In this review we address questions which must be considered if better attempts are to be made to treat all persons presently infected with human immunodeficiency virus (HIV). There are thirty million people in the world presently living with HIV, only 10% of whom are likely to be able to access currently available drug therapy. Even when available, such therapy causes considerable inconvenience and undesirable clinical side effects, and fails to eradicate virus from a small reservoir of latently infected cells. Thus, we must ask what forms of alternative therapy might be used. One strategy that may be considered is to reduce virus levels as low as possible using highly active antiretroviral therapy (HAART), followed by modulation of host immunity with immunotherapy in order to effect an appropriate and efficient response mimicking that found in long-term asymptomatic patients, with the aim of indefinitely maintaining the asymptomatic period following discontinuation of chemotherapy, or even of eradicating the virus from the latent reservoirs. In 1987, long before the advent of highly active antiretroviral therapy, J. Salk proposed the use of a 'suitable potent non-infectious (HIV) immunogen' to delay or prevent the development of AIDS in infected individuals (1). The objective of administering such an agent was to 'enhance and prolong the presence of (immunologically) protective factors'. The stated aim at that time was 'to destroy virus and viral antigen producing cells by the induction of the immune system's cytotoxic mechanisms known to rid the host of virus and virus producing cells'. Twelve years later, and after a quarter of a century living with HIV, and with the advent of HAART, is it time to use our knowledge of the host's own immune system to fight this seemingly intractable invader?