This has been in draft state since September 2013, and I’ve only just rediscovered it. Disclaimer: No Starch sent me a review copy of % Realm of Racket.
I tend to read a lot of books of a technical nature, and a couple of weeks ago I finished up [Realm of Racket] by No Starch Press. No Starch has put out lots of great books recently, including (but not limited to) Learn You a Haskell for Great Good, Learn You Some Erlang For Great Good!, and Conrad Barski’s Land of Lisp (which I haven’t read, but have flipped through and would love to do so).
And, while I think that the recent batch of No Starch books targetting functional programming is great, I thought Realm, overall, was just “good.”
Despite the “just good” rating, I loved the presentation and delivery of the material. As a reader, you assume the role of Chad, a sad soul, searching for meaning in his college career. His friends suggest he becomes a programmer, and he ends up in a dungeon maze controlled by Dr. Racket. In order to leave, Chad must be victorious in a series of games and challenges. This, of course, is enhanced by Dr. Conrad Barski’s signature comic book style.
The book starts out with a great introduction to Racket, and programming in general. It then dives right in to building games with simple user interfaces, focusing mostly on game state management. Managing the state of a program is, in principle, the most important part of a program. I would have liked to see a more functional approach taught here, rather than a purely imperative one. Big bang, the rendering and event handling system that the book uses, of course treats state in a functional manner, by insisting that a user returns the state as the result of callbacks, etc. But, within the pages of Realm, this isn’t emphasized, and we’re told to modify the state.
In later chapters, networking is introduced using Racket’s universe library and packages. While, it’s great to leave out some of the underlying, low-level details, the packages abstractions are, perhaps, just a little bit too hand-wavy. That being said, it’d be very hard to teach robust network programming for low latency games in a whole book, without packages, let alone a few pages.
Finally, creating an intelligent agent is presented using minimax. While certainly not a trivial concept in itself, I thought that the explanation was a little weak and deserved a bit more detail, especially when compared to the treatment of game tree evaluation and laziness in the previous chapters–all this build up for 2 or 3 rather simple functions to make use of it.1
Overall, I think Realm of Racket is a good book best suited for early career programmers and students2. It provides valuable lessons on deconstructing real problems into smaller, more manageable abstractions, does a good job at demystifying game programming in general, and does so with a fun and entertaining presentation. Those unfamiliar with Lisp shouldn’t worry, as the first few chapters do a great job of introducing what you’re getting into. And, while I would have liked to see a deeper dive into networking and AI, both concepts are rather advanced, and I understand the authors' intent–just enough to demystify.