Annual Review ofAutondc Progrtmning Vol. 16, pp. 161-174 1992 Printed zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA in Great Britain. All rights reserved 00664138/92/s15.00 0 1992 PergamonPress Lid EFFECTS OF CASE IMPLEMENTATION ON THE PRODUCTIVITY AND MATURITY LEVEL OF SOFTWARE DEVELOPMENT - A CASE STUDY IN A LARGE INDUSTRIAL COMPANY P. Kuvaja Department of Information Processing Science, University of Oh, Linnanmaa, SF-90570 Oulu, Finland zyxwvutsrqponmlkjihgfedcb ABSTRACT. Technological innovations have brought about substantially changes in the productivity and quality of the IS/SW development process during the last twarty years. Perhaps the most frequently refereed innovation today is Computer-Aided Software E+eerkg (CASE), which is highly recommended for incmasii the productivity and quality of the process snd quality of its products. This case study reparts empirical results of a CASE implementation in a large industrial company in Fiisnd with regard to productivity and quality of IS/SW development. The measurements include comparisons between development projects employing traditional development with Cobol. prototyping with the UFO application generator and integrated CASE tool environment IEW/ADW. 1 Introduction Traditional IS/SW development by hand has for a long time been felt to be too labour-intensive to be economical. When the expenses of a manually developed IS/SW sixed as 1000 lines of Cobol source code are equivalent in value to an automobile, 10000 lines of Cob01source code equal a private house and 1OoOO lines of Cob01 source code cost as much as a small offtce building (/32/). there are extremely powerful motives for improving the situation. The development of new tools, techniques, methods and methodologies for IS/SW development and attempts to apply these are recent remedies for the above situation. These technological innovations, when transferred into practice, are expected to being about substantially changes in the productivity and quality of both the product and process of IS/SW development. Application generators and prototyping (/l l/) form the next step in the chain of technological innovations after traditional IS/SW development with third-generation languages (as Cobol). Application generators were promoted enthustically on the markets in the mid-eighties and have now reached a mature state as IS/SW development tools, and now the second step in the chain of innovations is Computer-Aided Software Engineering or CASE. CASE tools of today already belong to the second generation of CASE tools, but even so, these tools and their potential effects on IS/SW development are a topic of intensive discussions among both practitionersand researchers. The productivity improvement effects of an IS/SW development technology are not self-evident, however, and this has motivated a number of empirical research efforts to study the productivity of various IS/SW development tools. The reported research results can be classified into three categork comparisons between programming languages (/la/, /l/ and /45/), investigations in data base environments (/47/ and /19/) and comparisons between the third and fourth-generation languages (/23/, /51/, /35/, /52/, /13/, /21/, /43/, /40/ and /37/).There is also a fourth category, which is still virtually empty, formed by productivity comparisons between CASE and third and fourth generation tools. To our knowledge, there is only one research effort that belongs to this category, the recent empirical case study by Low and Jeffery c/38/). in which they compared traditional IS/SW development with backend CASE tools, concluding that productivity gains can be elusive and that it may well be that quality or other characteristics of the software process or product gain more from the introduction of CASE technology. They also recommend further empirical research to assess the gains potentially achieved through the introduction of CASE tools. More general fmdings were reported by Herbert and Brody (/22/) in their survey among U.S. companies, in which they indicate that the expected benefits of CASE tools are higher quality, easier maintenance, faster development and lower costs, in this order. They also calculate that the main productivity factor from the companies’ point of view is the quality of the software systems produced, which in most cases is dependent on the quality of the software process. On the other hand, the complexity of the software code has been recognised zyxwvutsrqp 161