OSSGameBench: A La ge-Scale
Da ase o Con ibu o Ac i i ies in
Open-Sou ce Video Games
Online Appendix
1. Abs ac 2
1. Abs ac
Video games ha e e ol ed beyond en e ainmen o suppo educa ion, aining, and well-being
h ough deeply imme si e and in e ac i e expe iences. Ye , unlike o he so wa e sys ems
whe e s anda dized benchma k da ase s ha e accele a ed esea ch on downs eam asks such
as de ec p io i iza ion and localiza ion, he game de elopmen domain lacks compa able
esou ces ha e lec i s dis inc i e cha ac e is ics. Fo ins ance, de ec s in games o en a ise
om he complex in e play be ween code and non-code asse s, whe eas in o he so wa e hey
do no . The e o e, specialized da ase s a e equi ed o cap u e hese mul idisciplina y demands
unique o game de elopmen . This gap limi s he de elopmen and e alua ion o da a-d i en
app oaches o asks such as de ec p io i iza ion, de ec localiza ion, and au oma ed epai .
To ill his oid, we in oduce OSSGameBench, a cu a ed benchma k da ase de i ed om 970
open-sou ce game eposi o ies in Gi Hub, which con ains issues, commen s, commi s, pull
eques s, and e iew commen s, wi h explici mappings among hese en i ies o enable
ep oducible and comp ehensi e de ec analysis asks in game de elopmen .
