Game Development Reference
In-Depth Information
of this type. Since implementation costs are so straightforward, they are
almost always planned for, though often underestimated.
IMMATURITY BURDEN is the cost imposed on people who have to do
work that depends on incomplete parts of the game.
Most game developers don't have the luxury of working with stable,
mature software and game systems. Usually the tools are buggy and
poorly documented, the game mechanics are half-made, and the story
shifts daily. These immature elements impose a cost on everyone working
with them. Nonfatal but annoying software bugs slow down work. An
unfinished story slows down level design by injecting uncertainty into
level design decisions. Unbalanced game mechanics make it hard for level
designers to polish and balance challenges.
The dependency stack helps us avoid depending on immature de-
signs. But even with a clear dependency stack, it's impossible to completely
eliminate immaturity burden.
CRITICAL FAILURE RISKS are the costs imposed by critical failures of
immature systems.
Any immature system can reveal a fatal flaw at an inopportune time.
Fatal, difficult software bugs can throw tests into chaos. Unpredicted
dominant strategies can destroy months of balance work. Narrative holes
can lie unnoticed for months. These problems hide in the game like time
bombs, before going off and causing some critical failure.
The cost of these problems isn't so much the actual work of fixing
them, but their impact on other processes. A simple mistake by a tools
coder can force an artist in another building to spend two days tracking
down an asset import bug. Even if the bug fix is a one-line code change,
the damage has been done because the artist is now slowed, as is everyone
depending on him.
The damage is worst when these critical failure time bombs go off in
time-critical situations. Sometimes that one blocked artist was working on
a piece of art that was critical to the next test of a new game mode, which
was scheduled for the next day. The missing art forces a cancellation of the
test, which starves the designers of playtest data, which inhibits decisions
on the next code iteration, leaving gameplay coders without clear direc-
tion. This sort of chain reaction is not unusual in large, complex processes.
