An approach to a methodology which is easy to use and application domain independent

Abstract

This paper discusses the direction o f research to develop a system analysis and design methodology which is easy to use and application dom ain independent. In recent years of the inform ation systems field, several new m ethodologies for softw are analysis and design tasks have been developed. These m ethodologies can be generally classified into three categories: process-oriented, d a ta-o ri­ ented, and o b jec to rien te d. All o f these m eth­ odology types deal w ith how to decom pose a complex application dom ain in order to accom plish desirable levels of abstraction, m odularity, and inform ation hiding. The m ajor differen ce across the above categories lies in the kinds o f inform ation (i.e., process a n d /o r data) which each m ethodology tries to extract from the problem space. H ow ever, d espite o f the huge am ount o f effo rt put into the developm ent o f these m ethodologies, a question o f w hether these m ethodologies really help practitioners in system analysis and design tasks still rem ains. Most m ethodology research firs t identifies an ap plica­ tion dom ain then fits a m ethodology to c h a r­ acteristics o f the application. N evertheless, these works ignore that, fo r a m ethodology to be easily used, it should be constrained as little as possible by the characteristics o f an application dom ain. Based on this reasoning, an ideal goal is the developm ent o f a m ethodology w hich is ap plica­ ble and appropriate to any application dom ain. The overall fram ew ork for the developm ent o f an application dom ain independent m ethodology is iterative, com bining experim ents and theory-based reasoning. Each m ethodology is identified in term s o f its own strengths and weaknesses, and the strengths are m erged into a unified m ethodology. T he rationale for e x p e ri­ mental m ethod is that any currently available m ethodology should be thoroughly tested and understood. The basic experim ental procedure is: 1. Identify subjects with sim ilar experience with the knowledge of experim ental methodologies. 2. Subjects identify "difficult parts" of methods based on use. 3. R esearcher replaces d iffic u lt parts with analo­ gous activities …