Managing Ada development risk in a non-Ada-based workstation environment

• Authors: Kourosh R. Dinyari;Thomas M. Johndrew;Kenneth J. Lamarche
• Venue: TRI-Ada '90 Proceedings of the conference on TRI-ADA '90
• Year: 1990

There is increasing demand for systems with user friendly interfaces to complex automation. Workstations are especially well-suited for these complex systems because interactive graphics coupled with multiprocessing capability is offered in one small platform. Unfortunately, the current available graphics workstations are essentially a hostile environment for Ada developers. However, a successful Ada implementation within project cost and schedule constraints can be achieved when the problems that cause the environment to be hostile are identified, understood, and managed. We have identified the major problems associated with Ada development in a non-Ada based workstation environment through our experience gained on three Ada implementations. These major problems relate to vendor support of Ada, not problems associated with the Ada language itself. We also identified portability problems of Ada code developed in a non-Ada based environment. Portability of Ada code between platforms of the same vendor, as well as across different vendor platforms is addressed. Because Ada developers are forced to generate Ada code that interfaces with system dependent features of non-Ada based environments, implementation risk is high unless mitigated through a risk management process. Accordingly, we present a set of criteria to be used in the platform selection process. Application of these criteria exposes hidden problems of each candidate platform, and leads to selection of the platform with the least risk. We supplement these selection criteria with development guidelines for production of more portable Ada code regardless of which platform is chosen.