文档介绍：Generative ProgrammingPrinciples and Techniques of Software Engineering Basedon Automated Configuration and Fragment-ponent ModelsbyDipl.-Inf. Krzysztof CzarneckiA dissertation submitted in partial fulfillment of the requirements for the degree ofDoktor-IngenieurDepartment puter Science and AutomationTechnical University of IlmenauOctober mittee:Universitair Hoofd Docent Dr. Mehmet Aksit,University of Twente, herlandsProf. Dr. Ulrich W. Eisenecker,FH Heidelberg, GermanyProf. Dr.-Ing. habil. Dietrich Reschke (Chairperson of mittee),TU Ilmenau, GermanyOral defense held on August 9, 1999iiiAbstractCurrent object-oriented (OO) ponent technologies suffer from several problems such asthe lack of analysis and design methods for the development for reuse, lack of effectivetechniques for dealing with many variants ponents, loss of design knowledge due to thesemantic gap between domain abstractions and programming language features, and runtimeperformance penalties for clean and flexible thesis proposes Generative Programming (GP) as prehensive software developmentparadigm to achieving high intentionality, reusability, and adaptability without the need promise the runtime performance puting resources of the produced the area of analysis and design for GP, we investigate Domain Engineering (DE) methods andtheir integration with OO analysis and design (OOA/D) methods. The main difference betweenDE methods and OOA/D methods is that the first are geared towards developing whole familiesof systems while the latter focus on developing single identify feature modeling as the main contribution of DE to OOA/D. Feature modelsrepresent the configurability aspect of reusable software at an abstract level, . mitting to any particular implementation technique such as inheritance, aggregation, orparameterized classes. We give a precise and extended formulation of the feature diagramnotation and investigate the relationship between feature modeling, OO m