We need some consensus on what it means to ban someone, and which triggers should be applied. SMF allows us to ban based on username, e-mail address, IP address and/or hostname.
When a ban is issued, we wish to limit or restrict forum access for some identity. The subject for discussion in this thread is how we define the identity that is to be banned; for example, is a username sufficient? Is an IP address part of the identity we are banning? Once we have clearly defined an identity, we can apply ban triggers appropriately.
I am raising this in S&C because I think all members should have a say in how bans are orchestrated.
My view on this matter stems from the belief that a ban is always intended as a means of rehabilitation, not retribution. The purpose of a ban is to, where prior warnings have failed, discourage violation of the rules and encourage contributive posting. It is never to punish someone for past actions; nothing is to be gained by punishing someone for an offence they are no longer committing.
I propose that, for a first offence, an identity is always limited to the user account, not the IP address. That account is the thing that has been causing trouble, and it may be uniquely associated with a username and an e-mail address, but not necessarily an IP address. Therefore, a first ban should always make use of only the username and e-mail address triggers.
Suppose, then, that the user makes another account. There are two possibilities; one, that they use the same IP address as before; two, that they use a different IP address. In the latter case, an IP address trigger would have been useless anyway, so we shall only consider the former case. We can subdivide this into two further possibilities; either they create an account and continue to violate the rules as before, or they start posting in a contributive fashion and adhering to the rules.
If they are continuing to violate the rules, then we can consider the IP address as an identity that is causing trouble, as it is now creating multiple troublesome accounts. We may then apply an IP address trigger to resolve that problem.
The final possibility is that the user creates an account and starts contributing to the forum. It follows from the belief that bans are not for retribution that we do not care what the user has done previously; they are contributing now, so we don't need to restrict their forum access. On the other hand, if we had applied an IP address ban earlier, we would have prevented them from taking this positive step to recovery and lost a contributing member.
I conclude that, if you agree with my statement that bans are never for the purpose of retribution, using IP address triggers on the first offence is harmful rather than beneficial to the forum. I therefore propose that IP address (and hostname, which is almost the same thing) bans are never used unless it can be shown that they are chronically associated with troublemaking accounts.
Thoughts and comments welcome.