A framework for de-centralized use of computers

With the prices of microcomputers coming down and the systems becoming more and more user-friendly, use of microcomputers in a de-centralized set-up has been steadily increasing. PCs are used in a number of de-centralized data processing and decision-support applications. In this section we discuss two types of use in detail.

Systems for data processing

One of the most common applications of microcomputers in a de-centralized set-up is data processing. These applications are developed for either regular use or prototyping.

Regular data processing systems

Since software available on PCs makes it possible to develop data processing systems with less effort, one is often tempted to develop systems for regular use. These systems can, however, be successfully implemented only if the security and integrity features, which are weak in PCs, are achieved through externally imposed data access discipline, i.e., by following certain norms for accessing database files and by establishing procedures of back up and recovery. Normally such externally imposed discipline would function well if the systems are managed by individuals or by close-knit groups. Special efforts are, however, required to extend PC-based data processing systems to a general-user environment, because it is difficult to impose security and integrity disciplines externally on a large group of users. Advanced PCs which offer UNIX-like operating systems and advanced database management systems are one solution to this problem. Today, a large number of organizations are adopting this approach. Quite a few data processing applications are developed and regularly used for processing using PCs.

Prototype information systems

Development of illustrative systems in pilot projects is another popular use of PCs. Using end-user software packages, it is possible to develop, easily and quickly, a live model of an information system, involving all steps of information processing. Such systems can be subjected to field tests through installation in the pilot project areas. Experiences of using the system and suggestions for improvement can be documented. Subsequently, the main system can be developed using the appropriate technology, taking into consideration the experiences and suggestions resulting from use of the prototype. Prototyping in this form also facilitates user education and improves user participation in the computerization process.

Systems for decision support

From earlier discussions, it is evident that there is no dearth of software tools available on PCs to develop systems to assist the decision-maker. In fact, users have to prepare themselves to meet the challenge of utilizing the power offered by inexpensive and yet powerful information technology.

To design systems for decision support in planning, users should acquire model building and optimization skills. To design decision support systems for monitoring, users should acquire a feel for numbers and develop better indicators of performance using advanced statistical techniques. In both cases, development of aesthetic screen interfaces is an art to be acquired through experience. There are a number of instances where users have developed powerful decision support systems using end-user software without the assistance of systems specialists. However, beyond a certain level of complexity, users need to acquire system design and programming skills. More importantly, users should concentrate on acquiring the modern tools of problem solving in their problem domain to use the microcomputer technology to its fullest potential in a de-centralized set-up.

Systems for evaluating alternatives in project appraisals, project monitoring, profitability analysis, market research studies, production scheduling, inventory management, purchasing, portfolio management, advertising and engineering design are some examples of PC-based decision support systems.

System development strategies

Two approaches are generally used in the development of computer-based systems in a decentralized environment.

In the first approach - the traditional - a centralized department, such as a computer services department, develops systems and installs them on PCs. This approach has all the advantages of using expert skills in developing systems which are vital to the successful implementation of complex application systems. If systems so developed meet most of the user requirements, their acceptance should be high, primarily because:

· users interact directly with the systems for solutions; and
· users can enhance the system at local level by use of simple add-ons based on use of peripheral applications developed using the computer and end-user software on hand.

Major limitations in this approach are:

· it may not ensure total translation of users' requirements into the computer system because of imperfect communication between systems developers and users; and

· users may not wholeheartedly endorse the systems which are not developed by they themselves.

In the second approach, users themselves develop their applications with the help of end-user software packages like electronic spreadsheets and data management systems. In this case, since the problem context is very well known to the developers, the effort involved in completely translating the users' requirements into a computer-based system is minimal. It can be expected that such systems get implemented smoothly since users are the owners of the system. It should also be possible for users to introduce improvements to the system from time to time. The current trend in a number of organizations is to encourage this approach.

Some problems likely in this approach are:

· excessive de-centralization might lead to data indiscipline, making system integration a difficult task. Each department and individual might develop individual coding schemes, define their own data fields and the types and sizes of data to be handled. Sharing such data would be difficult if proper standards are not evolved and enforced;

· lack of systems analysis and design skills in users might result in the development of half-baked products. Systems which are not thoroughly tested might be put into use while users are totally ignorant of any bugs in the system. Systems developed must therefore be subjected to rigorous checking by others not associated with development before releasing them for regular use;

· de-centralization may lead to disintegration if each individual solves his or her problem in isolation. One might develop an efficient system within a very narrow scope of a department or individual, but, in a number of cases, such solutions probably turn out to be inefficient solutions overall;

· users may tend to be possessive of 'their' systems and databases, and may not share them with others; and

· lack of exposure to decision analysis and model building techniques might result in the development of mundane applications, where the value added to processed data is negligible. Users might waste their energy in playing with computers to developing more cosmetic features rather than objectively analysing results and taking the necessary development action.

In spite of the above dangers, development of applications by users is an ideal solution to increase users' involvement in information processing in organizations. Perhaps a mixed approach is desirable. Based on the organization culture, each organization will have to work out a strategy of information processing and cautiously blend technology with decentralization. A core group from management services, computer services and user departments could analyse the issues and work out a strategy to take advantage of developments in information technology.