Recently, I got ahold of a copy of a DS game called "Lost in Blue". This is not so much a review of the game as it is an analysis of certain design practices. Don't worry about spoilers. Seriously, it's about like saying "there are droids in Star Wars, and the bad guys kill people."
In Lost in Blue you get yourself stranded on an abandoned tropical island with a cutie one year younger than you, but you're kind enough to break her glasses, making her completely useless for things like gathering berries or walking around.
In the beginning, the game is mostly about desperately running around trying to find enough food to feed the two of you. However, as you discover various rare resources (coke bottles, green bamboo, etc) she spruces up the house a bit and makes things easier (no more fire lighting minigame? YAY!)
The thing is, this changes the focus of the game from survival to gathering enough food and twigs so that you can take progressively longer and longer walks in search for whatever the next resource will be, so you can bring it back and make better stuff.
In short, it changes the game from you and her to you and the cave.
This was a very safe choice, from a game design standpoint. It's hard to go wrong with a simple fetch-and-improve system. But it's a little disappointing, don't you think?
While the Swiss Family Robinson certainly had more than its fair share of extreme home improvement, in the end it was really about the Swiss Family Robinson, not about the House that Dad Built.
Building the house - and taming the elephants and packing explosive coconuts - wasn't done simply because it was possible to do so. It was done so that the various family members could bond and show off in various ways.
To me, I think the big problem with what Lost in Blue did is that they compressed time. For some reason, it has become very popular to compress time in any game involving little people that walk around. Presumably, it's a kind of gamer's short-hand for "things are actually much bigger": there's no way that a tree drops that many twigs or coconuts so often, so presumably the tree is a "stand-in" for a lot of trees in a lot of places.
While this works okay in some situations, it means that the act of walking a hundred feet takes half an hour of game time. That kind of pressure doesn't seem like it fits the idea of a tropical island survival game. "I can't go walking down the beach, I just don't have time to do that before she starves to death!"
This kind of compression is "the standard", so it's used quite often. But it produces a very specific feeling that time is your enemy. Not time meaning days or weeks, but time meaning every second your hands are on the controller. While this is suitable in some places, it's not suitable for someone who wanders around a beautiful island paradise looking for nice places to fish.
This may not seem related to the shift in focus, but I think it is: it's very difficult to give characters enough time to be themselves in any situation where a sentence takes five minutes to say. The pressure, in addition to being unsuitable for the setting, is also unsuitable for the dynamic. It actually makes it almost impossible for the gameplay to really revolve around the characters at all.
How would you fix it? How would you make a game that was "Swiss Family Robinson" style or "Blue Lagoon" style, but make it focus on the characters? Without being shallow?
The first thing I would do is ditch the ticking clock. I would make the clock not run at all, unless you were "fast-forwarding" to get to the end of a job. For example, if you were fishing, time would pass. If you were traveling from the jungle to the mountain, time would pass. If you were building an expansion to the tree house, time would pass. But if you're walking around or chatting with people, time does not pass.
This gives an impetus to walk around and talk to people. The second half of the equation would be to make walking around and talking to people worthwhile. Walking around is easy: by walking around, you discover various resources that can be used.
Talking to people is harder, because we've all gotten so used to dialog trees or flat responses from NPCs. Going up to someone and clicking the A button on them really doesn't put the player in any kind of control, doesn't give him any kind of choice.
It would be interesting, I think, to let the player take "snapshots". Snapshots of people, things, places - even ideas. Then the player could present a snapshot to a character and they would respond in a more-or-less adaptive, unscripted manner. A context engine and a little bit of elbow grease would make this work okay.
For example, let's presume Robinson. You're the father, and you show your middle son a picture of your youngest son. The context engine kicks in. If they've been fighting, the middle son says something like "fine, fine! I'll make peace with him." If the youngest son is missing, the middle son replies, "right, I'll go find him!"
Show him a picture of a place, maybe he'll comment on something in the area, or maybe he'll promise to go out and take a look, or maybe he'll say, "right, the berries in that area are good, I'll go get some tomorrow."
Show him a picture of a thing, maybe he'll come up with a new way to use it, or maybe he'll go out and get some. Maybe he won't have anything interesting to say about it, but now that it's on his mind he'll show the picture to whoever he happens to talk to, and since he's a different person than you, maybe he'll get a different response...
The idea is that the family (or any group of people) is as complicated and adaptive a system as any mad genius' underground lair. Building and maintaining it is just as important as building and maintaining your tree-castle.
I dunno, how would you do it?