COMPUTER AIDED SOFTWARE ENGINEERING VIA CONSTRAINT PROPAGATION

THE OBJECTIVE OF THIS PROPOSED RESEARCH IS TO SHOW THAT COMMERCIALLY AVAILABLE RELATIONAL DATABASE MANAGEMENT SYSTEMS (RDBMS) WOULD BE BETTER SUITED AS A PLATFORM FOR INTEGRATED, FULL-PHASE COMPUTER-AIDED SOFTWARE ENGINEERING (CASE) PRODUCTS IF THEY PROVIDED MORE SOPHISTICATED DATA INTEGRITY CHECKING. MOST CASE PRODUCT DEVELOPMENT INCLUDES ITS OWN DATA MANAGEMENT; THOSE FEW PRODUCTS WHICH DO UTILIZEA STANDARD RDBMS TEND TO BE OF SMALLER AND MORE LIMITED SCOPE. USING A STANDARD RDBMS AND AN OPEN ARCHITECTURE WOULD ALLOW CASE TOOLS TO BE DEVELOPED BY PERSONS WITH BROADER RANGES OF EXPERTISE AND EXPERIENCE, WOULD FACILITATENEW RDBMS ADVANCES TO BE MADE AVAILABLE IMMEDIATELY TO THE CASE PRODUCT WITHOUT ADDED EFFORT OR COST, AND WOULD ENCOURAGE THE CENTRAL CASE DATA TO BE INTEGRATED WITH THE REMAINDER OF THE CORPORATE INFORMATION. THE OBJECTIVES OF THE PHASE I RESEARCH IS TWO-FOLD: (I) TO SHOW THAT ALL THE FUNCTIONALITY THAT IS AVAILABLE IN CURRENTPRODUCTS CAN BE REPLICATED IN A RDBMS WITH AN EXTENDED CONSTRAINT MANAGEMENT SUBSYSTEM, AND (II) TO SIMULATE THAT SUBSYSTEM USING A COMMERCIAL RDBMS. A SIDE-EFFECT OF THESE OBJECTIVES WILL BE TO DEFINE THE MINIMUM SET OF INTEGRITY CONSTRAINT FUNCTIONS WHICH MUST BE IMPLEMENTED. THE GOAL OF A PHASE II PROJECT WOULD BE TO DESIGN AND IMPLEMENT AN EFFICIENT INTEGRITY CONSTRAINT SUBSYSTEM INTEGRATED INTO THAT RDBMS.