To reduce the cost of producing text-graphic materials, computer routines are used to generate illustrations and to automate page compositions. The present paper describes an on going contribution to this effort. Text-graphic material involved are from a narrative technical procedure description for the training of lower skilled mechanics of the U.S. Navy to operate and maintain complex technical equipment (e.g., an aircraft maintenance procedure). The procedure description text is written by the equipment designer and contains: (1) A description of the equipment, (2) a description of the procedures to be followed in using any part of that equipment. The graphic identification project is a module of the Expert Authoring System, which constructs a sementic network from the equipment and the procedure descriptions and then: (1) identify the relationships between the equipment and its components, between components and components, (2) identify the componeents needed for each task, (3) identify the results of each task, and (4) identify the relationships between the task steps. The paper describes the preliminary phases of the development of a graphic identification procedure from natural language processing of a technical text written in english. These phases involve the design of a semantic network from a given text, and the design of the support data base.
All Science Journal Classification (ASJC) codes
- Computer Science(all)