PO WEEKNAPPLICATION DEVELOPMENI APPLIED INTELLIGENCE | Pace 64 DECEMBER 18, 1989 RAD Designed To Facilitate Dynamic Change in Firms This-is the first of a series of articles on rapid applica- tions develop- ment (RAD), a methodology for development de- signed to be much faster than the tradi- tional life cycle. One of the ur- gent concerns in enterprises today i ; is the need for information-systems (IS) organizations to create and modify applications faster than is possible with the traditional de- velopment: life cycle. If applications take two or three years to build and the application back- log spans several years, businesses can- not: create and react to competitive thrusts guickly enough. The vital ability for dynamic change is lost. The Need for Speed The pace of business is increasing. To stay competitive, corporations are short: ening their manufacturing life cycles. They are under pressure to design new goods guickly and shorten the time re- auired to fill orders. Factories are striv- ing to reduce their inventories with just in time" inventory control. In many situations, the time available to make decisions is decreasing. To a large extent, this increase in the business metabolism is caused by the use of networks, immediate access to data and powerful decision-support tools on the desktop. Corporations are looking increasingly for ways to beat their competitors to market with new goods or services. American Airlines, for example, in- creased its market share by building Sa- bre, a complex system to put on-line ter- minals in travel agents' offices. TWA re- cently rebuilt its freguent-flyer program using a high-end computer-aided soft- ware engineering (CASE) tool and can now respond much faster to moves by competitors. United Stationers grew much faster than its competitors by guaranteeing 24- hour delivery of orders placed with ter- minals. Canon devised computerized technigues for designing innovative lenses and cameras faster than compet- ing companies. In the 19905, business will become a battle of high technology, often of com- peting computer systems. In such a - world, the corporation that can create the necessary computer applications fastest can win. If a company takes three years to build and debug a com- ana esem while its competitors mon! prergiea ths, the company will ln a society with a worldwide mesh of intercorporate computer networks, elec- tronic decision making, electronic funds transfer, computer-to-computer order placement and robot factories capable of building customized products over- night, it seems outdated to spend years handerafting applications. In the attempt to create applications faster, critical success factors for every IS department include: 6 tne speed of applications devel- opment; 6 the speed of applications mainte- nance; and 6 the speed of cutover of new applica: tions. 'There are many diverse aspects to achieving fast applications development. This series of articles discusses these new approaches and shows how they can be used together in new applica- tions-development life cycles. If these technigues are widely adopted, IS can move into a new era of fast develop- and motivated personnel; and % Management —use of innovative man- agement technigues to break down bu- reaucratic and political obstacles to igh-speed development. AE encompasses a set of technigues that can be used to build complex, stra: tegic and mission-critical applications in months rather than years. The RAD life cycle is highly flexible and can be adapted to the nature of the system and the tools being used. Today there is an astonishing differ- ence between the best applications de- velopment and that in the majority of IS departments. As shown in the figure, the produetiv- method, a function point corresponds roughly to 114 lines of COBOL code. . The productivity figures in the dia- suramarize measurements of the number of function points that were produced per person-month in a range of actual applications using widely vary- ing development technigues. The cost figures reflect the cost in dollars for each completed function point. The high productivity figures summa- rized in the figure are based on real or- ganizations that build systems guickly and can do so repeatedly. As in scientif- ic experiments, the success must, be re- peatable in order to be interesting. The examples cited use tools and tech- How the Methodology of Rapid Application Development (RAD) Pays Off 40 Broduetivity And Cost Comparisons U) $309 So Among Alterative Developitent Teehnigites To a large extent, this increase in the business metabolism is caused by the use o/f' networks, immediate access to data and powerful decision support tools on the desktop. si] j Hd ' ment that facilitates dynamic changes in the way corporations opetate. | | Rapid applications development refers to a development life cycle designed to give much faster development and high: er-guality results than the traditional life cycle. It is designed to exploit pow- erful Severne software. t has re: volv ni AD life cyele is based on signiti cant improvements in k tej 'e Tools—code genera! ga leg fourth-generation languages; e Methodology—life st optimized for "speed development; Ev raerikai o£ highly trained vm TU | ity of small teams using RAD technigues is more than 10 times greater than the productivity of those using well-man- aged third-generation language life cy- cles with productivity aids. The productivity figures are measured |in the form of a complexity metric called function points. Complexity is measured by counting the number of in- puts, outputs, storages of system data, inguiries, files and external interfaces in a program module. ma applles a weight to each element to pro- duce a measure of the complexity of each module, independent. of computer language or environment. Using this ] ". provements on virtually every project, nigues that are transferable to all IS or- ganizations, given appropriate manage- ment. Lower Cost Technigues for fast, applications devel- opment almost always result in lower- cost development. A key to fast development is using small teams rather than large ones, Small teams can be made highly produc- tive if they use methodologies for rapid analysis and power tools that generate code. A reduction in development time is thus combined with a reduction in the number of people needed. Integrated CASE tools that can gener- ate code for complete applications from design specifications are appropriate power tools for applications develop- ment in a RAD environment. The CASE tools on the market vary greatly in capability: With some, little improvement in productivity occurs. It is critical to select integrated tools that generate program code for complete ap- plications. The best tools are guite ex- pensive, but not compared with the cost of large teams of IS professionals using the old technigues. Fast development does not mean "guick and dirty." It is necessary to build applications of high guality—in- deed, much higher guality is needed than in many of the applications built with traditional methodologies. Some IS organizations using CASE tools have not improved productivity, lowered costs or shortened development: times. Only a small proportion of CASE users have learned how to achieve dra- matic improvements in these areas. Some IS executives achieve these im- and it is these users who are gaining a major strategic advantage for their orga- nizations. Next week, [ll discuss the actions that must, be taken to retool the IS organiza- tion in order to gain a competitive ad- vantage. B s: The embodied in RAD were introduced by James Martin at a live satellite conference on Dec. 12. For more information, call (800) 242- 1240. For information on seminars, please contact (in the United States and Canada.) Technology Transfer In- stitute, 741 10th St., Santa Monica, Calif. 90402 (213) 394-8305. In Eu- rope, contact Savant, 2 New St. Carnjorth, Lancs, LA5 9BX United Kingdom (0524) 734 505.