Preserving useful digital objects for the future

Abstract: This thesis deals with the preservation and presentation of physical structure (i.e. visual appearance) of digital information objects. The research is focused on digital information objects of a character that is common in e-services and e-government, namely web forms or receipts that exists in large numbers with similar form. The reason for this is the increasing number of e-government services that will produce records that need to be preserved. The main proposition is that physical structure is important to preserve in order to facilitate interpretation of information objects. Another proposition is that this physical structure is possible to describe with metadata. These propositions were challenged, and the concepts of trustworthiness, understandability, accessibility, availability and usefulness were discussed empirically with information experts (archivists), system developers, experienced users (genealogists) and occasional users, around a demonstrator object as base for the discussions. The demonstrator object was developed in an iterative process during the thesis work, according to inputs from both literature and empirical studies, and eventually a minimized metadata description of physical structure was found. This minimized structure represents a compromise between original look and manageability in the preservation process, as well as desires from the users and experts queried during the research process. Not surprisingly, trustworthiness was found to be mostly a matter of documentation of the preservation processes and the history of the information object, and even if an "authentic look" slightly improved the feel of trust, it was only considered as a "false sense" of trust. Regarding usefulness, the separation of data from physical structure, and the ability to present the material in different forms was seen as useful by all categories of respondents. The demonstrator object showed that the separation worked out sufficiently well, and as intended, it was seen as a suitable approach to presentation of documents of a textual character that exists in large numbers with similar form. So, in short, keep data and physical structure separated for maximum usefulness and flexibility; find a middle way of re-presenting the physical structure for the type of document in question; document everything that the information object is subjected to, in order to provide both trust and usefulness to the future users.