Game Development Reference
In-Depth Information
regression and load/stress tests on massive online environments. But even if online
MMP games are not your company's main output, you will still be on solid logical
ground arguing that highly repetitive testing is well suited to being replaced with auto-
mated testing. You will also be on firm ground arguing that to some degree, at least,
automated “sniff test�? testing should be introduced, which ensures that new code does
not contain obvious flaws before a commit is permitted.
Some of the predictable fixed costs when converting to automation are as follows:
Additional hardware (or at the least upgrades to existing hardware)
■
Middleware and tool licenses (you are likely to need at least some)
■
Scripting tool creation
■
New game code to add hooks, framework, and so on for automation
■
Tool training
■
Management software and support
■
Ramp-up costs
■
Additional team members who are devoted to automation
■
You will face some variable costs, too:
Test case implementation
■
Test case designs specific to automation
■
Maintenance
■
Results analysis
■
Defect reporting
■
Night-time system runtime
■
It should be straightforward to calculate the costs of these factors. They are based on
well-known variables and cost bases that already exist in some (at least similar) form
within your company, either in relation to existing manual testing or to game code
development.
In their 2003 presentation to the Game Developer's Conference in San Jose, Greg
Kearney, Larry Mellon, and Darrin West detailed their implementation of game test
automation in the development of
The Sims Online
(http://serious-code.net/
moin.cgi/AutomatedTestingInMmpGames)
. The problem they identified was that
developing and deploying massively multiplayer Permanent State Worlds (PSWs) has
proven to be very difficult due to the distributed nature and large scale of the system.
