Compass directions are concise, reversible, and encourage
the user to build a visualization of a space. For a game
system in which determining where an exit goes requires a
function call which might print a message, it is decidely
non-trivial to write a generic path-finding system. Some
games have done it on their own, but it is not in general
perceived as worth the effort. Having "GO <location>" do
path-finding but only move you one room per turn does not
do enough to make players lives easier; a "GO <location>"
which "teleports" directly does not require path-finding.
> (Where's the compass? What sane creature moves around an
> apartment in this manner?)
I've played a game called 'The Mystery of Arendarvon Castle' on the
C64 years ago which used a relative system of movement: Typing
'forward' would move you in the direction you would be facing. 'turn
left' and turn right' would change the way you faced.
Since it was one of the first adventure games I ever played it made
sense to me. But 17 years later I have yet to find another game that
uses this system.
The compass makes it much easier to map the game. A relative system
makes it very easy to get lost without good reason.
Compass directions have also become an convention, a habits, which is
all but expected by adventurers when they play IF. Even though
'Inventory' is a rather strange term to check what you have in your
pockets, every player knows that 'i' will list your possessions.
Also, automatic path finding kind of takes the fun out of exploring,
me thinks. I played a game called 'Knight Orc' by Level 9 which used
this system. If you knew the name of a location you could just type
'go <location> and the game would take you there, displaying all the
locations you passed through as if you just walked through them. The
program kept track of which locations you had truly visited so you
could not 'guess' locations to bypass puzzles. With this system you
did not have to make a map, just a list of location names. But thanks
to this the geography of the game became mostly irrelevant. I felt
less immersed in the game world because the computer did all of my
walking for me.
Just my 2cts worth...
It's a bird...
It's a plane...
No, it's... Gadget?
To send mail remove SPAMBLOCK from adress.
>Karl Filenius posted:
>>Being a newbie to IF I find some things, that seem to be
>>integral parts of the genre, strange and surprising. One of
>>those things is the use of directions for movement. It's not
>>just that I find typing "s,w,e" when wanting to get from the
>>living room to the bathroom ludicruous and unrealistic (Where's
>>the compass? What sane creature moves around an apartment in
>>this manner?), it's also the fact that most (all?) games seem
>>use compass directions as a pretext for not implementing even
>>the simplest path finding algorithms. What's your opinion on
>You can take a look at a recent example of no-compass
>navigation in action:
Erm, recent? Huntdark is a version of Hunt the Wumpus, isn't it? A
game that is almost as old as ADVENT...
It's faster to type "n" than to type "enter bathroom" (for example), but if
you're in your bedroom in Paris and you type "go to New York," I think there
might be a lot of story that you miss along the way.
Plenty of people who (today) can write a simple text adventure are not
experienced computer programmers, and thus would likely have no idea how to
implement the path-finding algorithim you mention. But since the textual
game arena is spatial, requiring the player to navigate through it with
compass directions applies an organizational scheme that, for many people
who think textually rather than spatially, is somewhat less chaotic than,
for example, the network of caves in "Wumpus" (which was supposed to be a
dodecahedron, if I recall correctly). YMMV.
Perhaps the "real" reason is simply that the creator of the first IF game,
Will Crowther, was an avid caver and explorer, who really did use a compass
while exploring caves. (I don't know how he got around his apartment,
Since then, the use of directions has become a convention, just the reading
of pages in sequential order is a convention in novels, or the use of "going
blurry" in a movie and having harp music is a signal of a flashback, or the
representation of motion by clicking arrow keys on the keyboard for some of
the early graphic comptuer games.
There are quite a few IF games that don't use the compass... some sections
of "Photopia," for instance, and Galatea, where the exploration of a
simluated textual space is not important to the plot.
By all means, if you would like to create a compass free game that presents
a better way to move around, then, go ahead! It might be, however, that
because of the time you spend working on algorithms, you don't have as much
time to devote to creating memorable room descriptions, characters, plot,
etc. (But go ahead and surprise us all, if you wish!)
