That answer is so short as to be potentially misleading. Please correct me if my understanding of SpamSieve and IMAP is wrong. I’m here because I’ve been hosing myself for reasons that are entirely obvious after a bit of reflection, after switching to IMAP. I came for confirmation, and I’m surprised not to find it.
If each of several Macs are running Spamsieve with the same IMAP, it becomes a race which Mac deletes a spam message from the server, copying it instead to the local Spam folder on that machine. Yeah, only that machine. As in, not the other machines. As in, you can’t see the message from the other machines. It’s not there.
For example, I didn’t get a PayPal bill today that I was expecting. In the old POP days, I’d look in my local Spam folder, see that this email was wrongly classified as spam, and retrieve the message. In the new IMAP days, the errant message could be in any machine’s local Spam folder, a horse race decided by who got to the message first. If I’m not at the right machine, I can’t access the message.
Spamsieve uses sophisticated probabilistic methods. Probability theory studies effects like when you give a monkey a football, it will run toward the wrong goal line 2/3rd’s of the time. This manifests itself using IMAP. If two machines are running Spamsieve with the same IMAP account, then misclassified spam messages will end up on the machine you’re not sitting in front of, 2/3rd’s of the time. People get Ph.D.s for studying stuff like this.
The moral is clearly to run SpamSieve on only one machine. Indeed, there’s a manual section on setting up a drone machine. However, that section baffles me. How can another machine even see a message that should be classified as good, after the drone deletes the message from the IMAP server? This is pretty funny, actually.