We present an automatic system for evaluating programs at programming contests which we have developed. We also discuss perspectives on the future of such systems.