Associative memory clearly deserves much more attention than it receives nowadays. When it was proposed in the 1950s, the available technology, the programming style, and the cost per cell combined to make it unattractive. Although this approach was chosen for database machine design and AI computers, no serious effort was made to produce computers with large associative memories.
This paper should help resuscitate interest in associate memories. It documents their advantages and disadvantages, the state of the art, and the most interesting applications. The paper is well structured, self-contained, and very instructive.