Similar to #2, it would be useful to have a temporary reject threshold in addition to the permanent reject threshold.
The idea being that a piece of mail arrives that is not quite spammy enough to outright reject it, but it looks like it could be spam. If it is re-evaluated later in time (i.e. when the remote retries sending it because it was only temporarily rejected) it's score could be higher due to more spam databases being aware of it.
I suppose this is a spam-score based grey-listing system.
The part that makes this a bit less straightforward is that a piece of mail that is in the grey-area, that does not increase in score after multiple attempts over a defined period of time should probably be accepted. So a database of temporary rejects would need to be kept to score the retries against.
Of course, I suppose this hold-and-retest-periodically could be done with MTAs that implement quarantine, but this solution would be applicable to MTAs that don't do that also.
Similar to #2, it would be useful to have a temporary reject threshold in addition to the permanent reject threshold.
The idea being that a piece of mail arrives that is not quite spammy enough to outright reject it, but it looks like it could be spam. If it is re-evaluated later in time (i.e. when the remote retries sending it because it was only temporarily rejected) it's score could be higher due to more spam databases being aware of it.
I suppose this is a spam-score based grey-listing system.
The part that makes this a bit less straightforward is that a piece of mail that is in the grey-area, that does not increase in score after multiple attempts over a defined period of time should probably be accepted. So a database of temporary rejects would need to be kept to score the retries against.
Of course, I suppose this hold-and-retest-periodically could be done with MTAs that implement quarantine, but this solution would be applicable to MTAs that don't do that also.