packaged im Berlin

I ended up spending more time working in my Berlin rental flat than in the rooms of the Intercontinental Berlin where WSC 2012 was taking place (just next to this canal along which I ran the next morning). Since Jean-Michel Marin was (fully!) available, we indeed tried yet another time to complete the Bayesian ex-Core soon-Essentials with R that we started more than two years ago near Marseille… For one thing, I wanted to write the bayess R package that should come along with the book. So I spent about twenty hours in Berlin working on the demos and on the maddening .Rd manual files that came with the fifty (shades of!) R functions that appeared throughout the book text. Without managing to reach the end of the list! Writing the manual and checking for each R function seems to take forever, actually, even in optimal working conditions such as the current one! I just cannot fathom how I did it in Italian trains for the previous mcsm…Anyway, tomorrow, I need to take a train ride to Grenoble and back, so hope to complete the tedious task!

The other “advance” we made about the book while in Berlin is that we will most likely give up on the (new) chapter about hierarchical Bayes models… The very one we have been pursuing for a long while, since we were already discussing the Scottish Lip Cancer dataset in Luminy more than two years ago. On the one hand, I was/am not completely convinced of the relevance of a hierarchical chapter, given that the theme is somehow transversal and pops in the mixture, dynamic and image chapters. On the other hand, despite Jean-Michel’s stronger commitments to the chapter, his advances were hindered by difficulties, both endogenous and exogenous, and I did not feel we could easily introduce some codes written in JAGS and BUGS in a book of this level. Add to that the recent highly relevant publication of the BUGS book. And the incoming new edition of Bayesian Data Analysis. And the tired feeling of having delayed way too long. And running out of steam. Then you get enough hints to explain our decision to cut our losses and move towards the foreseeable completion of the book by cancelling most of the hierarchical chapter.