The great success of silicon microelectronics has led advocates of many alternative technologies to aspire to replace it, at least in some applications. More bluntly, almost everyone who discovers some sort of instability or evidence of gain can think of an aspect of the discovery that has the potential to improve upon the performance of silicon electronics. These proposals are made, however, without an appreciation of the sources of the success of conventional microelectronics. An examination of current technology thus illuminates the barriers that new digital logic technologies must overcome. Digital logic systems must have large numbers of components. A practical system with a large number of components requires that the cost per component be low. The power per component must also be low and the reliability very high. Silicon microelectronics has been able to meet these requirements through miniaturization and integration. These same directions of progress have also conferred high speed on microelectronics. Transistors are favoured microelectronic devices because their high gain enables a standardized output to be provided with large noise margins. Transistors also easily accommodate varying fan-out and fan-in and offer excellent isolation of input from output. Optical logic elements do not share these desirable qualities. Optics may have many roles in the implementation of computer systems, in communication, in information storage, in input and output devices, and in lithography. However, there has been little progress towards devices suitable for a general purpose digital computer in the decades since the discoveries of the ability of one light beam to influence another, and of optical bistability. © 1985 Taylor and Francis Group, LLC.