Preservation Guidelines

The following points are based on a presentation by MacKenzie Smith at the CERN workshop on Innovations in Scholarly Communications in April of 2007.

What can go wrong...

• Systems component failure - server, network, software, 3rd party services, or entire data center

• Media faults (e.g. CD-ROM bit rot)

• Software and data format obsolescence

• Human error - deletion

• Policy changes with bad trigger effects

• Natural disaster

• Network intrusion and attacks

• Copyright restrictions - accidentally violating law by preservation activities

• DRM - encryption prevents preservation, reduplication, and migration

• Economic failure

• Organization failures - an institution or department can go out of business or eliminate positions or change mission.

Preservation Services provides for...

• Replication - multiple copies in multiple places (LOCKSS; use OAI)

• Migration - of hardware, storage media, collections

• Transparency - open source software, [meta]data (may need to be able to fix something)

• Diversity - use multiple formats (Word, ODF, PDF and ASCII)

• Auditing - checksum checking, file reads (detect and fix errors in data)

Repair Services

• Detect errors in the data and automatically restore it from secondary storage.

HDR Preservation Policies

Preservation Planning

• Monitor / interact with the designated community to understand requirements and changes

• Monitor the emerging technology and standards

• Develop and recommend preservation strategies

• Develop packaging designs and detailed migration plans and prototypes

• Implement administrative policies and directives.

Short-Term Preservation Guidelines

• Scan objects according to accepted best practices.

• Store objects in a central location.

• Create minimal metadata in a database that includes image and text filenames.

• Print copies of your digital images onto archival quality photographic paper

• Use Xena (xeni.sourceforge.net/) to convert your documents into a format that will be readable in the future.

• Make regular backups, ideally onto a removable hard disk drive.

• Update file formats when you upgrade the software that created them.

Selection Requirements

• Content must be curriculum or research related

• Content must be of long-term interest

• Content must have been created by the Hamilton community

• Content must be copyright-free

Recommendations for File Formats for Time of Submission

• Text documents may be submitted in their original formats (even if proprietary), such as Microsoft Word, but should also be submitted in PDF (150 dpi) and/or TIFF format (300 dpi).

• Images should be submitted in these formats (order of preference)

  1. TIFF format (200-600 dpi; slides: 3000-4000 dpi)
  2. JPEG2000 (ask about specifications)
  3. JPEG (high dpi as possible, 72 dpi min.)
  4. PNG (high dpi is possible, 72 dpi min.)

• PDF must meet these requirements
  • Files must not be password protected.
  • Text files must not have embedded fonts
  • If special fonts are required a copy of the font must be archived with the object

Recommendations for File Formats for Long-term Retention

The principles behind the file formats recommended for long-term retention (i.e., more than 5 years) of digital objects for in the custody of the Hamilton College Library are these:

• the format must assure the preservation of the content, that is the completeness of all necessary data and links between them
• the format must be widely acknowledged and accepted and can be used with hardware and software widely used and supported on the market
• the format must be directly ready for reproduction or migration to another format that can be directly used afterward in the new format
• the format must enable transformation from most commonly used formats (to this format) and will facilitate automatic reporting of any errors that occurred during transformations
• the format must is independent of specific hardware, software or other components of the operating environment
• the format must is based on international, national or widely accepted and, as a rule, open standard
• use of the format must not violate and laws

The current list of allowed long-term retention formats for all archival material in the Hamilton College Digital Collections Repository:

Images

• TIFF (TIFF 6.0, Adobe Systems)
• JPEG2000 (.jp2) ISO/IEC 15444-1

Text and mixed:

• ISO Latin-1 – ISO 8859-1
• PDF/A - ISO 19005-1
• XML - SGML – ISO 8879
• TEI - an application of SGML ISO 8879
• ODF – ISO/IEC 26300

Graphics:

• SVG – v1.1 – W3C Specification

Video/audio:

• MPEG-2 - ISO/IEC 13818
• MPEG-4 - ISO/IEC 14496