Speaker: Ben Miller, MIT.  Moderator: Nick Seaver, MIT. Abstract: Visual modeling of databases, when applied to the problematic of witnessing traumatic events, offers a way to structure testimony’s logical relations and text. Without the external logic offered by data markup languages and the textual structures of coding rubrics, witness reports of violence have proven vulnerable to insinuations of inaccuracy. UML helps turn a fragment, such as a witness's remark about unusual activity near a pond, into an evidentiary network of material regarding a mass grave. Precipitated by a process of visualization, this work of aggregation and association is altering the nature of collective memory and is exemplified in the way testimony on rights violations are categorized and measured using a rubric known by the acronym HURIDOCS. The Human Rights Documentation System offers a quantitative coding rubric for field workers responding to violence, and embeds that rubric in a system of relations focusing on "People" and "Acts." This widespread numerical frame is enabling the ongoing juridical response to the Cambodian Genocide of 1975-1979 by logically translating symbolic, mimetic, and indexical evidence into a database subsumed within a technical apparatus of fields, parents, children, siblings, relations, and codes, such as 09.1, "Possible Disappearance," and 05.781, "Change of repressor role to that of ally in order to disorient the victim." Reliant upon a visualized logic, this free play of translation of text to code to text, and the ability to formalize and elicit evocative relationships among that material is a hallmark of the database systems currently reshaping the collective mnemonics of trauma.