Practical Experiences of the

1

Practical Experiences of the Digital Preservation
Testbed Jacqueline Slats Public Record
Office April 3rd 2003
2

• Objective of Digital Longevity
• Securing the accessibility of reliable
  governmentinformation in the digital era.

3

• Objective of Digital Preservation Testbed
• Securing the sustained accessibility of reliable
  government information in the digital era.

4

• Target group
• The government

5

• Coming about of the program Digital Preservation
  Testbed
• Increasingly, the government is working with
  digital documents
• and other records
• Dutch Cabinet objective 65 of all transactions
  between government and citizens must take place
  using digital meansby 2006
• Continuity of the working processes will be
  maintained in the digital environmentas well.

6

• Coming about of the Testbed program (cont.)
• Digital information must be accessible, reliable
  and interpretable
• Strategies, methods and techniques are required
  to achieve this.

7
The problem of preservation Technological pace
of change and obsolescence there is no guarantee
that current computer environments and records
can be read in future.
8

• Digital Preservation Testbed is performing
  experiments on three strategies for preserving
  records without affecting the authenticity of the
  records
• - Migration
• - Emulation (Universal Virtual Computer)
• - XML
• Is assessing their practical use for theDutch
  situation.

9

• Experiments are taking place on
• text documents
• spreadsheets
• electronic mail
• simple databases.

10
Data Preservation is the first and simplest mode
of operation of the UVC approach
11
Research Questions

• Advantages of each preservation approach?
• Factors affecting each approach?
• Effectiveness of each approach?
• Basic Requirements for Preservation
• Which metadata are essential for preservation?

12
Experiment Process (1)

• Step 1 Definition of the process
• Step 2 Preparation for the process
• Step 3 Authenticity requirements and evaluation
  checklist
• Step 4 Design of the experiments
• Step 5 Resource specification
• Step 6 Go/no go decision

13
Experiment Process (2)

• Step 7 Development of the experiment
• Step 8 Test experiment
• Step 9 Go/no go decision
• Step 10 Run experiment
• Step 11 Evaluate experiment
• Step 12 Consider results

14
Testbed Team
15

• Migration
• the conversion of records from one hardware
  and/or software environment to another
• currently the most common preservation strategy
  for digitalrecords.

16
The UVC emulation approach defined by IBM
Research develops a flexible emulator used to
interpret both data format semantics and emulate
hardware platforms

• The advantages of the IBM Research UVC approach1
• The existence of the UVC drastically simplifies
  the emulation problem
• For data preservation, it eliminates the need for
  agreeing on standardised formats.
• Only the UVC interpreter will have to be written
  when a new architecture emerges. There is no
  impact on the archived information
• The UVC is made so general, and at the same time
  so basic, that its definition will remain
  relevant in the future

• Basic architecture
• Basic instruction set

Universal Virtual Computer (UVC)
1) Long-Term Archiving of Digital Information,
Raymond Lorie, IBM Almaden Research Center, RJ
10185 (95059), March 28, 2000
17
XML (eXtensible Markup Language) XML is
independent of the hard- and software used, and
is thereforeideal for information exchange for
future as well as presentconfigurations.
18
(No Transcript)
19
Text documents (1)

• Results

Process 1(C)

• For active records needs to be repeated every
  few years manual checking can only be done if
  migration is automated

• Migration from an older version of an application
  to a newer version of this application

• Migration from an application to a standard
  format PDF and RTF

• PDF is able to preserve text records RTF failed
  to represent the original record

• Migration of old records created in one word
  processor to another (WP to Word)

• Met authenticity requirements after manual
  intervention not a format for digital longevity

20
Text documents (2)

• Results
• Did not meet our authenticity requirements
  unsuitable for preservation

• Process 4 (C)
• Viewers

• Process 7 (A)
• Conversion to XML

• Results
• First results with Open Office look promising,

21
Email

• Results
• Met authenticity requirements binary attachments
  encoded (Base64)

• Process 2 (C)
• Conversion XML (wrapper approach)

• Process 3 (C)
• Conversion to XML (backbone approach)

• Results
• Met authenticity requirements attachments in
  their original file format

22
Spreadsheets
Experiment process
Results

• Emulation UVC data preservation approach

• Generating the logical data description is very
  difficult

• For active records needs to be repeated every
  few years manual checking can only be done if
  migration is automated

• Migration from an older version of an application
  to a newer version of this application

• Promising approach, but XML produced by Open
  Office and Excel XML not meaningful enough

• Conversion to XML

23
Databases
Experiment process
Results

• Content and structure can be preserved

• Conversion to XML

• Migration of large databases

• No results available yet

• No results available yet

• Migration of small databases

24

• What will the Testbed project provide
• advice on how to deal with current digital
  records
• recommendations for an appropriate preservation
  strategy or a combination of strategies
• functional requirements for a preservation
  function
• cost models of the various preservation
  strategies
• decision model for preservation strategies
• recommendations concerning guidelines and
  regulations.

25
For further information about the
Testbed Website www.digitaleduurzaamheid.nl
e-mail testbed_at_ictu.nl
26
Demonstration
27

Practical Experiences of the Digital Preservation
Testbed Jacqueline Slats Public Record
Office April 3rd 2003