Approximately 25% of all drug-like molecules exist as more than one tautomeric form. This affects all aspects of computer-aided molecular design. One challenge is that frequently biomacromolecules bind a rare tautomer, which may not be enumerated by a tautomer generation program. A second challenge is to devise a database scheme that will enable one to store observations on both the structures and the tautomeric ratios of a compound as observed or calculated in different environments. A third challenge is to decide how or whether to include tautomers as a result of a substructure or similarity search--should only 'predominant' tautomers be searched and reported, or 'all'?