A Holistic Approach to Bit Preservation

This thesis presents three main results for a holistic approach to bit preservation, where the ultimate goal is to find the optimal bit preservation strategy for specific digital material that must be digitally preserved. Digital material consists of sequences of bits, where a bit is a binary digit which can have the value 0 or 1. Bit preservation must ensure that the bits remain intact and readable in the future, but bit preservation is not concerned with how bits can be interpreted as e.g. an image. A holistic approach to bit preservation includes aspects that influence the final choice of a bit preservation strategy. This can be aspects of how the permanent access to the digital material must be ensured. It can also be aspects of how the material must be treated as part of using it. This includes aspects related to how the digital material to be bit preserved is represented, as well as requirements for confidentiality, availability, costs, additional to the requirements of ensuring bit safety. A few examples are:
• The way that digital material is represented in files and structures has an influence on whether it is possible to interpret and use the bits at a later stage. Consequentially, the way bits represent digital material influences the entire preservation of the digital material.
• The file formats can be more or less vulnerable to bit errors. Different file formats will therefore require different bit preservation solutions in order to obtain as the same maximum level of preventing risks of losing the digital material.
• The file formats can consume more or less storage volume, and the way the material is produced can influence storage volume. Thus the chosen representation and production can influence the requirements for storage volume needed in bit preservation, which influences the costs.
• There will be requirements for the availability of the bit preserved digital material in order to meet requirements on use of the digital material, e.g. libraries often need to give fast access to preserved digital material to the public, i.e. the availability of the bit preserved material must support the use.
• The digital material must be preserved in a way that satisfies general requirements for the digital material. For instance, confidentiality requirements for confidential material. Thus these requirements influence the requirements for the bit preservation.
The examples show that it is relevant to take a holistic approach and include aspects of digital representation, confidentiality, availability, bit safety and costs when defining requirements for the bit preservation. Analysis of such requirements and choice of the final bit preservation solution can be supported by the three main results presented in this thesis:
• First, a model to assist in analysis of a delimited and possibly shared bit repository which has to ensure preservation of bits. The model is the basis for defining terminology as well as the basis for analysis of functions within the bit repository and interface to the bit repository. This includes both technical and organisational aspects of a bit repository.
• Second, a representation concept to assist in analysis and design of representation of digital material under bit preservation. The concept is partly inspired by observations of possible conflicts between requirements for preservation and requirements for dissemination. The concept also includes aspects of representation assisting future use. It is furthermore inspired by results which show that decisions on the digitisation process can have impact on bit preservation costs.
• Third, a methodology to assist in evaluation of the best choice of a bit preservation solution, which can best meet the requirements for specific digital material. Considerations in selection of a bit preservation solution include many of the aspects from the holistic approach where various requirements for the bit preservation must be taken into account. The evaluation can be needed at different stages, e.g. in re-evaluation when a media migration is performed.