Computer-aided process planning has been under development for some years. Considerable interest has already been shown in the application of artificial intelligence to the automation of process planning. However, it has been found that the existing knowledge-based systems tend to be unrealistic and unreliable. A large gap still exists between what an intelligent system offers and what manufacturing engineers want. There are several crucial areas awaiting improvement; the methodology for knowledge acquisition and representation, the speed of searching and retrieving of information, and the user interface to the planning system. The system presented here is a symbolic-oriented, frame-based process planning system. It describes a component shape by using descriptive vector primitives in the design stage which can be extended to the stage of manufacture. It is designed to combine the merits of both the generative and the variant techniques. The generative technique enables simple and direct generation of both geometric models and process plans, provided that a suitable decision-making mechanism is available. The merit of the variant technique is that information in the database can be easily retrieved. A flexible, frame-based approach to manufacturing knowledge acquisition and representation, which is fully supported by an implicit, logical, decision-making mechanism, is another feature of the system. The information required by the system is direct and simple and can be modified or extended if necessary. Alpha-numeric symbols are used to represent knowledge internally. The selection of frame information leads to the production of clustered strings which are used as system control logic for the decision process. An attempt has been made to develop a system interface with the user which is as efficient and easy to use as possible, and through which the planning system can learn by instruction.