This is similar to bug 151887 (auto-response), but much more powerful. What I propose is that the filters be set up to prompt the user to select one or more responses from a list of possible responses; the user should then have the option of either send-as-is or to edit the message. I actually got to use such a system in 1995 when I was the UIUC webmaster, and it was *wonderful*. For "customer support"-type jobs, it can make a 10x improvement in response speed. (Really. If there's anyone at Netscape around still who was at NCSA then (alanb?), ask them about Ben's @ATS.) This would be a major competetive advantage over other email clients in the customer-support niche. EXAMPLE Here's an example. I had a rule set up so that if the string "admission" was in the body of the incoming message, it would suggest the "graduate admissions" response and the "undergraduate admissions" response. I also had a filter which suggested the "I need more information" response for all messages. When I read a message that had "admissions" in the body, at the bottom of the message would be three checkboxes, marked "undergraduate admissions", "graduate admissions", and "I need more information". I could click on the titles of the canned responses to bring up a window with the response. Below the checkboxes would be "Edit" and "Send". If I could tell that it was about undergrad admissions, I'd put a checkmark in the first box. If I could tell it was about graduate admissions, I'd but a checkmark in the second box. If I could tell that it was about admissions but not which type, I'd check both boxes. If it was some sort of word salad that made no sense, I'd check the third box. If it happened to be about a department that I knew something about, I might click "Edit" and add a line or two about the department. Otherwise, I'd just hit send. Boom, done. This is particularly useful when there are lots and lots of canned answers that you might use. It's much faster to choose from three answers than from three hundred. It also helps less-well-trained people do support. Instead of someone having to know how to explain a feature , instead of them having to know that a writeup on said feature exists (and where to find it), all they have to be able to do is read the question, read the suggested response(s), and recognize if the response answers the question. IMPLEMENTATION DETAILS Should you implement this, the response corpus could come from responses stored locally (see bug 21210). It would also be good to allow URLs (of all types, including file:) for the suggested responses -- so that a group of people could take advantage of the same answer corpus. It would also be good to be able to import/export filters on a per-filter basis (bug 151612) so that the team can share the same filter corpus as well. To best use this, there should be multiple actions per filter (bug 13145) or even better, an autodownload of filters (bug 126688).
Confirming, valid RFE. This one would be really neat and could be the next best thing after spam-filtering.
sorry for the spam. making bugzilla reflect reality as I'm not working on these bugs. filter on FOOBARCHEESE to remove these in bulk.
Filter on "Nobody_NScomTLD_20080620"