Gregory Yob's Hunt the Wumpus appeared around 1972. Will Crowther says he
created Advent in 1975 "give or take a year".
And Huntdark isn't simply a port of Wumpus. Try it!
(Sort of a spoiler for Hunter, In Darkness)
>Erm, recent? Huntdark is a version of Hunt the Wumpus, isn't it?
No. It's inspired by "Wumpus", it has the same premise, but apart from
And the navigation in "Hunter" is nothing at all like that in "Wumpus"
(in one game, you type commands like "up" or "right", in the other,
you type the number of the room you want to go to).
>A game that is almost as old as ADVENT...
Four years older, actually.
<PLUG HREF="http://www.cowlark.com/IF/bazic.zip">Hunt the Wumpus</PLUG>
+- David Given --------McQ-+
| Work: d...@tao-group.com | "The only thing to prevent what's past is to put a
| Play: d...@cowlark.com | stop to it before it happens." --- Sir Boyle Roche
+- http://www.cowlark.com -+
"s.nw.e" is fewer characters to type than "go to kitchen". The realism
doesn't bother me because it's an established convention; I never
wind up using higher-level movement commands even when they're provided.
"And Aholibamah bare Jeush, and Jaalam, and Korah: these were the borogoves..."
* Make your vote count. Get your vote counted.
> Plenty of people who (today) can write a simple text adventure are not
> experienced computer programmers, and thus would likely have no idea how to
> implement the path-finding algorithim you mention. But since the textual
> game arena is spatial, requiring the player to navigate through it with
> compass directions applies an organizational scheme that, for many people
> who think textually rather than spatially, is somewhat less chaotic than,
> for example, the network of caves in "Wumpus" (which was supposed to be a
> dodecahedron, if I recall correctly).
Yes, it was.
> Perhaps the "real" reason is simply that the creator of the first IF game,
> Will Crowther, was an avid caver and explorer, who really did use a compass
> while exploring caves. (I don't know how he got around his apartment,
Note that the original Adventure *did* have some support for typing
the name of a room to go there. (Was it limited to nearby or adjacent
rooms?) But use of such commands didn't catch on among Adventure
players; the MIT folk didn't implement them for Zork, and they've
basically been gone ever since.
Note that in _Hunter, in Darkness_, there isn't a whole lot of
movement. A large percentage of that is "up" and "down", which are
standard IF directions, and the rest is "forward/left/right/back" --
which rely on the IF habits of compass directions, just renamed.
The game "Battlestar" uses a similar system. Directions are "left",
right", "ahead", and "back", and exits are given relative to the
direction from which you entered the room. I don't see much advantage to
such a system. Unlike a more abstract "go to <location>" system, it
actually demands *more* of the player's spatial visualization skills than
the usual system, as the player has to keep track, not just of position,
but of orientation. I guess it would make the maze-of-twisty-little-
passages experience more realistic, but other than that, compass
directions are a welcome piece of abstraction compared to this.
But what about the more abstract route of "go to <location>"? I agree
that making the player walk through locations gives a better sense of
geography, and a convenient way to provide for exploration. But that's
not important in some games, particularly small games where all locations
can be easily reached from one another, or games taking place (in whole
or in part) in an environment that should be familiar to the player
There have been a couple of games with built-in path-finding (Knight Orc,
T-Zero), several with abstract room-based movement (Nord & Bert, Moon
Mountain, In The End, Shade), and quite a few now that avoid the issue by
taking place entirely in a single room.
I wouldn't call it "a version of Hunt the Wumpus". I'd call it an
original work in the Wumpus milieu.
Also, if memory serves, Hunter didn't really provide relative directions.
It just renamed the compass directions "left", "right", etc.
