The generic parts of an expert system are furnished by an expert systems shell. The human interface part includes questions and explanations for the user. The inference engine part is designed to facilitate problem-solving through the use of logic. The knowledge base, which is based on logic and plain English-language statements, is designed to provide answers to problems. Software shells are capable of non-numeric logic and can coordinate the decision-making logic of an expert with virtually all fields of expertise.
Computerworld. Nov 14, 1983, Vol. 17 Issue 46, p45
Insurance Industry, Artificial Intelligence, Workstations, Networks, DBMS, Database, Online, Future Technologies, User Interface, Information Systems, and Information Resources Management
By 1990 developments in data processing and communications technology will bring dramatic changes to the insurance industry and other businesses using stored information resources. Information will be kept in true data bases which will be accessed on-line for updating, inquiries, and transactions. Home terminals linked to a company's mainframe computer will provide for electronic purchases of insurance policies. Workstations will give users the ability to produce, store, retrieve, manipulate, and analyze information located in personal, corporate, and external data bases. All hardware will be seen as part of a communications network which the user will join when he sits at a workstation. In the 1980's methodologies such as prototyping, along with dedicated efforts by information system managers and end users, will be needed to bring about these advances.
Computerworld. July 8, 1987, Vol. 21 Issue 27, pS55, 2 p.
Artificial Intelligence, Productivity, Software, MIS, Market Segmentation, Expert Systems, Natural Language Interfaces, Software Complexity, and General Ledger Accounting Software
Artificial intelligence (AI) has two commercially valuable market segments: expert systems and natural language interfaces. Some of the productivity gains made by natural language interfaces include: applications prototyping, database applications design, development and maintenance, IBM structured query language coding, user self-sufficiency, and streamlined applications. Expert systems provide MIS with three main advantages: a new applications group, enhancement of existing applications, and the proliferation of intelligent debuggers. Natural language interfaces and expert systems can be used to enhance the productivity of general ledger applications.
Computer-aided software engineering (CASE) programs were introduced at the CASE Expo in Oct 1987. Products included Refine from Reasoning Systems Inc and Procap, Re-Source and Pro-Source from Promod Inc. Minimal training is required on Refine for programmers experienced with artificial intelligence languages. The program enables rapid prototyping and validation and has a specification compiler that provides declarative expressions and low-level procedure statements. Promod's new products are a family of integrated development tools. Procap is used for design, source-code development and maintenance, and documentation. Re-Source has a design library that feeds program elements into the modular design methodology, and Pro-Source is a code generator.
Computerworld. Oct 13, 1986, Vol. 20 Issue 41, p17, 1 p.
Expert Systems, Artificial Intelligence, Productivity, Efficiency, System Development, Analysis, and Product Development
What expert systems practitioners do and how they do it is no longer a mystery now that expert systems are available commercially, but while some individuals believe that the expert systems business is based on techniques similar to conventional systems, in reality the development of expert systems differs considerably from the development experience of conventional systems. A close examination of the expert system development reveals the fundamental difference between conventional systems and expert systems: a small prototype solving a small portion of a problem is the beginning of an expert system, a process different from any formal prototyping and in contrast with most methods for constructing conventional systems. Developers of expert systems are not prisoners of their systems, and they are constantly learning more about how the systems should be constructed. Expert systems are not magical, but they do represent an honest attempt to improve productivity in organizations.