Path

I've come to a conclusion about level design philosophy that's probably elementary knowledge for someone, say, studying game design in college, but it just congealed for me during my stint at TimeGate. Even when building a level that only provides the player with a strictly linear path, the designer should build the path through the place, not build the place around the path. It's about contrivance and cohesiveness.

I think that there's a natural inclination when laying out a linear shooter level to sketch an 'interesting,' abstract path first, then rationalize it by building out the appropriate geometry around it. In my experience, this tends to lead to spaces that feel very contrived and 'gamey.' Places built around a path are disconnected from a sense of purpose-- where is this curved hallway supposed to lead? Why do these storerooms feed into one another like a string of pearls? Why would they make people in this facility go up a series of ramps and catwalks to exit this room? Why is the layout of this place so convoluted?

I believe that the superior approach is to build a place first--a cohesive, functional space, with a purpose--then define a path through it by strategically restricting the player's movement. If it's a factory setting, build a small complex of storerooms, packing floors and shipping bays in an open structure that could simply be a place of its own, then start blocking off hallways, locking doors, collapsing staircases, and so forth to remove all means of egress that conflict with your chosen gameplay path through the space.

Hereby, the overall space inherently makes sense, while still allowing the designer to have a strong hand in leading the player. No part of the built space, and therefore valuable time, need be wasted-- the player can still be given controlled access to each room; or if a room isn't visited, it is at least visible, understood by the player as a part of the place he's exploring, showing him that there's more to the gameworld than the little path he's running along.