On the other hand, "go to kitchen" is fewer characters to type than
"sw.e.s.s.e.e.ne.e.s.s.ne.n.w", and is easier to remember. Large games
sometimes provide teleportation devices, such as the pendant in Mulldoon
and the pigeon statue in Zork Zerok, to reduce the necessity of long
strings of movement commands, and I certainly use them. Would I use
abstract movement if it were provided in these games? I don't know.
Without an object to remind me that I have the option, I might not.
Yes it did. It just did not allow you the option of turning.
Neil Cerutti <cer...@together.net>
> > "s.nw.e" is fewer characters to type than "go to kitchen". The
> > realism doesn't bother me because it's an established convention;
> > I never wind up using higher-level movement commands even when
> > they're provided.
> On the other hand, "go to kitchen" is fewer characters to type than
> "sw.e.s.s.e.e.ne.e.s.s.ne.n.w", and is easier to remember.
But also, in another post:
> I agree
> that making the player walk through locations gives a better sense of
> geography, and a convenient way to provide for exploration. But that's
> not important in some games, particularly small games where all locations
> can be easily reached from one another, or games taking place (in whole
> or in part) in an environment that should be familiar to the player
Bit of a double-bind, you see. If the game is small and every room is
one or two moves away from every other room, I'm going to want to type
one or two letter commands to move around. If the game is large, the
sense of geography becomes important.
And I really *do* find it pretty easy to type "sw.e.s.s.e.e.ne.e.s.s.ne.n.w",
if the center portion of that is a major transportation artery in the
game. By the time I've explored, I've learned large chunks of routing.
>>Also, if memory serves, Hunter didn't really provide relative
>>directions. It just renamed the compass directions "left",
> Yes it did. It just did not allow you the option of turning.
Nor did the game ever turn you. In any given room, "right" always
refers to the same exit. In that sense, the directions really were
absolute, not relative -- they were absolulte compass directions
>But what about the more abstract route of "go to <location>"? I agree
>that making the player walk through locations gives a better sense of
>geography, and a convenient way to provide for exploration. But that's
>not important in some games, particularly small games where all locations
>can be easily reached from one another, or games taking place (in whole
>or in part) in an environment that should be familiar to the player
>There have been a couple of games with built-in path-finding (Knight Orc,
>T-Zero), several with abstract room-based movement (Nord & Bert, Moon
>Mountain, In The End, Shade), and quite a few now that avoid the issue by
>taking place entirely in a single room.
For me personally, the big draw to adventures is the exploration. Go
<location> like in Knight Orc made me feel very disconnected from the
setting. Go Kitchen from the living room is another story, but even
that I don't use and find confusing. I like my maps nice and
traditional with compass points (call me conservative ;-) so I at
least know *where* the kitchen is, relative to the living room. It
makes it feel more real to me.
That is an unimportant implementation detail. ;-)
Neil Cerutti <cer...@together.net>
*** Your offworld investment in artificial dumbness paid $150 in
I think that compass directions are one of the places where "realism"
has been chucked in favor of usability. Quite simply, in a textual
respresentation of a world, using a fixed navigational system is *so*
much more useful than the alternatives. Consider:
Using compass directions...
You stand in the first room. Lots of interesting stuff is going on,
but we'll overlook it because this is a contrived example.
Doors lead north, south, northeast, east, and west.
A relative direction system (one wherein you navigage by "forward"
"backward" "turn left" "turn right") would be more realistic. Now, it
would also require retooling room descriptions to take into account
the direction the player is facing. That's more work for the author,
but it might be worth it, since it's not that much more work.
But think about what it's like for the player. How far does "turn
left" turn? 90 degrees? Isn't that unrealistic? As far as the next
exit to the room? THat's better, but now the player starts to lose all
sense of location:
You stand in the first room. Lots of interesting stuff is going on,
but we'll overlook it because this is a contrived example.
There is a door in front of you, one to your right, and two to your
left. Craning your neck, you can see one in front of you.
# TURN LEFT
There is a door in front of you, one to your left, one to your right,
and two behind you.
