Beyond Threads: Identifying Discussions in Email Archives

摘要

Email archives have the promise of serving as great resources for historians and social scientists. However, making sense of the information in these archives is a challenge. Email messages are often not self-contained and are instead a part of an ongoing discussion. The process of determining when discussions commence and conclude is a difficult task to automate. Threading messages by common subject lines and reply-chain information in email headers has been a common way of assembling messages into discussions. However, even though email provides this structured information, it does not usually reflect the user's behavior. Our work helps email archive explorers interpret the archived messages by providing access to the full scope of discussions that stretch beyond the thread. We present an interactive visualization that allows explorers to perceive a discussion and navigate over time and people to gain the context they need.