Though not especially media-rich, mailing lists remain in use and retain popularity for their built-in technological controls and their capability to “reply” to a message (“continuing a thread”). The motivation for extracting knowledge fragments from the unstructured text of mailing lists is compelling, though successes doing so may be considered only partial because it requires mental processing, or a certain cognitive effort, that complicates automation. Cognitive psychology distinguishes the Long Term Memory (LTM), which may be compared to text thread storage, from the Working Memory (WM), which initiates the retrieval of knowledge fragments stored in the LTM. Searching by subject, date, and time stamp ranges, and by keyword-inclusive fragments, constitutes the commonly used methods for executing sequential LTM retrieval. Retrieval can, however, be greatly enhanced by automatically gleaning certain classes of threads from the entire structure and displaying them alongside other properties. Here, we describe automatic “class” extraction and its effect on OM manufacturing and LTM retrieval.