A uniform approach to software patterns classification and software composition

Design Patterns are software problem solving techniques that package both a problem specification and its solution. Experienced software designers rely on their prior experience to identify similar problems and apply common design solutions. Since their introduction, hundreds of new design patterns have been discovered and documented to address a variety of design problems we encounter during construction. However, to use any one of those design patterns effectively and quickly, one must have an access to a design pattern description that unambiguously gives away its intended purpose, and find an easy way to choose one amongst related patterns. Existing research is either lacking or has not addressed these issues adequately to the degree that there are conflicting views amongst designers as to the definitions of some patterns. This thesis addresses this deficiency by proposing a software patterns classification model enabling us to develop uniform pattern definitions. We will propose a new process by which we can classify design patterns and allow for easy identification and differentiation between them.; Another different but related problem to the use of design patterns as a software construction technique is this: given that software requirement problems are identified, and a set of proper design patterns are chosen to solve them, how can we integrate the solution of each sub-problem in an independent and a repeatable way that comprises the entire software application? Not much attention has been paid to address patterns-based integration problem. This is a problem that both students and industry practitioners struggle with a great deal. We will propose a new technique in which we use the design patterns themselves as a software composition technique.