Validation Framework

download [Download]
Published on: 
Wed, 2010-05-26
Short Overview: Software and Release Report of the validation Framework This document is the release report of the validation framework developed in the Planets project, workpackage PP5. The validation framework connects the practical outputs of other Planets work-packages, mainly those of PP4, PC2 and PC4, within a common framework: Preservation requirements and criteria as they are defined in the objective trees in PP4 are mapped to the technical characteristics described by XCL in PC2/4, and to other criteria such as performance and format risks.

The validation framework consists of:
  1. Comparison measures as basis for computing simple measurements or aggregated measurements over several low-level characteristics
  2. A mapping mechanism for connecting criteria to measurements related to technical characteristics of both digital objects and tools,
  3. Extension of the measurements to cover other measurable criteria of interest such as format risks and runtime performance.
Aspects (1) and (2) were initially described in report PP5-D1, Specification of basic metric and evaluation framework [5]. In the final version they are fully deployed within Planets software components. (1) is a fundamental part of the Comparator tool, related to the XCL and its components. (2) and (3) are part of the Planets Preservation Planning Tool (Plato). Since all parts are strongly bound to Plato and the Comparator, the specifications of the comparison measures and the evaluation framework are also integrated in the documentation part of these tools ( and ).

This document briefly summarizes central issues of the validation framework and describes the most important advancements of its parts: The comparison measures have been expanded in order to enable comparison of text related information and to give a summary result, i.e. indicating if a migration of content A to content B has been successful or not. The evaluation framework is now integrated into Plato so that it extends to other technical aspects of interest, such as run-time performance and format risks. To this end, we have developed a prototype of a quality-aware migration engine (3) that is shortly described.