Most of these follow a standard formula. There’s an introduction, sometimes written by a guest, telling of Colossal Cave, Infocom and Adventure International. There’s a discussion of what a text adventure is (the term “interactive fiction” is used occasionally) – the books are unanimous that an adventure is, at heart, a puzzle. The idea that the game might tell a story is still firmly on the back seat. Then there are the sample games, usually listed in some monstrous home-computer BASIC, with annotations going into various levels of detail.
Peter Killworth’s How to write Adventure Games for the BBC Microcomputer Model B and Acorn Electron is one of the most interesting of these books.
It has some nice thoughts on puzzle design – which it calls “plotting” – particularly about building multi-stage puzzles where the player must solve a chain of logical steps one after the other. These do make good puzzles – the Babel fish vending machine in Hitchhiker’s Guide being the canonical example. (Killworth prefers to kill the player off or make the game unwinnable when they get one of the stages wrong, which has happily gone out of fashion, but other than that his thoughts on this type of design are still worth reading.)
On the other hand, the sample game Roman is deservingly cited in the Inform Designer’s Manual as an example of bad design, because almost the entire game consists of essentially one such puzzle-chain, so there’s only one thing for the player to do at a time. This sort of game needs to be broad as well as deep, and if I had to choose one over the other, I’d choose breadth.
Killworth also sees it as fundamental to the job of the IF author to make the player suffer (I’m not interpreting; he repeatedly says this in exactly those words.) One of the things Roman has going for it is the choice of setting, ancient Rome. It’s a setting that’s easy for the player to imagine – which helps when you’re limited to a describing any room, NPC or object in a sentence or so – and instantly provides some seeds for plot or puzzle ideas: gladiatorial combat, corrupt senators, Julius Caesar as an NPC. Then in one room there’s a “pilum” lying about. This would be a nice bit of flavour if the game bothered telling you what a pilum is. Not a Latin scholar? Well, you’ve got Google. But in 1984 you might have had to go to a library to find your nearest Latin dictionary. This isn’t an oversight – Killworth discusses putting in this information but deliberately leaves it out: “If we refer to it as a pilum, the player will probably not know its use, which is all to the good.”
This might make some sort of sense in, say, a time travel game where a Roman legionary tells you about his pilum and you have to figure out what he means – but no, you are a Roman, you see the pilum itself, and the game chooses not to tell you what is in front of your eyes. There’s no EXAMINE. You really do have to go and look it up in a Latin dictionary or a history book (or, and I fear this might be what Killworth is thinking of, have the right sort of education.)
The great cryptic crossword setter Araucaria described setting a puzzle as “a game in which the goal is to lose gracefully” – you shouldn’t want to stump the player completely, any more than a novelist wants the reader to give up after chapter 2. The idea that the IF author is trying to ‘beat’ the player is one that’s thankfully alien these days.
The book does include something in its introduction that is still certainly true after 30-odd years, though: “No matter how small an Adventure you write, it will take far, far more time and effort than you thought it would.”