Multilingual computer text handling appears at present unapproachably difficult. Despite considerable progress made over the past three decades, major open problems remain in encoding, interactively editing, programmatically transforming, and displaying text in many of the world’s writing systems. These challenges are especially pronounced with regard to ancient and minority or minoritized writing systems and are exacerbated by antiquated, retrofitted software and standards. In this paper, I offer a genealogical critique of how computer scientists have approached human writing hitherto, a new, clarifying perspective on writing as it relates to computers, and a sketch of how the problems I identify herein could be addressed.