The review begins by providing a brief typology of biological databases on the Internet, illustrated by examples of the most influential resources of each kind. We then take an insider look at one typical on-line genomic resource – the yeast genome database hosted at the Munich Information Center for Protein Sequences (MIPS) – and explain how and why it has evolved from a basic sequence repository to a multidomain knowledge base. The role of community efforts in curating and annotating genome data is discussed. The crucial role of data integration and interoperability in developing next-generation genomic facilities is underscored.