Encyclopedias

Having been asked to fill an entry on Monte Carlo for the incoming International Handbook of Statistical Sciences, I obliged by writing a short piece, whose utility is rather limited. Indeed, I am afraid I am not very much convinced of the use of such encyclopedias, as they try to provide entries on about “everything” but end up being partial, quickly obsolete, and not so informative… This may sound overly negative, but I never ever use this kind of books for my research or my teaching, so I wonder who does. Encyclopedias (encyclopedii?) were fine when the “whole” of Science could be crammed in three dozen volumes with a slow enough updating process. Current students most likely check on Wikipedia or at large on the Web and, given the price of those behemoths, it seems only libraries can afford them, and then this may be wasted money anyway! My misgivings actually extend to contributed volumes that are (were?) fairly common, mostly in connection with one conference or another, and whose utility is rarely demonstrated. It takes highly dedicated editors to turn contributed volumes into useful coherent books, one example being the 1996 MCMC in Practice book deeply edited by Wally Gilks, Sylvia Richardson and David Spiegelhalter, that still serves as a reference for MCMC methods. Even the traditional Valencia volumes, while giving a snapshot of the on-going research in Bayesian statistics, along with a collection of discussions, have lost some of their luster, when compared with the impact of a discussion paper in Bayesian Analysis… Having conference papers submitted to Bayesian Analysis and selected on the same basis as regular papers, as was done for Valencia 8, seems to me a much better idea than publishing separately a rather costly volume, not read by enough people. (The dates and locations of Valencia 9, the last of the kind, have been announced. It will be in Benidorm, on June 3-8, 2010.)