Throughout this brief review I have emphasized the unique biochemical and immunological properties of MUC-1 mucin that make this tumor-associated antigen a novel, exciting and tangible target for tumor immunotherapy. The tandemly repeating nature of the antigenic epitopes in the mucin polypeptide chain, the under-glycosylation of these epitopes, and the expression of the molecule at the cell surface, are all central to the immune recognition of this antigen and must be acknowledged in the design of MUC-1 vaccines. There are considerable difficulties associated with such a design. Short, immunogenic peptides that associate with MHC class I molecules on the cell surface to induce CTL responses are not useful here; MUC-1 must be expressed at the cell surface and this introduces the numerous problems associated with gene transfer. Furthermore, generating under-glycosylated mucin molecules that resemble tumor-associated antigen is not trivial. Competitive inhibition of glycosylation with PhGalNAc is often incomplete, and merely increasing inhibitor concentration results in cell toxicity. But treating tumors is also not trivial. MUC-1 mucin has many characteristics of the ideal tumor-associated antigen, and our understanding of the unique mechanism of mucin recognition on tumors and the appropriate vaccine designs to target this antigen is now advanced. Transgenic mouse and non-human primate models provide excellent preclinical models to test immunotherapeutic and vaccine strategies, and in vitro studies and early-stage clinical trials in humans provide considerable cause for optimism. The next few years in this field should certainly be productive.