R\'eflexion sur quelques points essentiels qui pourraient contribuer \`a d\'efinir la Science de l?information : posture \'epist\'emologique ; domaine d?\'etude de la science de l?information et sp\'ecificit\'e disciplinaire ; situation au sein de l?ensemble des sciences de l?information et de la communication ; cadre th\'eorique de r\'ef\'erence.

Information uncertainty is inherent in many problems and is often subtle and complicated to understand. While visualization is a powerful means for exploring and understanding information, information uncertainty visualization is ad hoc and not widespread. This paper identifies two main barriers to the uptake of information uncertainty visualization: firstly, the difficulty of modeling and propagating the uncertainty information; and secondly, the difficulty of mapping uncertainty to visual elements. To overcome these barriers, we extend the spreadsheet paradigm to encapsulate uncertainty details within cells. This creates an inherent awareness of the uncertainty associated with each variable. The spreadsheet can hide the uncertainty details, enabling the user to think simply in terms of variables. Furthermore, the system can aid with automated propagation of uncertainty information, since it is intrinsically aware of the uncertainty. The system also enables mapping the encapsulated uncertainty to visual elements via the formula language and a visualization sheet. Support for such low-level visual mapping provides flexibility to explore new techniques for information uncertainty visualization.