Not-So-Spatial Spaces

Let's think for a minute about how a game works. There are rules, and objects, and objectives, and we have to know the best possible way to play by the rules so we can complete the objective. How do we do that? We play with it, and we get a better and better mental model of how the game works, which in turn allows us to react faster to things in the game, and make better choices in our strategy.

For this reason of course, it's important to have a good interface. A person should be given a consistent and exact model of what is happening in the world; spatially graphics need to be consistent. A monster's avatar should increase in size consistently with each step it takes closer to you, a fixed rate of change; units in strategy games need to be represented correctly in how far apart they are from each other. It's completely rigid and mathematical.

So why am I posting about this? Anyone can see that in order to figure something out, feedback needs to be consistent. On the other hand, what does this kind of an interface suggest? It suggests that pretty much every game we play is of a consistent formula, and that we're experimenting with it until we boil the game down to pure consistent logic. Nothing wrong with that, there needs to be some basis for every system, however:

Why does our information need to be represented so spatially consistently all the time? Well, no other reason than that the logic of most games is so damned reliant on spatial factors. One of the reasons that books, plays, movies, etc. can be so imaginative is that there doesn't need to be this precise idea about space. As long as it looks feasible, we can assume that they're in a real-world situation, and their ideas don't have to be subordinate to the tedious and boring parts of reality.

We also don't always need to be loaded with information about hit-points, damage, bonuses, etc. Take the first Command and Conquer: even if there was no information on hitpoints and damage, it took about one failed assault to realize that you shouldn't send infantry against flame-based weapons or a group of undefended tanks into a field crawling with rocket launchers.

So I'm getting on a tangent away from my larger point, let's get back to it. The key information of many games shouldn't require such exact feedback on the screen. I'm not talking about first-person shooters or platformers, those are games that are about space; but for many other games, we're talking about how objects are related to one another, how one player interacts with another, alliances, etc. If we want games to move further towards concepts like story, relationships between characters, human behavior, etc; then there's no point in thinking so hard about things like space. In fact, I don't want to see such constraints on space in a game of this type because everything else risks becoming subordinate to it. On the other hand, having a space to move around in within a game makes it feel a lot more real. So, here's a proposal for the best of both worlds; take a look at this application of dynamic programming to shrink an image with 'retargeting':

http://www.milkandcookies.com/link/66481/detail/

Notice how there's no need for proper scale and distance between objects when they shrink it down. We probably wouldn't notice that so much was changed if we were just looking at the pictures casually. It preserves the exact same idea, and does a better job of it than if we scale it, where parts of the picture gradually become obscured. What I'd like to see is an RPG (or something like it) that follows this model. We can slow down or speed up the character accordingly to make sure that the player doesn't just shrink the screen to speed up movement (I think everyone would agree that that would mess with suspension of disbelief). This way, the player feels like he's moving in the real world, but we have a lot more flexibility with artistic representation of the game. What feedback on space the player is getting isn't important, this isn't a strategy game where we have to know where to place artillery and how close the enemy is to our base, we're simply traveling a world where we talk, buy, explore, and so on.

This image retargeting might be a bit of a silly example, on its own it doesn't mean much, but it also brings across a general point. The algorithm is able to preserve a sense of reality without being constrained by space. In doing so, it does a lot better of a job of making the shrunken picture convey the important information. We need to do this more often in video games, or else the important information will always have to do with space, something that really only matters if we're testing our ability to understand how things move in a space. Doing something else will bring the player's cognitions elsewhere and expand the medium. This clearly isn't off limits to video games, we know enough about computer science to make space look continuous without letting it dictate the rules of the game.