Two frameworks for task analysis: an application to the use of intelligent systems in the development of information systems

The productivity paradox of information technology has said to be one of the most difficult problem of computing field. We must try to improve productivity of both computing applications and software industry. It is therefore vital to identify and differentiate productive efforts from non-productive ones. In this paper we develop two frameworks for analysis of tasks in the development of information systems. When we automate part of the development, we at the same moment distribute tasks between man and computer. We shall show that this distribution will create some non-productive additional subtasks: Transportation, communication, inspection, coordination, coding/encoding and conversion. Our second framework consists of the initialisation and termination subtasks surrounded every process proper in a discontinuous, non-repetitive task. By making those implicit subtasks explicit and demonstrating their existence both in generative and building blocks approaches we want to increase our awareness, to improve productivity and to move consideration and management of those subtasks under conscious control.