Yale-Elsevier Mellon Project

1
Yale-Elsevier Mellon Project

• NISO/BISG Digital Archiving
• ALA, New Orleans, January 20, 2002
• Karen Hunter
• Senior Vice President, Strategy
• Elsevier Science

2
Context

• Yale University Library
• Elsevier Science
• previous joint efforts

3
Yale Library as a player

• history of electronic resources both on its own
  and with the NorthEast Research Libraries (NERL)
• electronic expenditures of 1.4 million in 2000,
  1.8 million in 2001
• clear that cannot endlessly continue to support
  duplicate collection of paper and electronic

4
Elsevier Science as a player

• e-journal interest dates to late 1970s
• formal archiving policy in 1999 as part of
  ScienceDirect license
• guarantee that would maintain archive or transfer
  to library-vetted repository
• customer feedback neutral archive now (i.e.,
  in library hands)
• also needed to protect authors interests
• explored various archiving arrangements

5
Joint efforts

• between 1997 and 1999 Yale and ES had explored
  the possibility of Yale being a local repository
  for some or all of the 1,100 ES journals
• end decision license for SD online because of
  fuller functionality
• throughout, working relationship had been good
  started this discussion in summer 2000, proposal
  to Mellon followed

6
Teams

• Yale
• Scott Bennett (PI), Paul Conway, David Gewirtz,
  Fred Martz, Ann Okerson (Co-PI), Kimberly Parker,
  Richard Szary
• Elsevier Science
• Geoffrey Adams, Emeka Akaezuwa, Haroon Chohan,
  Karen Hunter, Paul Mostert

7
Some of the starting assumptions

• digital archive gt100 years
• archiving content, not format or functionality
• archive responsible for migration
• archive not competing with publisher
• archive shouldnt be totally dark

• archive does not mirror pub. site
• archive does not create content not in original
  e-edition
• highly desirable for publisher to provide needed
  metadata
• standards are key
• archive not a hot backup for disasters

8
Work plan -- issues

• what does it mean to be an archive
• journal business life cycles
• economic issues
• contractual relationship between publisher
  archive
• metadata needs
• archival uses independent of day-to-day uses
• technical infrastructure

9
Work plan -- deliverables

• metadata elements
• model license
• a prototype

10
Sizing the problem

• Elsevier Science published 1,100 journals when
  project started
• acquired Harcourt (Academic Press, Saunders,
  Mosby, Churchill Livingstone) mid-2001, raising
  number of journals to 1,500
• backfile digitization project (v.1, no.1)
• guestimate 6.5-7 TB

11
What is an archive?

• publishers production archive
• normal customers with locally-held files
  (ScienceDirect On Site)
• self-designated archives -- national bias
• official archives -- formal relationship for
  perpetual care

12
Journal business life cycle

• Scott Bennetts starting premise
• ...information half-life, which is the point at
  which the commercial value of e-journal content
  to the publisher has declined to the point where
  the publisher hands off preservation and access
  responsibilities to an archiving agent

13
Trigger events

• failed to find such a half-life
• felt too early in the e-publishing process
• for example, we want to recover costs of
  retrodigitization
• extensive, fascinating discussion of trigger
  events
• points when archive can go bright to the public
• could not identify trigger events
• except if publisher goes out of business and no
  successor taking ownership of assets

14
Economic considerations

• inability to foresee a half-life and unlikely
  trigger event leads to question Does it make
  sense to establish an archive now, given costs
  involved?
• conclusion yes
• libraries need archives to go e-only and it will
  only be riskier, more difficult and more
  expensive later
• how to fund part of next phase

15
Contractual relationship

• issues different from normal license, including
• perpetual nature of an archive
• service level agreement
• trigger events -- public access
• financial terms
• format for submission
• comprehensiveness of archive (e.g., withdrawn
  material)

16
Metadata needs

• adoption of both OAIS and OAI as standards to be
  used
• detailed evaluation of the metadata work done by
  others, most notably the British Library
• also close analysis of metadata currently
  captured by Elsevier
• mapping of needs and gaps
• mapping to Dublin Core for prototype
  
  
  
  
  
  

17
Archival uses

• notion that there are uses made of an archive
  that are quite different from normal day-to-day
  researchers use of journals
• if could be identified, could open the archive
  for such uses immediately
• started discussion with history of science faculty

18
Technical infrastructure

• needed to understand the publishers e-workflow
  and what the publisher can delivery (now or with
  modification)
• needed to understand what others are doing
• series of site visits (to ES Amsterdam, Royal
  Library, British Library, Chase, etc.)
• want to learn from others and avoid re-inventing
  the wheel
• importance of standards (OAIS and OAI) clear

19
Deliverables

• metadata elements -- done!
• Description of Metadata Elements for the Yale
  Electronic Archive
• model license -- work in progress
• early version that needs more wordsmithing and
  legal polishing
• prototype -- done!
• working prototype that was registered with ARC
  for OAI harvesting

20
Next steps

• we have learned a lot, including that we can work
  together and must work together if this is going
  to happen
• now we want to build the 6.5-7 TB real thing
• part of that process will be further identifying
  ways to collaborate with other archives
  internationally

21
A word of thanks...

• to Ann Okerson and the Yale team for letting me
  borrow from the draft project report to prepare
  this presentation