On Thu, Sep 25, 2003 at 10:37:28AM -0700, WJCarpenter wrote:
> Docs (and books) for both Exim 3 and 4 say, in effect, "Here's how to
> allow 'fail' and 'freeze' actions in user filters, but don't do it if
> users actually control their own filters."
>
> I can see how there's not much motivation for allowing users to freeze
> messages, but what's the danger in allowing users to fail messages?
> Is this a philosophical position ("admins rule!"), a technical
> limitation ("if you allow that, there are certain mumbles that are
> very hard to detect and overcome"), or middle ground ("many users
> don't know what the heck they're doing and will embarass themselves or
> the entire domain if they try using this feature")?
>
> If a user wants to fail a message with a SpamAssassin score of 114 or
> more, why shouldn't I let them?
One of the reasons might be that it is expensive, because you have to
generate a bounce, which probably will be stuck in your queue as
brozen, because spam's sender addresses usually are not deliverable.
cu andreas