How long before the player gets disoriented? The player goes through a
sequence of rooms. How does he get back? In the standard paradigm, he
just reverses the compass directions. But in the relative direction
paradigm, there's no straightforward way; he might have to turn in the
opposite direction more or fewer times in each room to get the
direction he desires (Before you say it, no, a "backtrack" command
wouldn't help; it would only solve a small subset of the related
problems of which this is an example)
If you use relative directions, *every* area with nontrivial geography
becomes a maze, and we no longer really want that, do we?
Room-based navigation ("go to the kitchen"). Not bad. Pathfindinding
algorithms should be easy enough to implement or fake. But... WHat
happens to the player'ssense of spacial relations? What if something
intervenes, blocking the player's progress? It's unreasonable for the
move to abort, so the player is liable to end up in some room he
didn't want to go to, with no idea how he got there. What about rooms
the player's never been to? WHat if the pathfinding algorithm takes
the PC a way that the player wouldn't have thought of, or through
rooms he hasn't been in? What if something interesting is along the
way and the player wants to take a look?
SImulating reality is not always a laudable goal. In an area such as
navigation, in *reality*, people get lost quite easily. The evidence
we have seen so far suggests that a compass based system keeps the
player's chances of gettign lost relatively close to his chances of
getting lost in the real world. Other systems tend to greatly
increase his chances of getting lost.
On the other hand, if the game is small, keeping the movement commands
short matters less, because you're going to be entering a lot fewer of
them. Is the convenience worth the loss in mimesis?
Your own work, "Shade", has been mentioned repeatedly as an example of a
game without compass directions. You could have easily attached compass
directions to the kitchen and bathroom. You chose not to. How many
people have complained?
I was extremely careful to make movement commands totally optional in
_Shade_. You can play the entire game without typing "enter kitchen"
or "enter bathroom", or [*] "get up". The game always handles that
stuff for you.
I think that if I had not made that automatic, I would have gotten a
lot of complaints. Naturally, I have no proof of this assertion. :-)
But it would have pissed *me* off, to have to type movement commands
in that sort of game.
(Note that, contra your assumption above, you really do move around a
lot in _Shade_. Try counting the number of implicit "go to"s that
occur in a runthrough of the game.)
[* Except for the first move, where you have to type "get up"
explicitly before you can do anything else. This is a bit of
scene-setting / character-building, and I kept it minimal.]
>(Note that, contra your assumption above, you really do move around a
>lot in _Shade_. Try counting the number of implicit "go to"s that
>occur in a runthrough of the game.)
Hmm. Since basically everything is always in scope in _Shade_,
I'm wondering whether it would have worked to have had just one
My favorite moment  in Shade was when, after looking at the
contents of the fridge, I attempted to stop that ongoing activity
by typing "CLOSE DOOR".
This happened to work, although if we had a more sophisticated
system of timing the overrides and implicit actions, it's quite
possible it wouldn't have. 
1. Well, not really. Favorite according to some metric, though.
2. "DOOR" was interpreted as referring to the front door, which
was outside the kitchen nook and hence required departing the
kitchen nook; leaving the kitchen nook caused the refrigerator
door to swing shut automagically; all to be followed by the
observation that, in fact, the front door was already closed.
Like others, I find it most convenient to use normal compass directions in
most cases. The only time they become a nuisance is when I want to go a long
way and I don't remeber the exact sequence of moves I need to make -- this is
when I want to be able to "go to X". This doesn't happen in small games,
because in a small game there's never far to go. In a game large enough that
"go to X" would be useful, I am generally not able to remember the names of
Also, I think that in many situations the game world can be broken up
into pieces small enough that using compass directions is no problem. In
situations where this really isn't feasible (because of the nature of the space
being modelled), I'd say the trick is to have a few rooms with simple names
that can serve as "hubs", so if the player doesn't remember the name of the
Department of Interior Decorative Arts and Sciences but knows it was near the
Information Desk, he can "go to information desk" and find his way from there.
