Yale University Library

Digital Preservation Committee (DPC)

Meeting Notes prepared by Bobbie Pilette

Date: October 17, 2005, 10:30a.m.- 11:30a.m.

Present: Kevin Glick, Audrey Novak, E.C. Schroeder, David Gewirtz, Bobbie Pilette, Gretchen Gano

Absent: Rebekah Irwin

Updates:

Discussed the update Bobbie is to give at IAC later today. Agreed the policy was finished, except for the best practices section which is a work in progress. Will also report as to progress on the best practices.

Audrey presented a draft blurb reporting on the committee’s work that will appear on the DLF newsletter.

A proposal to review our charge at a future meeting to make sure we are on track and keeping the large picture in mind was made and agreed to.

Kevin clarified that the Fedora/University Records document should be understandable without having to go back to the original documents. He said to just send comments regarding the documents to him via email.

A Fedora Working Group for Preservation Services has been formed as part of the Mellon funded Fedora project. Kevin has been asked to serve on this group along with Eliot Wilczek from Tufts University Digital Collections and Archives, Grace Agnew at Rutgers University Libraries, Ron Jantz, Social Sciences Data Librarian at Rutgers University Libraries; Sandy Payette, co-director of the Fedora Project, Computing and Information Science, Cornell University; Bill Parod, Academic Technologies, Northwestern University; Karen Miller, Digital Projects Catalog. A draft of the charge is attached.

David reported on the European Fedora meeting saying that the last round of funding will be 2007 and they are looking for a sustainable model for Fedora. Also that with the remaining funding they are looking for a final product(s) such as Archival Workflow and Preservation Services, hence the working groups.

Best Practices & Use Cases:

Discussion then turned to the main topic of best practices. Having looked at various use cases and raised a number of issues associated with each, the committee then needed to decide how to find the common and unique issues among the use cases.

One approach based on OAIS was discussed.

Pre-ingest

• Requirements:
• Managed environment
• What a selector would consider
• Trusted vendor
• Who is the audience
• Selectors
• Faculty

· Source of files

When vendor of leased/purchased files agrees to preserve or maintain long-term access falls under 3^rd party scenario.

When vendor of lease/purchased files guarantees preservation by providing some form of the files for us to maintain that falls under managed environment.

Example Best practice: Vendor should provide perpetual access and be able to confirm access. If YUL gets the files, we should receive them in the following formats [state formats] and on these media [state media].

• Creation
• What degree of influence on formats
• In case of records may skew the creation
• Lease/purchase; very limited control
• Ingest
• Metadata
• Formats accepted
• Normalization [change formats to that which can be preserved]
• Transmission [the process of receiving]
• Intellectual property rights
• Authentication & Authorization
• Rendering/Access
• Administration
• Migration
• Back-up/recovery

David then suggested that we look back to the ECAR Life Cycle (see the ECAR powerpoint).

4. Preservation Services Architecture

· A description of the suite of preservations services

· Service operation – how it works

· How to interoperate with the service – APIs, utilities,

· Based on services in Fedora 2.x

5. Further Discussion. We need to decide in the WG how much we can undertake and how the work maps to our schedule. The following are items that can be included in our WG activities:

• A plan for federated preservation within the Fedora community. Preservation is a difficult task and needs the economies of scale that a shared platform should leverage.
• RLG and OCLC are working on a trusted repository certification process. The WG should collaborate with RLG/OCLC as reactors and testers of their certification process.
• A definition of how Fedora qualifies as an OAIS platform, demonstrating conformance with the model proposed by OAIS. This will help also with marketing Fedora to savvy archives.
• Recommended best practices for preservation and provenance metadata to support collaborative preservation.

6. Related work (groups, technologies, standards the WG should review)

· NDIIP, NLA, CEDARS, etc

· Content models – how integrated with an object (e.g. digiprov metadata)

· Persistent Ids (CNRI Handle, ARKs, Fedora PIDS as Handles, etc)

· Object and format validation (JHOVE). What is the digital original? Does this object have the standard format for a jpeg, wav (etc) file?

· PREMIS (preservation metadata – data dictionary from OCLC/RLG)

· Migration events

· Audit trails and versioning

· Intelligent agents and semi-automated preservation processes

· Canonical forms for archival masters

rcj – 8/23/05