This dissertation investigates data integrity in Ada 95. The hypothesis presented is that Ada 95 programs must be designed under the control of the programmer in order for data entities to be protected from internal corruption. The designers of the language made a conscious decision to add object-oriented features by extending the existing definitions of Ada 83. While the new implementation provides the object-oriented features of inheritance and polymorphism, the language implementation of these features, along with that of hierarchical libraries, introduce an ambiguity of object-oriented design-to language constructs that can result in data integrity problems. Coding techniques are presented for various program design dilemmas and emphasize the protection and consistent use of data entities within and between the various components of an Ada 95 program. The results of testing the coding techniques indicate that different encapsulation organizations have different impacts on the various aspects of data integrity. During this testing, a flaw in the compiler was revealed with respect to the inheritance of private primitive operations. Through investigating the language from the perspective of data integrity, it was found that the object-oriented paradigm, as well as the protection of critical data entities as dictated by the problem domain, can be achieved through a combination of Ada 95 features. This is fully dependent on an increased intervention and control of the program code by the programmer.
| Date of Award | Jan 1 1996 |
|---|
| Original language | English |
|---|
| Supervisor | S. Rollins Guild (Supervisor), Marlyn Kemper Littman (Advisor) & Michael Moody (Advisor) |
|---|