Bälter captures keystrokes of people using email, and from that, develops a mathematical model of how long it takes people to do things. The key insight is that for large volumes of email, it takes more time to file messages than it saves in later retrieval.
Bälter and Sidner prototype a system that allows all messages to remain in the inbox, but to be categorized into five groups (which was as many as Lotus Notes let them do), and the inbox sorted by category. Their categories were a mix of who the message was from, who it was to, and the contents of the message. Interviews indicated that people quite liked it.
The authors do quite extensive studies of how people keep track of their to-dos (not in email specifically) and describe a prototype to-do management system.
In this important paper, the authors motivate and describe an application that shows tasks and email messages simultaneously in one list-of-messages.
The authors describe the process of designing an integrated PIM, which is basically email + sticky notes + reminders + grouping.
The authors describe grouping messages in the inbox based on which folder it is likely to end up in, very similar to Mock (2001).
The authors studied how the type of a message (request for information, request for action, etc) and the social relationship of the sender affected how the reciever would deal with the message. They got lots of quantitative information.
The authors find that managers organize their folders differently than non-managers: in a more relationship-centric way, with more folders, and with more hierarchy.
In this widely-cited paper, the authors discuss the results of fieldwork that confirms Whittaker and Sidner's contention that email is used for many things besides email, and further shows that people "live in" their email client.
This is an early, oft-cited paper. The Communicator was an email system that mandated that users tag messages with what kind of message they were -- request, offer, decline, etc -- as a way of helping people keep track of the progress of a conversation. (While it's not in the paper, I've heard that it was a smashing failure: it demanded too much overhead from senders, and removed some essential ambiguity from conversational transactions.)
This paper has much of the same stuff as Kerr et al (2004). Read that paper instead.
This paper is about two things: an email interface and also an interesting way of testing users. The email interface (TaskView) is based on TimeStore, but has explicit indications of tasks in messages. Users had a number of abstract cognitive abilities tested, then were tested on email tasks on two different clients. Broadly speaking, the hypotheses were that better interfaces would help people who scored lower on the cognitive abilities more than the better interfaces would help the high-cog users, and that TaskView was a better email interface than Outlook. TaskView turned out to be better for finding Dates, but worse at a "Header" task. Some cognitive abilities helped, others didn't.
This is more-or-less an earlier and briefer version of Gwidzka and Chignell (2004).
The authors describe a Bayesian technique for determining how urgent an email message is, and briefly discuss its use in a prototype Outlook plug-in called Priorities.
There are quite a few papers that have come out of the Remail project; this one is a good representative. The Remail system has just a ton of new features, too many to mention here. This paper is worth reading (and is easy to read.)
DriftCatcher is an email client that displays messages differently depending upon the strength of the social relationship the user has to the sender (based on duration, frequency, and reciprocity) and what type of transaction (based on content analysis of the message).
Mock has an Outlook add-on with a learning classifier system that assigns a category to a message. The categories are learned by examining the existing user folders. This is similar to many systems (e.g. iFile), but Mock uses the categories to group by category in the inbox (similar to what Overcome Email Overload advises).
The authors find that if two strangers communicating over email share some personal information first, subsequent negotiations are more likely to be mutually successful. This has nothing to do with my CSC2536 presentation, but it is a really interesting paper.
This is an interesting paper about a groupware tool that is part IM, part email, part contact management, part shared whiteboard, and seems very well put together.
The authors describe a sidebar with correspondents, listed in social-relevance order, that shows how many messages there are from each correspondent. This social relevance information alters how messages are displayed; clicking on a person takes the user to messages from that person.
The authors report statistics on how people triage their email.
The author distills a number of findings from interviews and observations, as well as academic research, and converts them into specific, Outlook-specific techniques for lay audiences to deal with email faster.
Stern shows their technique for extracting dates and times from email, and claims an 80% success rate. She appears to use a much more deterministic technique than Horvitz et al do in Priorites.
The authors discuss an agent for processing email in the context of mobile users. Not very interesting to me.
This paper, like SNARF (2005), comes out of Microsoft Research, and does similar people-based organizing, but Inner Circle is much simpler.
The authors present an algorithm for teasing out what people are in a "community of practice" from email records.
This paper isn't all that different from Venolia et al (2003), though it reports more on the observations leading to their models.
The authors describe an email client which aggregates thread messages, strips redundant messages, and presents them in a tree format.
The authors produced an email client that was people-centric instead of message-centric. Several views are available to show different aspects of social networks, including an email view.
This is a seminal, must-read paper. The authors make a number of insights about how people use email, including: a) that they use email as a to-do list, and hence want to see all their "to-do messages" in one place; b) that people divide pretty neatly into "no filers", "frequent filers", and "spring cleaners"; c) that people need conversational context (i.e. better threading).
Zest is a hierarchical thread browser, which strips quoted material and which allows expanding/collapsing a subthread.
TimeStore is an email client that shows the list-of-messages in a 2D grid form, where time is on the x-axis and correspondents are on the y-axis.