In a pen-based computing system, multi-modal data is transferred between a paper domain and a digital domain. Data initially generated in the paper domain is captured by a smart pen and a digital file including the captured data is generated. For example, a computing system coupled to the smart pen generates a digital file including the captured data. A paper representation of the digital file is subsequently generated. The digital file can subsequently be modified by editing the paper representation of the digital file using the smart pen. Edits to the paper representation of the digital file are captured by the smart pen and converted to the digital domain where they are used to edit the content of the digital file.