Sustaining Engineering Informatics

1
Sustaining Engineering Informatics

• Towards Methods and Metrics for Digital Curation
• Joshua Lubell, Sudarsan Rachuri, Eswaran
  Subrahmanian, Mahesh ManiNational Institute of
  Standards and Technologylubell, sudarsan,
  eswaran, mahesh_at_nist.gov

2
NIST Workshop April 24-25, 2007

• Long Term Sustainment of Digital Information
  for Science and Engineering Putting the Pieces
  Together
• Over 30 participants
• Implementers of OAIS
• Government (NARA, Library of Congress, Navy,
  Govt. Printing Office)
• Universities
• Third in a series of workshops
• NIST (March 2006)
• Bath, UK (February 2007)

3
Breakout Issues Archival Information and
Technology
4
Breakout Issues Standards and Specific Domains
5
Overall Workshop Conclusions

• Facilities for archiving should be available at
  the source of information creation
• Archival systems must deliver the right
  information for the task at hand to the end user
• Archival system design is a socio-technical
  problem

6
Archival Challenges Unique toEngineering Design

• From Kopena, Shaffer, Regli CAD Archives Based
  on OAIS, Proceedings of ASME Computers and
  Information in Engineering Conference,
  DETC2006-99675, Philadelphia, September 2006
• Capturing all aspects of a design project
• Preserving data generated by software tools
• Predicting how data will be used over long term
• Package schemas tailored to CAD preservation
  needed
• Standards for representation information helpful,
  but not a silver bullet

7
Library of Congress Digital Format Sustainability
Factors

• See digitalpreservation.gov/formats
• Disclosure availability of documentation,
  validation s/w
• Adoption - popularity
• Transparency analysis possible without
  specialized tools?
• Self-documentation metadata included in digital
  object?
• External dependencies specialized software
  needed to use objects?
• Impact of patents
• Technical protection mechanisms access
  restrictions

8
Sustainability Factors Applied to ISO 10303 (STEP)

• Disclosure
• ? International standard
• ? Third party documentation
• ? Validation software
• Adoption
• ? CAD vendors
• ? Other domains
• Transparency
• ? ASCII and graphical formats
• Self-documentation
• ? Rich model-based representation
• External dependencies
• ? None as long as software does import/export
• Impact of patents
• ? None
• Technical protection mechanisms
• ? None

9
So What's Next?

• Sustainability factors not domain-specific
• Potential future access scenarios not taken into
  account
• We need more metrics for STEP and other
  engineering digital objects
• Quality and functionality factors a start
• Question How would you measure the quality of an
  engineering archiving and/or records management
  strategy?

10
Access Scenarios The Three Rs

• Reference
• Preserve information in its original state
• Example (product data engineering) 3D
  visualization
• Reuse
• Allow for future modification, re-engineering
• Example ISO 10303-2031994 (STEP AP203)
• Rationale
• Encode construction history, design intent,
  tolerancing info, lifecycle management info, etc.
• Example STEP AP203 ed.2
• Ontologies and/or other representations needed

11
The 3Rs STEP
12
Extended Functional Model
EI Engineering Informatics DOP Digital Object
Prototype METS Metadata Encoding Transmission
Standard
13
Future Goals

• Formalize 3Rs
• Define EI sustainability metrics
• Create EI implementation framework
• More generic than LOTAR (a specification for long
  term archival of 3D-CAD and associated product
  data), but more domain-specific than vanilla OAIS
• Develop EI archival testbed