The parser in Adventure is harder to use than most interactive fiction parsers, because this is the first of its type and fairly primitive. This is almost immediately evident when a user makes it to the building. One obvious thing to do would be to input “enter building;” doing so, however, returns the message “That’s not something you can enter” (Crowther and Woods). Likewise, one puzzle requires the user to sic a bird they have captured on a snake, but the command “throw bird at snake” does not work. Instead, simply “throw bird” will take care of that problem. Another puzzle requires users to realize that a certain object (a wand) is useful on a certain object (an uncrossable gap) in a certain way (by issuing the command “wave rod,” not “use rod”); these are probably not ideas that are immediately thought of by users. Another aporia that exists is the game’s geography. As the annotation notes, the game doesn’t utilize Euclidian geography: “a north exit from one room to the next doesn’t necessarily imply a south exit will return you back to the original room” (Jerz). This makes map-making, a standard interactive fiction tool, somewhat difficult. —The Problem is Choice: Aporia, Epiphany, and Conflict in Interactive Fiction Endings (ENGL 668K: Digital Studies (University of Maryland))
Some good points, but the “enter building” problem is an issue with one particular edition of Adventure — the Inform port by Nelson (1994), after Ekman and Baggett (1993). It’s not accurate to ascribe this particular difficulty to Crowther and Woods. And the original Adventure parser was only built to recognize a noun and a verb (in either order). The Inform environment that enabled the 1994 port permitted the entry of more complex sentences, but the game itself doesn’t recognize those complex sentences. since the game was written for a parser that only recognized a noun and a verb. And the practice of mapping an adventure game on an Euclidean plane only became possible after the Adventure genre took hold. Another classic cave-explloration game, Hunt the Wumpus, took place in an icosahedron, if memory serves.
While I’m quibbling with this passage about a digital text I know well, overall I thought this posting did a good job applying Aarseth’s concept of ergodic to the chosen examples.
Another geek correction: Frank, an icosahedron has 20 triangular faces, but only 12 vertices (rooms), as opposed to “lots more”. This is a consequence of Euler’s formula for regular polyhedra (V+F=E-2) and the fact that the icosahedron and the dodecahedron are “dual” to each other (swap the faces for vertices and they switch – but both have the same number of edges). Go grab your old bag of D&D dice – mine are right here in my office – and check it out!
My (very old and worn) copy of Ahl’s “More Basic Computer Games (1979)” has a Wumpus variant with 6 different caves: the dodecahedron, a mobius strip, a “string of beads”, a hexagonal net on a torus, a dendrite, a set of one-way only tunnels, and a “free-for-all” (that you design at the start of the game, dynamically). It’s neat :)
Pwned!
Thanks, Frank. I should have paid more attention to that little voice that whispered to me that I’d better add “if memory serves”. I was thinking that five exits sounded like too many. The general point is that Adventure’s divergence from strict grid was not new, although Adventure’s imitation of the real twists and turns of a natural cave was new. Wumpus did have a rigid structure that was part of the game (to those who noticed it and learned to exploit it).
You should have known some geek would correct you :)
“Hunt the Wumpus” was set on a docecahedron with 20 “rooms” at the vertices, but only 12 imaginary pentagonal faces. An icosahedron has 20 triangular faces and a lot more rooms.
The easy way to tell is that each room in “Wumpus” has three exits. If it were an icosahedron, each room would have five exits.
I still remember laboriously typing in a BASIC version of Wumpus back in the late 1970s. After all that, the game itself was pretty dull.