Method of using redundancy in very large computer stores

The organisation of a very large random-access computer store which uses redundancy in the form of error-correcting codes is described. The improvements in manufacturing yield which can be achieved by means of such redundancy are emphasised in the form of examples. However, with typical basic store sizes in the capacity range 107¿108 bits, some of the most significant improvements in manufacturing yield are achieved by programmed assembly which makes use of the ease of testability of such redundant systems. The conclusion is that widespread use of the proposed technique of design and manufacture can result in significant savings, both to designer and customer.