there's a reason why people on forums and chans consistently recommend land of lisp over other books, it's deliberately imperfect in a way which appeals to a certain kind of young hacker's mind. it doesn't assume basic lisp knowledge, it assumes you're going to type the program in without understanding and then figure it out from exploration. it also assumes that you won't understand the whole program immediately or even after reading the whole book. it's a different didactic approach, and is modeled after author's personal experience of learning how to program by typing BASIC programs from magazines. as such the programs are deliberately "misarchitected", they are somewhere between scripts and hacks that present a variety of ways in which common lisp the standard allows you to solve this or that problem. they are also designed to be "tasty", compact and quirky, as opposed to well structured, plodding and dogmatic. they are lifted directly from a wizard's laboratory and they invite play.
by comparison touretzky is a college level slog, solid, dependable coursework, that takes you over all the proper subject matter in a way that's certain to be approved by any computer science department committee, but that fails to create any sort of motivation to engage.
by comparison touretzky is a college level slog, solid, dependable coursework, that takes you over all the proper subject matter in a way that's certain to be approved by any computer science department committee, but that fails to create any sort of motivation to engage.