Pace 50 This is the sec- ond. in a series of articles on key industry trends affecting computer hard- ware, software, envi- ronmenis aii communications environments. Trends in com- puter software are changing rap- idly, posing chal- lenges for many organizations. Some of the trends are the growth of end-user computing, the transi- tion to graphical user interfaces; the in- troduction of powerful development: tools, such as integrated computer-aided software engineering (I-CASE) tools; and the growing commitment to integrated computing environments, such as IBM's Systems Application Architecture (SAA) and Digital Eguipment Corp's Network Application Support (NAS). These architectures and tools will in- creasingly be used to build enterprise- wide cooperative-processing systems that distribute data and processes across multiple hardware platforms (see chart). At the same time, the wide- spread adoption of software standards will provide consistent operation of graphical user interfaces, programming interfaces, communication interfaces and repository standards. Software Strategies By the end of the 19805, many large organizations had adopted a strategy for software use within the firm. Off-the- shelf PC applications were commonly used by office workers and other end users, guery languages and decision-sup- port systems were used by business ana- lysts, fourth-generation languages were used for the creation of departmental applications and some corporate sys: tems, and COBOL was still pervasive as the de facto language of information systems. But the language infrastructure will erumble in the 19905. While end users will continue to re- guire the functionality of off-the-shelf products, they will demand that the functionality of several products be inte- grated and user interfaces be standard- ized. Much of this integration will be achieved through the use of multitask- ing operating systems such as OS/2 and graphical user interfaces such as Win- dows 3.0 and Presentation Manager. While both Windows 3.0 and Presenta- tion Manager offer many of the same in- terface and multitasking features, the oi operating systems differ radi- The Windows 3.0 interface within the DOS-environment, As such, applications are limited to single tasks and restricted to a base memory of as a platform fo lon čannot be used rm for the development of Software's Future Lies in Cooperative-Processing Systems large, complex applications. Already, the basic kernel software of ICASE tools can barely operate within the base memory of 640K bytes. If the high-speed loop of these products exceeds 640K bytes, the bank-switching functions pro- vided by DOS extenders and products such as Windows 3.0 will not provide a solution. Because of the fundamental limitations of DOS, high-end CASE tools are likely to migrate rapidly to OS/2. As the hardware computing resources increase in functionality, DOS and Win: dows 3.0 will not be able to fully exploit the capabilities of the hardware; OS/2 and Presentation Manager, which sup- port both multitasking and direct ad- reguest. In addition, the use of artificial intelligence and natural language tech- nigues will assist the user. These more powerful—yet easier to use—guery and decision-support systems will become the basis for commonplace executive in- formation systems. Fourth-generation languages will con- tinue to evolve in several significant ways. First, the languages will become more portable—moving to smaller com- puting platforms. They will also im- prove in power, performance and inte- grated capabilities. COBOL will continue to retain its place in the IS organization, but will take on a new image. Rather than users The Building Blocks of Enterprisewide Cooperative-Processing Systems Integrated Computing Environments Facilitate Seamless Interaction Among a Vartety Of Platforms — Common communi- calions protocols data access via SOL Common user intertace Common programming | services John Avakan Users will need the functionality of several off-the-shelf products to be integrated. Much of this integration will be achieved through multitasking operating systems. zA of large memory applications, 'The use of guery languages and deci- sion-support languages will enable users to more easily access and analyze large databases. Except for sophisticated mathematical analyses, much of the data access and manipulation will be performed using graphical and forms- based user interfaces. It's not like the early 1980s, when a user had to understand both the physi- cal data structure and a precise lan- Buage syntax for performing a guery; newer facilities allow the user to sz Centrate on the semantic meaning of the coding applications by hand, COBOL ap- plications will be generated automatical- ly via CASE tools. Current CASE prod: ucts let the user. enter specifications us- ing graphical metaphors and symbols. 'The specifications are checked for accu- racy and completeness, then used as in- put to source-code generators. While these technigues are in their infancy to- day, they will become the standard means by which applications are devel- oped and maintained in the future. More important than the elimination of hand coding and the development of COBOL applications, CASE products will also assist in the maintenance of exist- - elient/server-based applications. In a ing COBOL systems. Reverse-engineering and re-engineering of existing applica- tions through the use of CASE products will continue to preserve software in- vestments. Perhaps the most significant trend in computer software will be the ongoing development; in the field of CASE. CASE products have evolved from simple graphical editors with data dictionaries to robust front-end tools that enfranchise systems-development: methodologies. Several CASE products are now capa- ble of supporting the entire systems-de- velopment life cycle, from business plan- ning through testing and enhancement of the application. Users will be assisted with expert-system rules that ensure methodologies are being followed. Once the application has been defined and specified, the user will select an appro- priate language version, and the applica- tion will be automatically generated. CASE products that. provide much of this functionality are already in the market. The products will continue to improve by adding facilities that mathe- matically or logically validate the sys- tems design, incorporate AI and expert. systems technigues, and generate multi- ple language versions from a singular set of specifications. Client/Server Applications GASE software products will also evolve to support the development of elient/server application, the applica- tion, data and computing resources may be distributed throughout, a network. Developing applications that are opti- mally tuned for network distribution is today a complex task, inhibiting the growth of the client/server architecture. A current obstacle in the CASE mar: ket is the lack of standardization among vendors, particularly in the repository. It is not currently feasible to mix and match facilities from several vendors into an integrated CASE environment. The repository becomes the focal point for linking diagramming facilities, design analyzers, code generators, documenta- tion generators and project-management aids. Without; a common repository for storing specifications, users are unable to select a preferred tool to support a specific function or phase of the sys- tems-development life cycle. As a result, single-vendor [CASE products currently offer the only solution for complete life- | V cycle coverage. Next week's focus will be the major trends now occurring in database-access software. M The concepts embodied in this article are described in the High-Productivity Technology volume in The James Mar- tin Report Series. For more informa- tion on this volume, call (617) 639- 1958. For information on seminars, contact (in the United States and Can- uda) Technology Transfer Institute, 741 10th St, Santa Monica, Calif. 90402 (213) 394-8305. In. Europe, con- tact Savani, 2 New St., Carnforth, Lancs., LA5 9BX United Kingdom (0524) 734 505.