Abstract

Medium- to long-term archiving of digital documents, beyond the lifespan of the authoring software/hardware, is a challenging problem. Magnetic and optical media are susceptible to environmental influences and deteriorate over time, often to the point where the archived documents can no longer be retrieved. Previous attempts to address this problem include migration and emulation, both of which have their attendant difficulties. It is the contention of the present study that an Open Source approach offers several advantages. More specifically, by archiving the Open Source application programs (in source code, not executable form) along with the documents in question, in both plain and compressed form, significantly increases the likelihood of being able to retrieve such archives at some future time. The application source code can be recompiled to a form suitable for reading in (Open Source) viewers, thereby presenting to the user the archived document as the original author envisaged it. One set of experiments was undertaken distributing documents together with their (Open Source) authoring software via a Portable Virtual Machine (PVM) program to unused disk space on a network of SUN workstations. The success of this approach was evaluated using the following four measures: (i) lossiness of conversion, (ii) edit-ability, (iii) ability to save back to the original format, and (iv) functionality retention. Another series of experiments was conducted in which artificial (‘speckle’ or salt-and-pepper) noise was deliberately introduced to the archived documents in order to mimic degradation of the storage medium over time. It was found that survivability was heavily dependent on file type: simple text files and MPEG movies were impervious to even 18% introduced noise. Source code programs and JPEG images, by contrast, were intolerant to even the smallest noise levels (it has to be said however that straightforward re-editing of the former led to error-free compilation without much difficulty). Lastly, it was found that decompression (specifically the publicly available RAR decompressor) further enhanced the file recovery process. We conclude that an Open Source approach to the preservation of digital archives has considerable potential.

Share

COinS