The compass, location descriptions and mimesis

Everyone makes excellent points.

However, having had time during my horribly long commute to think about this, I think I phrased the question wrong. Or, maybe the conversation needs widening.

There is a lot of discussion on ‘mechanics’ and Andrew makes the excellent point that this has been discussed since pretty much forever. I vaguely remember being part of a discussion with Gareth Rees (whatever happened to him?) and others around compass directions and whether or not there was a better approach back on r.a.if in the nineties (I can’t remember the conclusion - probably ‘no’) However, I’m talking about the text. The text is all there is in parser IF, and, unlike any other narrative text which needs to establish a million things like voice, context, narrative, plot, character, etc…etc… IF text is also required to cue the player/reader - to do stuff - to interact with the text in a way that drives the narrative. Which includes giving the player enough information to move around the game.

Andrew made the point that “I guess my view is that in traditional IF, the compass directions are not competing with the prose, nor distracting from it.” but I disagree slightly. For a ‘traditional IF’ game in which puzzles and goals are everything, fair enough. But in a game in which narrative voice and story are at the forefront, maybe so.

It feels like the ‘structure’ of an IF narrative has remained the same forever. Take a location description. In general, I would argue that 90% of location descriptions take the form of :

Going from a Scott Adams ish


Exits : NSEW

You can see: tree.

…has the way the author present the text to the player really changed in the last 30 years? The argument is: For us, this works. We’re used to it. We’ve always done it that way. Doesn’t make it right.

My original question was : ‘How do I elegantly embed a directional mechanic in the text without it being clunky and breaking player immersion, while still giving the player enough information so they don’t get irritated’

Given Twine and it’s ilk, which really does enable complete integration of game mechanic and text, I think the discussion really ought to be, for parser IF, ‘Does the traditional structuring of narrative elements within IF need a re-think?’ ‘Is there a way of more closely integrating game mechanic within the text?’

I don’t know the answers to this, of course. But I thought it interesting.


You can’t say that story and narrative voice are the forefront but, oh yes, the player has to run around this complicated building in an adept way. But that’s not important. Story and narrative voice! But if you get lost in a hallway you’ll never finish.

That’s fundamentally incoherent.

What Twine shows us is that if you don’t want to think about navigation issues in your game, take out all the navigation. That’s a valid choice. But if it’s in there, it’s part of the design.

The other thing to add is that interactive story telling is still (in my mind) in its infancy. If you can envision it, build it, and find happy “customers”, then it’s a good thing. We’re still experimenting with all aspects of the medium and although compass-directions are currently the standard for parser-based IF, it doesn’t mean future experiments won’t change that.

So experiment. Just keep trying crazy stuff and see what sticks.

The traditional world model of parser games involves objects in rooms, connected to one another by directions. The textual conventions that we have arose in order to communicate that model: often a room description in an Inform 7 game looks fairly similar in its outlines to the code for that room. So the conventions are traditional, but not arbitrary.

Of course there are exceptions already: there are one-room pieces or heavily plot-driven things where the player doesn’t do a lot of room navigation but is just moved to new locations as the story demands. Those designs cut down on the need to communicate space, because spatial navigation is in fact not a thing you’re asking the player to do.

In any case, I’d be the last to discourage experimenting if you have some ideas you want to experiment with, but as a design principle I tend to hesitate when I see a suggestion that boils down to “let’s work harder to obfuscate the underlying model of this game.” In my experience, that much more often leads to player confusion and pain than to a good and coherent experience; I typically find that players are willing to look past a fair amount of scaffolding if the result is a coherent story in which they have a strong sense of agency.

So for that reason I personally would be more interested in exploring not “can we narrate this same world model in a different fashion?” but “can we do parser IF that does not make use of that type of world model? what types of story would we tell if we did?” And in fact I think we are seeing a little experimentation in that direction, and that said experimentation does perhaps reflect the influence of Twine. I’d point to Castle of the Red Prince and Toby’s Nose and Enigma here.


I’d again urge you to check out Beyond Zork, then.

You might also like trying Guardians of Infinity, which is this crazy multi-window real time commercial game from the late 80s.

Some T-Zero room descriptions from the opening:

I’m fairly conversant in compass directions in my everyday life (trained as an architect) but for some reason my mind always orients whichever direction I go first as “north” unless there’s an accompanying map. (There’s a neighborhood in the city I’ve lived in for ten years in which I still consistently get north/south reversed! But just that one.) So I always kind of hope for games to have a valid north exit out of the first room.

The compass direction alternatives I love are “go to [name of room],” or hyperlinks with the same meaning, and clicky maps. Those are both pretty great.

I can’t see that it does… it doesn’t seem to be an issue, ever. In practical terms, I mean - outside of conjecturing.

We do have the means to do away with compass directions in the text, though. Sure we do. Automaps, compass roses on the status bar. But it’s so practical as it is, and navigating a game isn’t a fanciful narrative event - it’s practical. Like zarf said, players need to be adept at movement. They should be able to have a mental map of their surroundings in some way, and the absolute movement of the cardinal directions is perfectly suited to this, as it was when Woods first used it to navigate a colossal cave in a grand adventure.

And if, in a certain game, or a certain part of the game, movement is indeed fanciful and part of the broader narrative, rather than a practical thing, then it’s easy enough to adapt the game or sections of the game - but the author will be the first one to say “Ok, for THIS section, instead of compass directions, we’ll have something different”.

I mean, we have a convention and some games go against that convention and the vast majority of games embrace it. That’s what a convention is. It’s a good, practical convention. We can’t really do away with solid groundwork like that - isn’t it good enough that there are plenty of ways not to use compass directions if the author doesn’t want to?

Personally, I find navigating by landmarks extremely confusing, and as soon as Blue Lacuna gave me a compass I never stopped using it. However, in Hadean Lands and Counterfeit Monkey, I found myself making extreme use of the “GO TO” verb - which is merely a shortcut, though, and not a substitution for a navigational system.

I mean, if you have to navigate a game where instead of single-letter directions you have to type the full name of the place you have to go to for more than ten turns, or if you have to keep turning yourself to face something, you go bonkers, really.

Like MTW, I’m a bit confused by this statement. What would be “right”, then? If there is such a thing in this context.

My favored nav system depends on the context. In caves and forests and other wildernesses larger than three rooms I think a compass is appropriate. I might use a real compass in real life in that situation. But indoors, room names are easier. When in my character’s own home, KITCHEN is more natural. The only exception might be indoors of a large unfamiliar space in which compass directions are OK until I learn the rooms. Or if I’m going to be chased through an area, I would like to know what’s a dead end and what isn’t. But beyond that, navigation by itself rarely tells a story, so I’d prefer not to need a map, mental or otherwise.

I only liked compass directions in Bronze because of the un/visited-colored compass rose, but I started using room names pretty soon afterward because I couldn’t remember the whole layout. It was just too large.