Steps Towards a Theory of Information Preservation

Giorgos Flouris
Carlo Meghini

Istituto della Scienza e delle Tecnologie della Informazione (ISTI), CNR
Via Giuseppe Moruzzi, 1, 56124, Pisa, Italy
{flouris,carlo.meghini}@isti.cnr.it

Abstract
The rapid obsolescence of large volumes of digital data is one of the 
most difficult problems faced by modern archivists. This problem 
(information preservation) is not fully understood to date; one of the 
major gaps in related research is that there is no general agreement on 
a formal model to describe the problem or on a formal description of the 
required properties of a good solution to the problem.
This work's aim is to fill this gap by developing a formal, mathematical, 
logic-based description of preservation as a scientific discipline, to 
the end of deriving a methodology resting on solid theoretical ground. 
Thus, this work is not about how to perform preservation, but about how 
to descibe formally what it means to perform preservation.
We will present a number of preliminary results towards this aim, 
including the identification of three types of preservation, some 
connections with the related fields of belief change (belief revision) 
and ontology evolution, and some desirable properties of a formal 
description of preservation. Our study will show that it is not usually 
necessary (or possible) to preserve the whole digital object (e.g., a 
database) but a set of properties (questions and their answers) related 
to the object, and will verify the need to separate the data from its 
environment (i.e., the external information which gives meaning to the 
data). Both issues are considered in our proposal towards a formal 
description which handles both the static and the dynamic aspects 
of the problem.