ISO Archiving Standards - Tenth US Workshop- Minutes

The Tenth US Workshop on Data Archive Standards was held at the National Archives and Records Administration's (NARA) Archives II Complex in College Park, MD on January 28-30, 1998. The list of attendees is available at: http://ssdoo.gsfc.nasa.gov/nost/isoas/us10/participants.html.

Action Items
Materials Distributed
Future Meetings
Discussion Items

• Agenda
• Liaison Activities
  • CODATA Conference
  • National Association of Government Archivists and Record Administrators (NAGARA)
  • JISC and British Library
• Action Items Status Report
• Detailed Discussion of Reference Model
  • Section 4.1 -
  • Section 4.2 -
  • Section 4.3 -
  • Section 5 - Migration Perspectives
  • Annex C - Federated Archives
• Writing Assignments
• Workshop Planning
• Other Business
  • OAIS Status
  • Joe King Comments
  • SOMO Report

Action Items

Materials Distributed

1. Meeting agenda
2. Planning for US/International Data Archiving Future Standards Workshop
3. NAGARA Conference Announcement
4. Request from King's College London
5. Open Action Item List
6. OAIS Sect 4.1 - Functional Model
7. OAIS Sect 4.2 - Information Model
8. OAIS Sect 4.3 - High Level Data Flows
9. Analysis of Information Migration
10. Preservation Description Information example matrix
11. OAIS Annex B Federation of Archives
12. Object Modeling - Practical Tips
13. Joe King's comments re. OAIS
14. International Council for Scientific and Technical Information

Future Meetings

Discussion Items

Agenda

Liaision Activities

CODATA Conference

• Committee held large conference on scientific and technical issues
• One splinter session was on archiving
• Don Waters and Cliff Lynch gave a presentation of paper on Preservation of Digital Information paper
• Don Waters and Cliff Lynch expressed interest in getting involved with upcoming workshop

National Association of Government Archivists and Record Administrators (NAGARA)

• NAGARA is a group of people at all levels of government
• NAGARA is holding a meeting in July in Philadelphia with a session on Standards (Item 3)
• Don Sawyer is to provide a presentation on OAIS to them
• This is another opportunity to interact with RLG who will also be giving a presentation in the same session

JISC and British Library: (Item 4)

• A Mr. Neil Beagrie is conducting a study on Guidelines for Digital Preservation for ISC/BL
• He is interested in OAIS and is also looking for a bibliography of Archive-related papers

Action Items Status Report

Detailed Discussion of Reference Model

Section 4.1 - (Item 6)

Major Changes (as relayed to Europeans)

• Nomenclature: Access and Dissemination Issue
• Order execution flow; End-to-end
• Data Mining; level of discussion

Meeting Discussions

• Mike Martin reviewed the changes he had made in response to comments from the IWS in November:
• Admin is involved as to what is being done under Review
• Need report card from Ingest to Administration; a science rewiew of what was agreed to versus what was done
• In 4.1.5, add Data Archiving Engineering entity to include scientific review
• Move peer review from Ingest to Administration and rename it

Archival Storage

• Need to address requests for "Bit-error-rate" performance
• "Levels of Service" could be requested by Ingest to Archival Storage
• Expand example where Reed-Solomon is used for correction and not just detection of errors
• *Mike Martin to develop a organizational architecture of an OAIS roles and responsibilities to see if this might be included in the Reference Model
• Need to talk more about preparing and providing Finding Aids
• Prepare is proposed to be in Data Administration while Provide is in Access
• *Mike to modify this section
• Access does not mean getting the data; this is under Dissemination
• Need to agree on functionality of Data Management
• *Find a new term for Access
• Data Management deals with keeping records up to date
• What selection criteria has to be stored in Data Management
• Lou Reich: Data Management is concerned with triggering actions
• Data Management is the home for orders
• Access deals with metadata
• Dissemination deals with DIPs
• Administration deals with everything else
• Access and Dissemination are not specifying interface(s). Interfaces are an implementation issue and they could be implemented as a single interface.
• Does order come from Access or Data Management in going to Dissemination?
• List of documents is in Access
• Say in begining of 4.1 - this is not an implementation
• * Action for Mike - he plans to walk order execution through the way PDS does things to gain insight. He will generate a proposal that works for both standing and ad-hoc orders.

