Understanding published research results should be through one's own eyes and include the opportunity to work with raw diffraction data to check the various decisions made in the analyses by the original authors. Today, preserving raw diffraction data is technically and organizationally viable at a growing number of data archives, both centralized and distributed, which are empowered to register data sets and obtain a preservation descriptor, typically a `digital object identifier'. This introduces an important role of preserving raw data, namely understanding where we fail in or could improve our analyses. Individual science area case studies in crystallography are provided.
