Session 12

Digital Preservation, Dale Flecker, Harvard University

Digital preservation is emerging as a major problem. Digital objects are extremely fragile because they require mediation by hardware and software that is continually changing. Older technology can and will break, while new technology frequently does not understand old formats. Consequently, digital objects require great care to maintain and preserve over time.

A sound digital preservation strategy will pay close attention to format, metadata and technical infrastructure.

Format and Metadata

In attempting to address the problem of digital preservation it is important to think up front about the formatting and documentation of the digital objects that one creates. In a best case scenario an archival master, a production master and a use copy of each digital object should be created. The creation, format, format options and use requirements for these copies should be well-documented with good technical metadata. In addition structural metadata should be created which outlines the potential effects of any changes made to a finished digital project and administrative metadata should be created to detail exactly who has the right to make decisions about a completed project. Good metadata will also note the provenance of a given digital object, whether or not there are other versions of the digital object and how these other versions might affect one’s preservation strategy.

Technical Infrastructure

Backing up one’s digitization project to a disk or cd is not a sufficient preservation strategy. A more rigorous technical infrastructure is required, one which might include a digital repository. The functions of a digital repository are to:

It is also important to maintain a formal and disciplined approach to operations within the technical infrastructure. This includes:

There is no concrete definition of what a sound digital preservation program is, outside of the understanding that digital preservation requires an on-going process of auditing and refreshing the materials in one’s repository, while taking care to make multiple copies of the data that one creates.