Finished Analytics&OLAP in SQL by Joe Celko today.
A very “blog-style” book, some very good ideas, some repeats (read “Trees&Hierarchies in SQL” previously), sarcasm and a set of good examples. Not a starter-book, but a recommended reading for anyone interested in data analysis.
“Tape-based” approach and Von Neumann logic still dominate IT. SQL is bad for analysis, but what’s better? Celko’s MDX-criticism seems right. And we definitely need a pack of open-standards in OLAP.
And I, specifically, need to study stats more. All those probability things, quantum things 
…
Liked the book format very much (wide-enough fields to write comments). Enough for review.