Lately, I've been replaying some of the Zelda games Nintendo released for older Gameboy handhelds, specifically Link's Awakening and the Oracle titles. They're ingeniously designed, and part of what I find fascinating about them — particularly in light of the 15 years of videogame history since — is the way that they deploy the basic conventions of the franchise in order to build a sense of place. The original Legend of Zelda was all wilderness, and its sequel on the NES, The Adventure of Link, shifted perspectives radically to round out Hyrule as a kingdom, but the model for the world-building in that early Gameboy installments is A Link to the Past, first released on the Super NES in 1991.
With all of that swimming around in my thoughts, it's hardly surprising that a recent Gamasutra series on A Link to the Past's level design caught my attention, and particularly the latest post on how the game manages progression. What stood out to me about it, though, was the language it used to frame that design. The player is "railroaded from [Link's] home to Hyrule Castle," then "locked down" until they've achieved a goal in the castle. The world "opens up" by more than half after that, but even so, the player can't interact "in a meaningful way" with large portions of it. The usable items found in dungeons are mentioned mostly as tools to "access" the rest of the map.
Most of this falls under the general heading of "gating," and it's part of the language we've developed to talk about videogames that have typically sold under the heading of adventure. The "metroidvania" genre, for example, is characterized by a gated world that gradually opens up as the character "unlocks" new skills or tools. The entire open-world style of adventure game design is formulated as a contrast to such "closed" worlds, allowing (in principle) unfettered access to an elaborate world without the sort of step-by-step progression dictated by the conventions of older, more technologically constrained videogames.
The terminology of gating is virtually sanctified by tradition, but it bears with it a number of implications that I've come to think are wrong-minded. One is the way that it throws emphasis on accessibility as both the goal and reward of play. Descriptors like "railroading," and "locked down" seem to suggest that the ideal state of the in-game world is one that would allow the player to travel freely to any point on the map. If the designer's task is the judicious placement of obstacles to limit the player's access, then the focus of play is on uncovering the world in full.
Playing the Zelda games, I do find myself wanting to see the entire world, but I don't play as though unfettered access were an end unto itself. It isn't that there are obstacles standing between me and a God's-eye view of the world. Rather, those obstacles are opportunities to play. Both the narrative goals and territorial ambitions are rationalizations — just as I have no actual vested interest in the rescue of Zelda or the waking of the Wind Fish, the game's real estate is not real enough to motivate me. I play because there is something satisfying about exploring the behaviors provided as solutions to those obstacles. The progressive revelation of the world is an effect of that process. I want to see it in its entirely because that represents the totality of play the game affords.
To that end, when we talk about controlling an adventure game's "flow," gating may be a misleading metaphor. Reserving the tools needed to meet certain challenges need not be about restricting access. It can be, rather, a way of directing attention to challenges that the player is already equipped to meet.
Granting that perspective, the conventions of adventure game design will, in the best of circumstances, function not to frustrate access, but rather to parcel out the game's principle activities as a sequence of discrete, yet overlapping challenges. The conceit that those challenges occur in "a world" is built up by arranging that sequence into a coherent spatial unity. The more creative expression the designer brings to that structure, the more distinct it feels as a place. Thus, the Oracle games give the territory they cover dimensionality by depicting the changes that occur as the player moves back and forth through time. That dimensionality is an aspect of play, as well, and some of the series' challenges can only be addressed by moving from one season or age to another.
Both the world and the sense of place are provisional, though. That's a point driven home by Link's Awakening, no doubt the most self-reflexive installment in the franchise (though it also lends itself to themes of mortality). There, the level bosses are continually reminding the player that Link's quest to awaken the Wind Fish will ultimately unravel the world of the game — a world that is, in the first place, articulated only in service of the quest. Like Thomas More's Utopia, the world of the adventure videogame is no-place, making total access a dream with no substance.