problem solving methodology by computers requires application domain expert to develop programs in a programming language. Software tools designed to ease this task and the expansion of computer use in practically all aspects of human life lead to an increasing software complexity. In this paper we present a solution to the problems resulting from increasing software complexity by developing a methodology for problem solving with the computer where problem solving process is split between application domain expert and software expert as follows: (1) AD expert develops solution algorithms using the natural language of the application domain; (2) IT expert implements universal algorithms that characterize the application domain; (3) the AD is provided with an interpreter that implements AD algorithms in terms of the processes implementing the components of the AD algorithms. This is realized by Computational Emancipation of the Application Domain (CEAD). Here we illustrate these ideas using linear algebra as application domain.