Yazılım yöntemi ile kalitesini değerlendimek için kullanılan ölçümlerin (metrics) fonksiyonel yöntemler için incelenmesi ve bir bankacılık organizasyonunda seçilenlerin değerlendirilmesi

ABSTRACT Defining software quality which has many characteristics each with different significance depending on the application and customer needs is a complex process that has been subject to many researchers for many years. Many quality factor sets defining various characteristics expected to be in a high quality software have been developed to evaluate software quality. Many important organisations like HP and IBM have developped their own quality factor sets. Factor sets convey our expectations from software but do not express whether those were met. Numerical expression of software quality factors is possible by defining metrics which measures quantitative characteristics of software. Although these expressions are not valuable if applied only to one program or process, when applied to more programs they form an historical database and enable comparisons and evaluations. Organisations must establish a metric model to control quality of programs and software development process and a data collection mechanism to collect data concerning development process and programs to be used in metric model. As an Information Systems Department of a bank aiming to take part in the market with its high technology and computer aided service, we must form a base for a metric model and data collection process which are main requirements for high quality software and software development process. That's why, Lines of Code (LOC), Function Point Analysis and COCOMO and Putnam' s estimation models which all have great importance in project planning, time and cost estimates and determining software quality are studied on the data of a small project. These analysis and model are based to form a simple metrics model. This metrics model is not a static model. It will improve itself according to the organisational structure and requirements depending on the data collected. Placing the data collection process and various estimation models into software development process will minimize the cost of collecting data. Additional information that would be obtained from collected data on project planning and management will meet this cost. Data collection mechanism for this initial metric model is presented with some information forms that computer support may be available. 11