The Task Segment Framework (TSF) is a tool to analyze full typing flows of translation tasks as keylogged with timestamps recorded for keydown, keyup, mouse clicks and moves, and actions performed in other applications. The TSF assumes that intentional pauses flag stretches where subjects concentrate on unrecorded cognitive processes such as planning and assessment. The interspersed typing stretches are task segments, with or without text, where basic subtasks may be observed, mainly adding new text, changing existing text, and searching for information. Accumulated experience and planning allow translators to lump strategically similar activities together, in order to spare efforts and task switching costs while maximizing efficiency. Hence, task segments may contain activities of just one such subtask or many. Translation fluency is a key notion of the TSF, operationalized through many indicators such as typing speed, prior pause length, TS (task segment) length in events, text length as full words, number of typos and respites (=mid inter-keystroke intervals), subtask(s), and the like. The approach seems particularly sensitive to translation expertise levels and may be applied with variations to other multilectal mediated communication tasks. This article lays down the conceptual basis of the TSF and summarizes its basic notions and constructs.