Other Details ------------------------------------ There exist numerous similarity measures, but there is no generic measure that applies to different types of data. This paper presents a conceptually uniform, generic approach to measuring similarity: count common neighbourhoods. This approach has resulted in novel similarity measures for multivariate data, sequences and trees. Evaluation shows that they outperform a range of state-of-the-art measures. This work formed the basis of an EPSRC proposal on structural information retrieval. Although the proposal was not funded, it was viewed as highly innovative and ambitious. A revised proposal has been prepared and will be submitted shortly.