We try to make decisions as a group. In the TShock project, we have a two-man rule on code review. This is discussed in more detail in that project, but:
- Two people must sign off on code for it to go in. If the code is internal (from another leader), you just need one more sign-off to get it merged. Please respect if someone requests changes: that doesn't mean get approval and go around, it means work with them to turn their denial to approval.
- Two leaders must review external code. This is technically three sets of eyes. If you're the second leader reviewing code, and you feel comfortable, you can merge it. If you don't -- that's okay, just approve it and move on.
In non-code activities, we try to reach an "ad-hoc" majority consensus. This is a really loose, trust based decision system:
- We poll people who are available, in a reasonable amount of time, to make a decision in high charity.
- If the majority of the people polled are in favor of an action, it happens.
A reasonable amount of time is defined as "about three days." If you're the one trying to get a decision made, it's your job to reach out to people to get them to chime in.
Finally, we err on the side of trust with our decisions. If a consensus was reached and we made a decision, we try not to go back on it later. We haven't been the best about this in the past, especially with kicking and banning troublesome members, but we should strive to be more consistent. Unless outstanding circumstances apply, we should strive to not reverse decisions made as a collective unless absolutely necessary.