Section 4.2 - (Item 7)

Major Changes (as relayed to Europeans)

• Take Representation out of IO Taxonomy and split into IO-generic and Content Specific
• What is appropriate view of Representation Lineage/Net/Container
• OMT of Access Aids
• Remove Accesss Methods; included specific type of Access Aids

Meeting Discussions

• Time was allowed for participants to read the section The afternoon was devoted to discussing the revised text. Statements made include
• The Information Object is a big point of the document
• Representation Information is confusing and should revisited
• Paul Grunberger: only describe Information Object
• Move discussion re. Representation Information forward in the document
• Representation Information is key to successful long term archiving
• Put generic text re. Representation in Section 4.2.1.1 and specific text re. Representation as an Archive Information object later in the document
• Retitle section 4.2.1.2.1 into a Taxonomy of Archival Information Objects and Content Information parts of 4.2.1.2.2 become part of Content Information.
• Change title of top box in Fig 4-6 "Archival Information Object"
• Delete first half of paragraph under Fig 4-7
• Don't talk about software under Representation Information
• Will have to talk about Software somewhere
• There are two concepts: PDI and PDI container
• Proposed to do this with Representation
• But there are no classes in Representation like there are in PDI
• Move generic part of Representation Information forward to Information Object
• Rewrite section

Section 4.2.1.2.3 - Preservation Description Information

• * Lou Reich will try rewriting this first paragraph and add a reference
• Delete "In addition to the content object"
• OAIS is to focus on the digital, but to also work with the physical
• Context, as stated in Preserving Digital Information paper, included what we now have in Packaging Information
• Extended discussion re. Reference Information
• Include in Provenance Information, history/description of instrument and processing history, who has had control of the information
• Lots of overlap in the contents of these entities
• Content Information can include algorithms to be applied, calibration data, tracking data
1. You created this information and you now have the provenance. You created it
2. When you get lots of data, when does something move to context and out of provenance
• Lou Reich feels we should discuss Provenance information at a higher level.
• * Mike Martin will work up a matrix of PDI examples for various types of archival information.

4.2.1.2.4 Packaging Information

• John Garrett: feels some of the description in second paragraph is wrong; if it has to be preserved, it is Content information
• Function of Packanging Information is to delimit the content like a test-tube or CD-ROM, at a minimum
• Packaging information by definition is outside the Content Information
• If we say it must be preserved, it is Content content
• * Lou Reich to rewrite second paragraph

4.2.1.2.5 Descriptor Information

• Associated with the locating, analyzing, and ordering information
• Keep the four categories at the same level
• Haven't introduced AIP yet so make it IP
• Remove Access Methods as a separate category - consider it to be under Access Aids since it was here to provide higher performance access to Archival Storage contents.

Section 4.2.2 - Logical Model of Information

• Need to develop a philosophy regarding acronyms since some sentences have a great number of them.
• It was suggested that acronyms covering only two words should be avoided
• Last paragraph: strike cardinality and optionality

Section 4.2.2.3 AIP

• Fig 4-11: PDI Container is confusing; "container" is an aggregate of things
• *Look for better name than "Container"
• Fig 4-11: Change AIU to AIP
• Under Fig 4-12 re. "virtual" can't assume that all the Packaging Information is on the same media with the Content Information, but this doesn't make it 'virtual'
• Need diagram or example
• Expand packaging description to address split information
• Need word to describe content which is split across multiple volumes (and so is its packaging description)
• Term virtual doesn't work
• * Lou Reich to rework
• Content representation information stays with Content while Packaging Representation discussion is moved to migration

