Workbook on Digital Private Papers > Digital preservation strategies > Degree of preservation

Degree of preservation

A digital object, such as the kind of typical object found in a personal archive (e.g. a speech written in a word-processed document, a budget in a spreadsheet, or a digital image of a family) is defined in the PREMIS Data Dictionary as a discrete unit of information in digital form which is comprised of three different levels:

  1. Bitstream: in its simplest form, a digital object consists of a bitstream, i.e. an ordered sequence of bits (binary zeros and ones). This binary inscription will usually be stored on a physical medium of some kind. A computer system with the correct combination of hardware or software translates the bitstream into something meaningful.
  2. File: a named and ordered sequence of bytes known by an operating system. The format of a file is laid out in the format specification, which transforms the file from its binary ones and zeros into something which makes sense to a user, i.e. stipulating the proper encoding, sequence, arrangement, size and internal relationships which enable the construction of a valid file of the relevant type (e.g. gif 1989a file). This level represents the transformation of the input bitstream into output for presentation purposes; the physical medium on which the bitstream is inscribed is therefore of no consequence at this level.
  3. Representation: denotes the set of files needed for a complete and reasonable rendition of an Intellectual Entity. This is defined by PREMIS as a coherent set of content that can reasonably be described as a unit; this is essentially a conceptual object, or something that a human can understand as a meaningful unit of information, e.g. a website, a report, a photograph. An Intellectual Entity or conceptual object may have one or more digital representations or encodings; for instance, the text of a politician's speech might be saved as both a Microsoft Word document and a PDF file. The underlying encoding of each will differ considerably, but the textual content of each item is identical. A Representation may also be made up of one or more Files, and it is important that the relationships between such component files are clear.

Simply preserving the bitstream therefore does not guarantee ongoing access to a digital object. The digital object has an existence separate from the medium on which the bitstream is inscribed, and successful preservation is only complete when all the significant properties are maintained and the digital object can be displayed in a meaningful and understandable form. In some cases, this might mean that only the intellectual content (e.g. the text of a word-processed document) is preserved, and that original formatting and layout is not retained in the preserved object. However, in other cases (notably, but not exclusively, complex objects such as interactive resources) it may be important that the 'look and feel' of the original object, and its functionality, is retained or recreated as part of the preservation process.

This point is usefully illustrated by the definition of digital records propounded by the National Archives of Australia: digital records are viewed as 'performances' - the result of interplay between technology and data. A digital archivist is not primarily interested in preserving the physical bitstream, though without this nothing else is possible. The aim is to capture an acceptable representation of the fleeting and temporary performance of the record on screen as it was originally created, viewed and edited by the individual whose archive is being preserved.