Methods and apparatus, including computer program products, for managing data. The technique includes defining a secondary data set in a primary data set, linking the secondary data set to state information of the primary data set, transmitting the secondary data set to an external entity, receiving a modified secondary data set from the external entity and linking the modified secondary data set to a version of the primary data set stored with reference to state the information. The state information defines the state of the primary data set at the time the secondary data set is defined. The link between the modified secondary data set and the primary data set allows access by a user of a modified primary data set if the primary data set has changed with reference to the state information.