Self-Contained Digitally Signed Documents: Approaching "What You See Is What You Sign"

The "what you see is what you sign" challenge has been part of digital signatures since the very start. Digital signatures apply to the bit level. Users see a higher level, so how can they know what they sign? A sample of real-life applications indicates that the issue is still open. We propose a method for improved assurance based on simple tenets. The document to be signed is a well-defined visual impression. Exactly that visual impression is signed. After signing all parties have a copy of the signed document, including its signatures. PDF makes it possible to store signatures and metadata in the document. The method is being implemented in an e-government web platform for a major Swedish city.