Towards a Digital Architecture
The thoughts and opinions expressed are those of the writer and not Gamasutra or its parent company.
I occasionally wonder what type of games famous architects would play, or even what they would design. I think Le Corbusier’s fondness of climbing levels, ramps, and looking down on surrounding areas would make him very talented at designingfirst person shooters. Frank Lloyd Wright’s “refuge” centric architecture; with the hearth being the center of thehouse and setting buildings among shaded cover; would be perfect for tower defense games. Likewise, I.M.Pei’s large museum entry spaces; providing previews of upcoming floors and exhibits that visitors will soon explore; play out like a 3D Mario game, wherethe first star in a world typically requires players to traverse the whole territory. Louis Kahn’s National Assembly Building in Dhaka, Bangladesh and its journey of concentric rings plays like a Zelda game. Finally,when I was once asked by a professor what game the architecture school I studied at resembled; with its lack of spatial terminus points, free flowing spaces, and sandbox-esque studio; I answered “Grand Theft Auto.” (He took it surprisingly well…)
"Sniping's a good job mate!"
The somewhat long winded point I am trying to make is that ina time when gaming is trying to find its own way to slip itself into the somewhat ambiguously defined realm of human creation known as “art”, architecture may not be a bad place to make the connection. (Actually, I also agree with the viewof some that games can also be like performance art, but I don’t have as much expertise in that area) But haven’t we tread this road before? Haven’t we had lots of discussion on this “games and art” topicalready? Yes, and I don’t intend to restart the debate with this post. Rather, I’d like to look at the similarities between these two art forms and find ways that games can learn from architecture, specifically in the realm of level design.
Games, as we all know, are interactive user experiences created by designers. While many architects will design buildings from a formal “parti” (a basic form from which a building design is conceived…basically an architectural core mechanic), and analyze them from a sort of “top-down” approach through plan, section, and elevation drawings, they become interactive the moment a person steps foot inside and looks around. It is in this very simple way that games and architecture are very similar. Architects who intend to create useful building spaces look at how the motion of occupants can be channeled by the spatial design decisions they make (“form follows function.”) When it comes to video games, the rules of play are often constraints on player movement and action specified in the programming code. While this is a somewhat simplistic view, most of these constraints are viewable to the player as the abilities their avatar has and the level design itself. In this way, it can be said that the architectural forms of a gamespace are an embodiment of the rules of agame…game design and architecture are very close indeed.
Often, when real buildings are brought up to game designers,they are dismissed as not having the same design constraints as good gamelevels. A good example is the re-creation of Frank Lloyd Wright’s Fallingwater in Half-Life 2. While a virtual tour of this seminal piece of architecture is great in theory, it bombed miserably as a Counterstrike: Source map. Likewise, my own recreation of the Basilica of the National Shrine of the Immaculate Conception in Portal succeeded in its intended purpose; allowing designers of a mosaic to see what it would be like to interact with their work; but didn’t offer much gameplay-wise when the Portal Gun was added to my own…ahem…test versions…
These examples don’t work because the spatial scaling of real buildings does not often translate well to games. Game engines present space to gamers with simulations of wide angled lenses that require level designers to model things bigger than they would actually be. In this way, the dimensions the spaces within of Fallingwater were very poor for the large-scale gunfights that define mostmodern FPS’s. What do work, however, are the “under the hood” reasons of why these buildings are greatpieces of architecture. Great buildings are great buildings for many reasons, but many of them live on as great experiences because they employ some element of spatial psychology designed to evoke certain emotional responses.
In his book, The Origins of Architectural Pleasure (which I think should be required reading for all level designers), Grant Hildebrand makes connections between elements of human survival instincts and the experiences of great buildings. I don’t want to spoil the book (buy it and read it…really), but I will say that most games; with the exciting exception of some of the more artsy/indie titles; use survival as a primary motivation for player actions, making the subject matter incredibly relevant to game designers. Topics like the use of refuges (the architectural equivalent of cover) or prospects (the architectural equivalent of open areas for sniping), light quality, shadow, materiality, and others can have a great dealto teach designers making effective levels.
Likewise, elements of spatial psychology, especially using behavior theory to lead players through danger to rewarding safe zones, cangive players small feelings of accomplishment, despite the lack of a more tangible reward such as a new gun or health pack. Spatial or visual rewards are used as masterfully in many of Valve’s games; notably the Half-Life series, as they are in buildings such as thosebuilt by Louis Kahn. In the developer commentary to some of Valve’s games or discussions with the designers themselves, it is not uncommon to hear them discuss “providing rewarding vistas” to players who reach a new area, such as the energy core of Episode One’s Citadel or the White Forest base in Episode Two. These types of concepts can allow designers to better control level flow and even lead players down intended paths, or provide hints towards hidden surprises, a la Metroid.
