Opinions sought about door differentiation

That was Savoir-Faire, Emily Short’s XYZZY-award winning game and another great example if Amanda wants to take a look at it.


…prompting the player to pick the first listed one, then the second listed one, followed by the third, and so on. Because how else are you going to remember whether you tried the “nondescript” or “unassuming” door, particularly if you save and quit and go do something else in the middle?

Maaaaybe you’d get players going for the plain door first, just because it’s the easiest to type. But making multiple choices identical except for a single, unique adjective which is synonymous with all the other adjectives seems like an invitation to lawnmowing to me. Or at least I’m fairly confident that’s how I’d approach it as a player.


Without anything differentiating one door from another, no matter what their description is, you’ll have a player lawn-mowering through his options.


I made that up for the post. One of my use-cases for the Easy Doors extension was recreating the hallway full of doors in Alice in Wonderland.


If you use the code I posted above, OPEN NORTH will work fine without disambiguation. (Inform will prefer any other thing over a direction, for any action that doesn’t apply to directions specifically.)


Right. Which implies that naming them “unassuming door”, “nondescript door” and so on doesn’t satisfy the stated criteria.

If the goal is that the player should choose an arbitrary door with no particular cluing/hinting, then that probably needs to be implemented as an actual behind-the-scenes mechanic implemented as part of the door logic (that is, the door descriptions might be in static order but the destinations are randomly assigned when a door is first opened).

Alternately I guess you could pseudo-clue the doors in a way that just turns out to be irrelevant. Like provide elaborate and evocative descriptions of the doors and the sounds, smells, and vibes each are emitting, which will (presumably) engage the player more than “blue” or “north” would, leading them to coming up with a rationale for picking one door over another other than just grabbing the first one listed. This will probably only work if they think they’re just going to get one bite at the apple, and if they come back later to make a second choice if they’ve subsequently discovered that all the atmospheric nonsense didn’t really hint at what was behind the door, they’ll probably just fall back on lawnmowing.

I don’t really know what the underlying puzzle is, so it’s difficult to come up with a specific recommendation for shaping (or elaborately failing to shape) the player choice. For anyone interested in this sort of thing in the abstract, magic (as in pick a card and so on) offers a lot of strategies for approaching this sort of problem.

1 Like


If you want the choice to be arbitrary (instead of being a puzzle to be solved or something like that, you just want them to freely pick a door more or less at random) then you would probably have to implement this in the code for the door. That is, instead of the choice itself being random, the results are random. So if you have four doors, and four destinations behind them, then 25% of the time, the north door leads to destination a, 25% of the time it leads to destination b, 25% to destination c, and so on. Of course unless this is a magic door or something that only applies to the initial reveal.

Basically if the result you want is players picking a random door (such that roughly 25% of players pick each door as their first choice, the distribution is equally uniform for the second choice, and so on) then you’re probably going to have to handle that via game mechanics. Because almost any “honest” selection mechanism presented to the players this way is almost certainly going to be biased.


Thanks, everybody, for all this lively discussion. It has convinced me to change course a bit and instead of having a door be chosen randomly by the player, to have puzzles connected to the doors (not directly, indirectly) and whichever puzzle the player solves first picks a door for them. Now I just have to make sure that the puzzles are of roughly equivalent difficulty.


I am not sure which system you are using but surely it would be also possible to have door descriptions reflect either the PC’s prior knowledge of the game world or dynamically reflect the PC’s in-game history when they are supposed to have no prior knowledge.

In the former case an example might be ‘the door to the kitchen’ in the PC’s house. In the latter case something like the ‘north door’ becomes ‘the door to the great hall’ once the player has previously used that door and would know where it leads.


It’s literally the same problem people have with menu-based conversation trees, except this time it’s worse. :smiley:

1 Like

I’m partial to color or material, but to me the more important thing is consistency–if there’s a “north” and “south” door in one room and a “red” door in another, there better be a good reason that door’s red, you know? But if you’ve gone to the trouble to make a “plain” door and a “worn” door in one room, now I want unique adjectives on them all (even if it might get a bit silly).

1 Like

one door
two door
red door
blue door


It would be cool if systems had a function to easily shuffle choice order on a page, like within a tag to discourage sequential lawnmowering.

[Go to the store] 
[Go to the beach]
[Go take a nap]

I feel like a [rocket scientist] because I just figured out how to remap my dead right bracket key to left curly-brace. [cue angelic choir]


I know you like Hybrid Choices; it looks like all one would need for that would be to sort the table of currently available choices in random order (and adjust the index column appropriately).


Yeah, Hybrid Choices does all kinds of amazing things if you learn it and figure out its intricacies and are also tricky with I7. I was able to use Inform’s standard text variation to do some fun things with extensively updating and changing choice labels in Fair - so encountering the same list had variation because the label text could cycle.

I know the effect of that, for me, would just be a lot of confusion. If I come to a room that has an oak door, a pine door, a maple door, and a cherry door, I’m going to mentally sort them as “door 1, door 2, door 3, door 4” and remember them by the order, not by their names. If that order then changes up randomly, well…

1 Like

Yeah, order or compass directions would really be by far the simplest thing for me – plus if there are that many closed/locked doors in one room, it already feels a little game-y, and if for some reason I need to believe that the architect decided to make them all out of different materials or paint them different colors that would probably seem kinda silly to me.