From the Intro:
The selection of appropriate file formats is likely to be a significant factor in the survival of information objects. Digital information depends on hardware and software to make it comprehensible, so changes in hardware and software mean that long-term access requires a degree of coordination between original data and current facilities. There are different ways to create this coordination to make data accessible, but all approaches assume a firm grasp of the conventions used to encode it in the first place.
This report summarizes trends and issues in file format selection as they are articulated in the digital preservation literature at the time of writing (2008-9). This reflects the current status of digital preservation implementation: some of it concerns current practice from functioning repositories while more developmental approaches continue to emanate from the research community. Pronouncements by operational archives are more frequently in the form of their submission guidelines than a full statement of their ongoing preservation formats, though the two are related. As a result, whilst this report concentrates on formats of files being managed by archives (Archival Information Packages in OAIS terminology), it also considers literature about preferred submission formats (i.e. within Submission Information Packages or SIPs). Dissemination formats (Dissemination Information Packages or DIPs) are given less consideration.
The report considers file formats in the context of other elements of a digital preservation infrastructure, especially representation registries, archive governance and digital preservation strategies. There is interdependency between these issues. Some of the current literature appears to minimise this fact by treating format migration or „conversion‟ almost in isolation1. To understand why this might be, it is helpful to review the Open Archival Information System information model and particularly its terminology. OAIS does not generally concern itself with the issues discussed in much of this report, but this does not mean that the deeper implications of the OAIS information model do not need to be taken into account when considering file format issues.
Download the report. [pdf 869 kb]
See other DPC tech watch reports.