A number of activities are underway related to understanding R performance in practice and to exploring techniques for improving performance generally or in important special cases. I'm writing directly on behalf of an informal group centered in the Computer Science department here at Stanford, but also on behalf of similar activities elsewhere. We would all be helped by realistic examples of serious computations with R, where substantial computing is required and the results are important to users. The more and the more representative the data available, the more likely that useful performance tools can be designed. If you can provide us with examples that we can use and share, please send a description to: Justin Talbot <jtalbot at stanford.edu> As examples accumulate, the plan is to make them available to the community---details to be worked out. Thanks in advance, John Chambers