Identification of PDF preservation risks with Apache Preflight: a first impression

By johan on 19 December 2012 – 3:15pm

The PDF format contains various features that may make it difficult to access content that is stored in this format in the long term. Examples include (but are not limited to):

  • Encryption features, which may either restrict some functionality (copying, printing) or make files inaccessible altogether.
  • Multimedia features (embedded multimedia objects may be subject to format obsolescence)
  • Reliance on external features (e.g. non-embedded fonts, or references to external documents)

A more exhaustive overview is given here:

http://www.openplanetsfoundation.org/system/files/PDFInventoryPreservationRisks_0_2_0.pdf

and also here:

http://libraries.stackexchange.com/questions/964/what-preservation-risks-are-associated-with-the-pdf-file-format

When creating a PDF, it is possible to minimise these risks by using one of the PDF/A standards, which delineate a number of PDF feature profiles that are unlikely to result in any long-term accessibility problems. However, the simple fact is that most PDFs that are out there are not PDF/A.

PDF Profiling

For assessing risks in existing collections, it would be helpful to be able to screen or profile PDFs for specific ‘risky’ features, such as encryption or font embedding. Since PDF/A was specifically designed to eliminate these ‘risky’ features, one would expect that PDF/A validators (i.e. software tools that check the conformance of a PDF file against the PDF/A specification) would be able to provide some useful information on this.

In a first attempt to test whether this approach is feasible at all, I did some tests with Apache Preflight, an open-source PDF/A-1 validator that is part of the Apache PDFBox library.The specific objectives of this work were:

  • To get a first impression of the Apache Preflight (part of PDFBox) PDF/A-1b validator.
  • To investigate if Apache Preflight is able to detect unwanted (from a preservation point of view) features in PDF files (i.e. PDFs that are not necessarily of the PDF/A sub-type) such as password protection, encryption and non-embedded fonts.
  • To provide a comparison with the Preflight module of Adobe Acrobat 9.5.
  • To decide if doing more work on Apache Preflight (more elaborate testing, possible involvement in its development) are worthwhile.

The results can be found in the report Identification of preservation risks in PDF with Apache Preflight: a first impression.

The Archivist’s PDF Cabinet of Horrors

The report’s findings are to a large extent based on a suite of small, simple test files that were created especialy for this work. Each file contains one ‘risky’ feature, with focus on the following feature classes:

  • Encryption
  • Multimedia
  • Scripts
  • Fonts
  • File attachments
  • External references
  • Byte corruption

The dataset can be found here:

http://www.opf-labs.org/format-corpus/pdfCabinetOfHorrors/

Link to full report

Identification of preservation risks in PDF with Apache Preflight: a first impression

Update (January 2013)

Since the report was published, a number of improvements have been made to Apache Preflight which should fix some of the reported issues. I haven’t tested the latest version yet, but will try doing this some time soon.

Preservation Topics: 
Characterisation
Preservation Risks
Representation Information
Corpora
Tools
SCAPE