I seem to have found a suitable solution to this problem owing at least slightly to a suggestion I got in response to my original posting.
The solution is to use a single Junk folder but alter my spam-training process to instead of learning from all messages in the Junk folder, to skip messages that are already flagged as spam (i.e. "NOT HEADER X-Spam-Flag YES" in IMAP search language).
Additionally, the messages which were used for learning (i.e. that were put there by the user) get deleted so the user doesn't have to sift through them as they look for false positives.
Much thanks for all of the input.
Cheers, b.
--- 8< --- detachments --- 8< --- The following attachments have been detached and are available for viewing. http://detached.gigo.com/rc/4N/wTxLbvw5/signature.asc Only click these links if you trust the sender, as well as this message. --- 8< --- detachments --- 8< ---
List info: http://lists.roundcube.net/dev/ BT/aba52c80