An important thing here, though, is that the Information Desk shouldn't be put
there just to be a hub; it should actually be geographically positioned to
serve as a hub, so that the player will come to think of it as a hub and will
remember which rooms are in its vicinity.
--OKB (Bren...@aol.com) -- no relation to okblacke
"Do not follow where the path may lead;
go, instead, where there is no path, and leave a trail."
For those who would nonentheless like to have such an algorithm in
their games, there is a TADS implementation called dijkstra.t on the
IF archive. I don't know if there's a corresponding Inform
implementation; if not, there should be.
>Note that the original Adventure *did* have some support for typing
>the name of a room to go there. (Was it limited to nearby or adjacent
I don't think it was limited to nearby rooms, but my memory is a bit
hazy on this point since I never used those commands myself.
>But use of such commands didn't catch on among Adventure
>players; the MIT folk didn't implement them for Zork, and they've
>basically been gone ever since.
I think one reason people didn't use these commands in ADVENT
was that they were implemented in a rather confusing way.
IIRC, you can go to certain rooms - but only to a small number of
rooms, not to any room - by typing the name of that room. But this
is rather arbitrary - how do I know that I can go to room A but not to
room B in this way? And IIRC it was also rather poorly advertised -
it was mentioned in the docs but it remained an obscure feature.
And there is also the confusing fact of the magic words, like XYZZY,
that actually teleport you to a different place. In the game world,
there are rather different from the shorthand movement verbs, which
are just commands to walk to a room the usual way. I've seen evidence
of this confusion here on this newsgroup, where people have been
referring to a large number of magic words that would teleport you
around but which didn't always work. Those words weren't supposed to
be magic words at all...
>>(Note that, contra your assumption above, you really do move around a
>>lot in _Shade_. Try counting the number of implicit "go to"s that
>>occur in a runthrough of the game.)
> Hmm. Since basically everything is always in scope in _Shade_,
> I'm wondering whether it would have worked to have had just one
In terms of the Inform implementation, that's what I did. The kitchen,
bathroom, futon, and desk are containers in the room; the shower is a
container in the bathroom.
Then I messed obsessively with the library until everything Worked
Right. It wasn't pretty. (Hm, I guess I should put together a
one-year-later source code release, like I did for _Hunter_.)
But the point of all that implementation was to simulate a standard
multiple-room game, with two additions: everything is in scope, and
movement is automatic. I didn't want a standard one-room game
experience; that didn't give enough of a sense of geography.
Stuck in the Library with a Globe - An Interactive Geography Lesson
> > Plenty of people who (today) can write a simple text adventure are not
> > experienced computer programmers, and thus would likely have no idea
> > how to implement the path-finding algorithim you mention.
> For those who would nonentheless like to have such an algorithm in
> their games, there is a TADS implementation called dijkstra.t on the
> IF archive. I don't know if there's a corresponding Inform
> implementation; if not, there should be.
There is also a related module, trackActor.t, that allows you
to set NPCs in motion, giving them location nodes and letting
them use dijkstra.t to calculate their own best paths from
room to room. I don't know that I've yet seen a released game
that used these modules, but they seem to work well.
> And I really *do* find it pretty easy to type "sw.e.s.s.e.e.ne.e.s.s.ne.n.w",
> if the center portion of that is a major transportation artery in the
> game. By the time I've explored, I've learned large chunks of routing.
And yet, in another thread, he says:
> I have a preference
> (when appropriate) to have the game fill in the details [...]
> -- there's no need to make the player
> walk through a mechanical series of commands once he's proven he's
> gotten the point.
So don't get the idea that I'm the only one here capable of expressing
seemingly contradictory preferences.
The reason these feel different to me is that the former *becomes* a
single command, in my head and (more importantly) in my fingers.
I believe at least one pre-King's-Quest Sierra On-Line game used this
approach, but only in a maze.
John W. Kennedy
(Working from my laptop)