; EDEN AA ONE ANI o ler io Pace 86 This is the last of a series on IBM'S recently announced, Re- pository, a sig- naficant compo- nent of its ap- plications-devel- opment strategy Jor the 1990s. 'The adoption of a standard re- pository for de- sign information is vital to a suc- cessful strategy for computer-aided software engineering (CASE) for both end users and vendors: In accepting a repository standard, users of compliant CASE tools will be able to share design specifications, while GASE vendors will be allowed to compete in an open environment, Its that simple. The introduction of the Application Development (AD)/Cycle by IBM clari- fies many of the issues in defining a strategy for CASE. Third-party vendors of CASE tools now can decide whether or not to adopt the de facto standard for design specifications defined by IBM. Users of CASE tools also can decide whether to limit their purchase of CASE tools to those compliant with the pub- lished standard. Still, there are other is- sues relevant to the implementation of a successful CASE strategy. These include support. of top management, selection of highly integrated CASE tools, use of a methodology optimized for high-speed development, and issues related to the training, education and motivation of the development staff. A ehief objective of a CASE strategy is a substantial boost in the productivity of the applications-development process. As shown in the figure, a large improve- ment in productivity depends on four factors (depicted as [our tall pyramids): igols, methodology, personnel and man- agement. Correct judgments in all of hese areas must be made to be success- ful in CASE. The management issues and the people issues are especially crit- ical. There are many examples of CASE strategies that have failed due to a lack o aHention to these issues in making the transition to a new development en- — The announcement. of the IBM Reposi- Vory provides information-systems IS): development. technigues will reduce Cosls and increase iveness implementing prdi o p Organizations planning for CASE should use di: technigues, such as entity-relationship as well as data-modeling and process-mod- eling technigues. Above all, they should plan to use the most powerful develop- ment methodologies available, such as information engineering, to build the PO WEEKNAPPLIGATION DEVELOPMENT APPLIED INTELLIGENCE | strategic applications of the 19905, In- formation engineering is a methodology (hat was specifically designed to sup- port the development of the complex, mission-critical applications reguired for the computerized company of the future. Managers need to learn how to imple- ment new business thrusts faster than their competition through the use o£ high-end, integrated CASE tools and methodologies that are optimized for high-speed development. A large part of the success of high:-speed development depends on building an infrastructure. It is possible to develop applications on a project-by-project basis with no infra- IBM's Repository Is Key to a Successful CASE Strategy ticulate a strategy for CASE. The man: ager must understand the fast applica- tion life čycle and reusability, and be — able to select CASE tools that meet the evolving needs of the organization. It is important to select tools that are highly integrated and that support the IBM Repository standard. Unfortunately, the IBM Repository is not available; even in June 1990, it will still be a frag- ment of what is available now in the re- positories of highly integrated CASE tools from third-party vendors. 'The re- positories in these tools are being rapid- Iy brought into compliance with IBM's Repository. Some of them are also supported by — Fout Essential Aspects | Of Fast Applicatioi Development — | High-Speed Applicatioti Developmetit Tralned People Forimblized, , (Preclse Methotlology HM. With a repository standard, users of compliant CASE tools will be able to share design specifications, while CASE vendors will be able to compete in an open environment. structure at all, but development will be much faster if resources are devoted to an infrastructure that a high level of i Roni ječe zadi - em uni ue achieve fast developmeni i Sea t, the. odology zom resemble a finely clock, eh all perfectly, If any of the fo man meh pment will not poi rapid pace, Will not proceed ata An important aspect o£ strateg CASE is selection of the ie sij The manager can be deseril vice president of CAS; BA clearly understands the eriti factors of the Organization ad sa ar- AI O Ž';..ŽŽ EEA very detailed methodologies, such as the Information Engineering Methodology from James Martin Associates. Productivity improvements of two to three times over conventional tech: nigues are guite achievable with high- end, integrated CASE tools, including the effort devoted to extensive front-end design activities. Such improvements do not reguire sophisticated development aje ere are many examples of corpora- ira that have improved productivity A tor of 10, and many of these do ton en a sthodologies such as informa- rather, they were the result of single Project, development. It OOCTOBER 23, 1989 is much easier to AK zane : uic] with single project development OZENU iswiti Mreaaaen engineering, because information engineering re- guires approximately two to three years to build the infrastructure. However, once it is in place, it is possible to build applications fast consistently. To effectively monitor productivity, it must, be consistently measured, Top management should be informed that a metrics program is being put in place and that, measurements of productivity will be reported to them on a regular basis. This may alarm some personnel because it gives the CEO a means of measuring the success of the IS depart- ment. However, consistent, measurement and reporting of productivity gains are a powerful way of attracting and retain- ing the interest of top management in a CASE strategy. Some CASE vendors are beginning to provide metrics automatically from in- | formation stored in the repository, but | much more needs to be done. As thean- | alyst creates design diagrams, the sys- | tem can track the number of inputs, ; outputs, inguiries, files, interfaces and 4 algorithms. These measurements can be | expressed in the form of funetion šč G points, which are a measure of the com: V plexity of the module. As the design progresses, the function-point estimate is Ci refined. Before entering the construction TI phase, a detailed function-point compu- » tation is made, which can be combined with other project-management informa- pi tion to provide produectivity measure- bi ments. All CASE vendors should provide 4 these metrics automatically. k Choosing the Right Vendors i d Another important step that organiza- | tions can take now to prepare for the si introduction of advanced applications- d development environments such as AD/Cyele is to select CASE vendors | u who have committed to the common Re- s pository standard. As of last week, 32 U CASE vendors have pledged support for IBM's AD/Cyele Repository. lna suramary, the successful implemen- tation of a CASE strategy relies on many factors. Primary among these is attracting the interest and support of top management. Additional factors leading to success include selection of a highly integrated CASE tool and a meth- | odology that is optimized for high-speed development. Issues involving training, motivating and educating the develop- ment staff may be more important to | the successful introduction of CASE than technology issues. The management. of cultural change within an organiza- tion will be a major challenge for the computerized company of the 19905. 8 To learn more about the subject of these articles, please call The James Martin Report, an information service updated guarterly, at (800) 242-1240. For injormation on seminars, please contact (in the United States and Can- ada) Technology Transfer Institute, 741 10th St, Santa Montca, Calif. 90402 (219) 394-8305. In Europe, con- tact Savant, 2 New St., Carnforth, Lancs., LA5 9BX United Kingdom (0524) 734 505. | 4 , I TU COF RA TTTI | ET