In searching for game-design tools (in the abstract sense) I came across this:
There are a lot of neat ideas in this presentation, the one that has helped me the most is something I'm adapting. Rather than describe gameplay in a linear fashion, we can describe it in terms of causal chains. So we define an objective, and then figure out what sub-objectives there are beneath it:
[Win Soccer Game]
[Shoot At Net]
[Move Ball Down Field]
[Deke Out Opponents]
[Protect Ball Carrier]
[Deke Out Goalie]
[Don't Get Scored On]
[Don't Let Opponent Get A Shot]
[Defend Against Area Around Net]
[Screen Opponents Shot]
[Position Player Near Net]
[Goalie Makes A Save]
[Position Goalie Between Attacker And Net]
This is a nested diagram showing objectives to win a soccer game. To win you have two main objectives, carried out in parallel: don't get scored on, and score goals. Although you might be focusing on one before or after the other, they both are valid objectives at any point of the game if you want to win, so they are parallel. Now, to score a goal we need to do a few things in parallel: shoot at net, and potentially deke out the goalie. Deking out the goalie might have it's own requirements, not listed here, but to shoot at the net, we need to move the ball down the field, protect the ball carrier, and so forth. Although it might not be too useful to deke out the goalie when the ball is a long ways away from the net, it's still something to do that helps towards scoring a goal.
This is useful in game design because we can chart out the player's decision making process, and evaluate what decisions they have to make as well as what options they have. If we have too many goals in parallel, then probably the player will be confused and unable to choose. This is what I did with Venture the Void, I'm afraid. If I were to redesign it, I would (before programming the real details of gameplay) think about the general ideas for gameplay that I have in mind (e.g., trading, shooting aliens, questing) and try and see how they break down into an causal chain. Then I could see where there was maybe too much paralellism, or not enough, and I could also see where to focus communication with the player as to what they should or shouldn't be doing.
If we don't have enough parallelism, however, then the gameplay will likely be percieved as too linear. This is what can make certain old-school style of adventure games frustrating. Usually there is only one or two puzzles you can complete next, and so the game is an extremely long causal chain. Figuring out each action requires some insight, and in case we miss something, we get totally stuck.
I've drawn the soccer example above in text; so things nest only a certain way. But we can also draw each objective in a box, and then draw an arrow to any and all prerequisites. This allows us a more complex model of gameplay. That said, there is a risk of confusing the player with a diagram that isn't hierarchical enough (i.e., too graph-like, not enough tree-like).
A great game like Zelda: Link to the Past works because it has a great balance between paralellism and linear depth. Your main objectives are to solve the dungeons, and to get stronger. Solving the dungeons requires first finding the dungeons on the overworld, which requires items from previous dungeons. Each dungeon is itself a complex causal chain. So we could, first of all, draw a causal chain of what dungeons require solving what other dungeons; this would be more or less linear. Awesome, creative speedruns notwithstanding, that's how most of us approached this game.
But the other main goal-- to become stronger-- is highly nonlinear, with many options open to the player. Because the overworld is constructed in such a way that you require items from dungeons to explore certain areas, what happens is the "solve dungeons" chain ends up gradually opening more and more nonlinear choices to the player in the way of becoming stronger. By the end of the game, you have a lot of tricks up your sleeves!
Anybody who has played Zelda: LTTP knows that it has an amazing sense of exploration; of choices you didn't imagine earlier on opening up to you. And yet, there are plenty of games with more secrets, larger overworlds, items, and so forth. The reason that it works so well in LTTP is the carefully structured causal chains. And while this rationale might sound a bit abstract to someone who hasn't played this game, I think it's actually pretty straightforward.
So then, in conclusion, I'm putting together the gameplay elements and ideas I have for The Real Texas into causal chains. This is work best done on paper, sipping tea and looking out my fabulous front window at all the snow falling softly outside.