Game Development Reference
In-Depth Information
P
ROBLEMS WITH
C
ATEGORIZATION
Difficulties also arise when you are trying to categorize things. Often, the definition
of a category is entirely subjective. In our color example, people may have different
understandings of what a color “should be.� My ex-wife used to ridicule me for
claiming that one of the four colors of tiles in our Rummikub set was orange. She
asserted that it was yellow. In fact, neither of us was “right.� Orange and yellow are
examples of invented classifications that we humans have devised for the sake of
simplicity. By being able to say “orange� or “yellow� we spare ourselves the headache
of trying to communicate color through an expression of relative wavelengths of
light. (On the other hand, there is a growing portion of the computer-graphic-
savvy population that can speak quite fluently in red-green-blue [RGB] triads.)
My ex and I were looking at the same color wavelengths. For the sake of argu-
ment, let's assume that our eyes worked well enough to perceive those wavelengths
properly. The problem arose because we likely had differing opinions of where
yellow ended and orange began (or vice versa). In fact, the color was somewhere in
between the commonly accepted definitions of those two colors. However, not
having a box of Crayola crayons handy to help us negotiate a compromise such as
“sunglow,� “dandelion,� “goldenrod,� or even simply “yellow orange,� neither one
of us was willing to back down on our assertion about the color of the tile.
(Full disclosure: Wikipedia lists the color as yellow, but the instructions on the
official Rummikub site list it as orange. I feel better about myself now. Kinda
vindicated, ya know?)
This concept is surprisingly important to modeling not only perception systems
in games, but decision systems as well. Let's look back over some of the classifica-
tions we have to make for our engagement decision: What is “low health?� How
many enemies need to be present before we begin to get nervous about being “out-
numbered?� What is the actual linear distance away from our leader that no longer
constitutes “close enough to feel safe?� These are examples of decisions we would
need to make if we wanted to measure and compare attributes in a meaningful way.
P
ROBLEMS WITH
U
NDERSTANDING
The issue that generates the most inconsistency and subjectivity when people address
attributes is based more in psychology than perception, however. Even when obser-
vation and measurement can be exact, therefore taking the vagaries of perception
out of the loop, there are other issues in play. Extensive studies have shown that not
only do different people put differing values on things, but the same person may put
differing values on things based on their understanding of the situation. By stating
the situation another way or in terms of another scale of measurement, the answers
that people give may change significantly. The facts don't change, but the meaning
people ascribe to them may differ significantly.
