CRE paper

Friday 4/6/07 we reviewed Rao et al. (2004) IEEE Trans Info Theor, V 50 (6) “Cumulative Residual Information […]”, here.

We decided that while the motivation for the paper was valid, that it was undesirable for a number of reasons — mostly that the CRE of many well-behaved distributions (power laws notably) diverged. We’re all currently working on better generalizations.