[CentOS] A Group is Its Own Worst Enemy (was: EL 6 rollout strategies? (Scientific Linux))

[Emmanuel Noobadmin]

On 5/17/11, m.roth at 5-cent.us <m.roth at 5-cent.us> wrote:
> If we *really* need a moderator, here's an option: soc.religion.paganism
> has a robomoderator; on topic posts get autoapproved, obviously off-topic
> get bounced, and if there's any question, they get randomly bounced to a
> configurable number of human moderators, allowing for load balancing on
> the humans (and vacations, sick time, etc).

As already said, this sounds really complicated. Coming from an IRC
and vBB admin background, I'll suggest moderation using the reactive
approach instead of a automated process.

In the context of a ML, the signature could include a link with an
unique subscriber/msg hash that triggers an email to moderators only
if a threshold is reached. This limit actual moderation only to
threads/posts that are really offensive to the point people are
willing to actively flag it rather than rely on some automated process
to decide if something's really off-topic or offensive.