Another concept I am very fond of is Christopher Alexander’s idea of a “Zen View” from his book A Pattern Language: Towns, Buildings, Construction. In this topic, he describes a wall built by a monk to hide a beautiful mountain vista, making it only visible from a certain spot at a certain angle through a small slit window. The thought behind this is that the view is wasted if you have immediate access to it at all times, and that it becomes more rewarding when you get small experiences of it. Salen and Zimmerman would refer to this as denial; keeping a reward from a player but letting them know it exists. Such techniques can help give players a preview of either long or short-term goals in game levels,but deny them direct access and instead require them to traverse the territories where the creatures with nasty-big-pointy teeth live.
While this sounds like it SHOULD be common sense, there are many examples of levels that are frustrating for one or more design issues. One of my favorite examples to cite is the Library level from the original Halo: Combat Evolved. Many players describe it as one of the most frustrating levels in the series. While the Flood gets much of the blame for this, the level space itself is a giant open area with little to no safe zones from enemies. Making it through some of the prospect-like area is often done largely through sheer luck, with the only reward being access to an almost identical floor elsewhere in the complex. “Pancaking” floors are relatively common through many games (but a characteristic of bad architecture), and while good level design is sometimes not a concern for the team designing certain games (the Halo team refined FPS controls after all), frustrating gamespaces can even turn playing an amazing game into a labor.
"How not to design a level"
I felt this way about the recent Final Fantasy XIII, when endlessly linear tunnels replaced the beloved sweeping landscapes of my childhood. I realize, of course, that JRPG’s have always been linear, but it was that world map that really created the illusion of non-linear exploration for most players. Likewise, the player could geographically track their progress through the game: were they in Midgar or Wutai? Were they capable of only crossing rivers in the buggy or could they fly anywhere on the Highwind? I used to know these things. I have no idea where Lake Bresha is in relation to Palumpolum, and cannot revisit either one if I wished.
Even outdoor areas in games can benefit from architectural thought. I often find myself telling students that despite their natural appearance, their world maps and outdoor levels are still spaces with aspects that have to be properly tuned and craftedby a designer for the best experience possible. Chaim Gingold, noted for his work on Spore, based parts of his dissertation, “Miniature Gardens and Magic Crayons: Games, Spaces, and Worlds”, (yet another work that all level designers should read) heavily on David Slawson’s Secret Teachings in the Art of Japanese Gardens: Design Principles, Aesthetic Values, and likens the knowledge from the book with the design aesthetics of many of Shigeru Miyamoto’s worlds. As I stated before, some of these aesthetics; such as giving new occupants an initial preview of a space; are prevalent in the works of architect I.M. Pei, such as in the Rock n’ Roll Hall of Fame in Cleveland, Ohio.
I.M. Pei provides enticing previews in many of his buildings
Recently, I have been experimenting with the Unity engine for an upcoming game project I am creating with several students. To prepare for the project, I have been following the book Unity Game Development Essentials by Will Goldstone; which teaches the engine by having students create and island then populate it with mini-games. I followed the directions for the island from the book, then realized that much of the action occurred in the central area, where there was a hut for the player to occupy, and that the rest of the island was largely wasted. I referred to some of the Chaim Gingold and David Slawson concepts to save my island. To make things more interesting, I modified the game’s first task; finding 4 batteries to open the hut’s electronic door; to require the player to search the entire island for the batteries, rather than lay them out around the hut like the book suggested. Adding to the “tour” aspect, I placed one of the batteries on a high hill that let the player see the whole island. Basic? Of course. Extraneous for an engine tutorial? You bet! But I thought that giving a tour of the island to new players during their first task would not only introduce them to my level, it would make it more fun.
This post merely scratches the surface of ways that paying attention to real architecture can help educate game developers to pay better attention to the experience of their own digital worlds. It would be, however, outside the scope of this blog to go into everything. This is a topic that I believe has many varied and interesting facets and I hope it is a topic of further reading for those reading this post, and possibly more writing on my own part. What can the principles of Wright, Pei, Corbu, Kahn, and many many others teach us about making great video game levels? Understanding of such ideas can only make great level design appear more often in games and populate the gaming frontier with more masterworks of digital architecture.