Format risk analysis: a preliminary interface implementation

The result of a file format identification, or an indication of a format change, can suggest a file is at risk of becoming inaccessible or corrupted. Preserv has developed a repository interface to present formats by risk category, and to enable information on high risk formats and files at risk to be displayed.

The facility for adding risk assessment data is provided in the latest version of PRONOM (v7), which became available to the project during March 2009, the very end of the project. However, the functionality of this version was emulated in advance with the PronomStubCode developed by Dave Tarrant at Southampton University. The risk interfaces illustrated below were first developed using this code, after which it was a simple procedure to re-direct the repository classification data (results from the Preserv2 EPrints toolkit, including DROID) to the live version of PRONOM and test this. Two minor changes were needed to make the risk analysis module in EPrints work with the new service, but Preserv2 has shown that providing risk scores through the live PRONOM service is feasible.

The risk scores presented are entirely hypothetical, however. It was expected that risk assessments of selected file formats found commonly in repositories (see pie chart on the previous page for the breakdown of file formats in the 'typical' repository) would be included in this release, but format risk assessments are very detailed (with dozens of risk factors), requiring skill and expert scrutiny, and this was not done within the project timescale. It is hoped that format risk scores will be provided through PRONOM or other publicly available services in the near future.

Hierarchy of (hypothetical) format risk scores for an EPrints repository

How high risk formats and files might be presented