There's been a lot of talk on the Design list especially wrt various ways to improve the management of mailing list and rss feed items. Power users especially, some of whom subscribe to hundreds of lists have expressed a desire to be able to differentiate between mailing list folders that have "unread" messages and mailing list folders that have "new, unread" messages. The reason being, most mailing list messages go unread and as a result, most mailing folders consistently show hundreds of unread messages which the user is not interested in knowing about. However, when new messages come in, they ARE interested in scanning them to see if anything worthwhile came through.

Most suggestions have centered around providing some kind of visual feedback (color, flash, icon) to indicate that a mailing list folder has received new items (new as defined by the user to be some period of time, ie. 15 minutes).

However, in the Chandler world where the notion of Triage (get this item out of my face and into, for example a mailing list folder) and Filing have been decoupled so that mailing list items can still live in the "Inbox" or "Dashboard" view so to speak, even if they've already been assigned to a mailing list collection.

As a result of this decoupling, we can offer a much simpler solution that:

1. Doesn't burden the user with a lot of visual feedback clutter to decode
2. Doesn't burden the user with the need to go to the mailing list folder in order to check out recent messages

Instead, all new mailing list messages come through the user's Dashboard. The user can have a filter in place that automatically marks and moves such messages as Done so that they are moved out of the users' face, aka Inbox or Now area OR for users who are concerned about keeping on top of their mailing list items, they can simply be marked as Done, but remain in the user's Now area until the user explicitly clicks the Purge button, at which point, all items marked as Done are moved away.

This allows users to keep on top of things as they come in, without burdening them with a lot of mouse clicks and DnD to put things in the right place.

Also, there should be no reason why we can't model auto-triaging after spam-filtering-learning-mode, where the client learns over time the kinds of things users generally mark as Done right away (ie. PTO announcements, lunch is here announcements, mailing list items). Many of the users we interviewed expressed a desire for "rule-building" functionality but didn't really know where to begin to find the feature. However, if learning mode can work for SPAM, it should also work for Triaging and perhaps even Labelling and as long as items aren't automatically moved away, there is always a human element (the user) to correct mistakes.