A system of HIV surveillance based on AIDS case reporting is no longer adequate to monitor the epidemic of HIV/AIDS in the U.S. We are now faced with the challenge of designing an effective system of HIV surveillance. The "New Approaches to HIV Surveillance: Means and Ends" conference emphasized that there are several alternatives, each with strengths and limitations. The CDC has recommended that all states adopt a system of HIV surveillance based on case reporting. Although it has not specified that such systems need be name-based, CDC appears to reward states that adopt name-reporting systems. The rationale for this stance should be reviewed and made explicit. Name reporting may be superior in some respects to a system of case reports based on unique identifiers (UIs), especially in its greater ability to link surveillance activities to follow up at the individual level. Neither a name-reporting nor a UI approach to case reporting would provide HIV incidence data. The only currently envisioned means of providing incidence data is statistical estimation based on "snapshot estimates" of HIV incidence in sample cohorts. Calibration of this new instrument for HIV incidence estimation against existing data or through field trials is of critical importance.