Typically we treat Apress /No Starch press and especially O'Reilly books (with those black-white animals on cover) as the highest authority in IT. We naturally anticipate that those books are written by hard-core professional that can mentor of all. Well, that's not the case with Juliet Kemp book "Linux system administration recipes" - the whole book has the "unwritten concept" - "Hey, check this out!" . That means that it needs to be considered as a sort of article in Linux magazine that covers some Linux/sysadmin concepts that got Julie Kemp mesmerized, but not the solid reference material.
For example, page. 188
ldapsearch "(uuid=testuser)" | sed '/^#/d' | sed '/^$/d' - well , that works but any decent experienced sysadmin will tell you that it is not cool to pipe sed to sed, Instead you use '-e' command -
ldapsearch "(uuid=testuser)" | sed -e '/^#/d' -e '/^$/d' - that is much nicer. I am not splitting a hair here, but IMHO books like this need to have more elegance.
Also the whole idea of centralization using LDAP /NFS / puppet is not clearly illustrated in this book - a few Visio / Omnigraph diagrams will be a HUGE help to understand the entire concept and to see whether it can be applied to your network. I am reading between the lines that Juliet Kemp has discovered LDAP /NFS /puppet / bash completion / perl -and she is trying to tell us "See, what I am using... isn't it cool or what?"
We need less "discovery channel", but more analysis here. I am not telling you that the book is useless - it has some nice ideas, but the 2nd edition (if Juliet Kemp cares) need to be seriously revised