1. Juul J. Half-Real: Video Games between Real Rules and Fictional Worlds. IT University of Copenhagen, Glentevej 67, 2400 Copenhagen NV: MIT Press; 2005:225. Associated webpage: http://www.half-real.net/.
An interesting book that looks at defining what makes a game a game, starting with a 'classical' model and then looking at the ways in which that has been expanded by recent video games. He examines the interplay between rules and fiction, where the rules can reinforce the fiction and the fiction can teach about the rules.
In discussing the fictional world of games he defines 'coherent' fictional worlds as being ones where it is possible to explain all of the things that happen in the game in terms of the fiction, whereas 'incoherent' fictional worlds require the player to discuss the game rules when talking about the things that happen. An example he uses is the multiple lives of Mario - the only way a player can explain these is to say that the rules say Mario starts with three lives. In terms of our game, a coherent fictional world would mean that all of the consequences for players actions are described in fictional terms (e.g. weeding your crops a lot means you get greater yields) rather than in points (e.g. sending your kid to school gets you 20 points at the end of the game). He does point out that players don't seem to object to incoherent worlds, but that the inconsistencies tend to make the rules more obvious than the fiction.
One interesting point he makes is that rules can be used to create emergent behaviour (e.g. behaviour that the game designer didn't expect, or more complex scenarios than would be suggested by the rule set). This is contrasted with progressional games, where it is only possible to do the next thing, or solve the puzzle in one way. He discusses that some games (e.g. Grand Theft Auto) combine progressional and emergent play, by allowing the player to do their own thing in between quests, or to go through the quests however they choose, provided they reach a given end state.
One interesting point he made is that subtle tweaks in the rules make for very different player behaviours. One example discussed is the comparison between Quake III Arena and Counter-Strike (p 88-90). Apparently they have very similar rules and both support multiplayer gameplay, but Counter-Strike tends to lead to team-oriented play while Quake III Arena doesn't. He says this is because in Counter-Strike the rules have been tweaked to make it more important to have other people covering your back - subtly, by things like changing the respawn points and increasing vulnerability of players.
Possibly more interesting for game design than thesis!