Associative processing based on content-addressable memories has been argued to be the natural solution for non-numerical information processing applications. Unfortunately, the implementation requirements of these architectures using conventional electronic technology have been very cost prohibitive, and therefore associative processors have not been realized. Optics has the capability over electronics for directly supporting associative processing by providing economic and efficient interconnects, massive parallelism, and high-speed processing. This paper presents the principles of designing an optical content-addressable parallel processor called OCAPP, for the efficient support of parallel symbolic computing. The architecture is designed to fully exploit optics advantages in interconnects and high-speed operations, and is potentially very suitable for applications where the number of data sets to be operated on is high. Several search and retrieval algorithms are mapped onto OCAPP to illustrate its ability to support parallel symbolic computing.