Towards a Semantic Preservation System

TL;DR

This paper discusses a system that preserves access to file content by extracting meaningful logical structures using a standard format description language, reducing long-term preservation costs.

Contribution

It introduces the Defuddle parser, a generic tool that utilizes DFDL to extract logical structures from various file formats, enhancing preservation capabilities.

Findings

Provides a universal approach to logical content extraction

Reduces software and maintenance costs for long-term preservation

Separates bits, formats, and logical content issues effectively

Abstract

Preserving access to file content requires preserving not just bits but also meaningful logical structures. The ongoing development of the Data Format Description Language (DFDL) is a completely general standard that addresses this need. The Defuddle parser is a generic parser that can use DFDL-style format descriptions to extract logical structures from ASCII or binary files written in those formats. DFDL and Defuddle provide a preservation capability that has minimal format-specific software and cleanly separates issues related to bits, formats, and logical content. Such a system has the potential to greatly reduce overall system development and maintenance costs as well as the per-file-format costs for long term preservation.