4.2.3.1 Specialization of AIP

• In Fig 4-17, delete Access Methods and describe need in text
• Storage mentioned only in Storage and Management
• Under Fig 4-19 use different term for virtual - like temporary

Section 4.3 - High Level Data Flows and Transformation (Item 9)

• Significantly expand Section 4.3 to address the PDI implementation issues - like collections of collections
• Some comments had been received at November International Workshop
• Lou Reich plans a total rewrite of this section to better describe how the various structures are formed
• Mike Martin had some minor comments which he provided to Lou Reich

Section 5 - Migration Perspectives

Analysis of Information Migration

• Don Sawyer had prepared a concept paper to incorporate concepts of content and packaging information (Item 9)
• He felt his views on concepts were in agreement with present positions such as Randy Davis's treatment
• This is not intended to be a replacement for anything
• Mike Martin: Suggested a media format picture as a way to introduce and explain the volume/file structures
• Mike Martin: There were a lot of good, important issues addressed in this paper
• Lou Reich: Are there any technical shortcomings in this document? what needs to be put in Section 5? He had trouble with flow.
• Mike Martin felt all items were applicable to real life situations
• Under Device Usage; change to Performance Drivers
• Under Migration, what is done in Storage and what has to be outside of Storage?
• Where does Migration end? If strict transmutation is not possible, the new product will have to go through re-ingest process
• BA feels this is a major expansion of the normal term Migration
• Reasons to migrate include: when HW, SW or media is no longer supported or you are out of touch with the user community
• Does name have to be changed
• *Put boundary on discussion
• Would like to get to where Migration can be done in Storage and not go external for renaming, etc.
• New management systems can give trouble
• How do you migrate error recovery techniques
• other edits provided to Don Sawyer

Annex C - Federated Archives

Federated Archives

• Lou Reich and Paul Grunberger had developed a paper on this subject (Item 11)
• A number of architectural arrangements were presented
• Paul Grunberger indicated that these were intended to stimulate thought; his approach was to leave OAIS intact
• First design simply represents a mutual agreement between two OAIS; doesn't rise to level of a Federation
• Mike Martin liked the first paragraph and felt it should be the basis of further development in the balance of the paper. Don Sawyer agreed.
• Gives examples of what one can do as a function of your requirements
• Change title to Interaction of Archives
• Paul Grunberger doubts that Federation is the answer
• Lou Reich feels federation is a combining of local and remote archives
• PDS has a Global Ingest operation: Send request to a data engineer who would determine where the data resides and you would then interface with that archive
• PDS data is distributed/stored by type
• Easy to conceptualize from Consumer view; hard to conceptualize from Producer point of view
• APL project has a type of Common catalog system: receives notices of data but does not store data
• On figures, show areas of overlap
• Lou Reich explained his thinking re. his developing the table in the document
• Generate one global description and send to a central source is one type

Writing Assignments

• Lou Reich is in favor of scratching Illustrative Scenarios unless we can get Randy Davis or someone else to update it
• Federation of Archives
  • *draft (include comments, define classes) - Lou Reich
  • *edit (give examples, include pictures) - Paul Grunberger

Workshop Planning (Item 02)

• Initial plans regarding a large workshop (100-150 people) to discuss OAIS Future Archive Standards
• Tentative dates of June 22-26, 1998 was set.
• Planning committee will be formed and will provide more details shortly.

Other Business

OAIS Status

• Hope to get final edits at JSC
• Editors of Sections are encouraged to attend
• Not sure if there will be RIDs
• Hope for Draft DIS by end of JSC Workshop
• Six month cycle for RIDs
• Agency RIDs will be reviewed in July

Joe King Comments (Item 13)

• These comments regarding the OAIS Reference Model were discussed
• Attendees were asked to read it

SOMO Report

• Mike Martin stated that the final report is about to come out
• Don Sawyer is on the SIS working group
• SIS has been interviewing projects to identify where cost drivers are in an effort to save money

The Workshop was adjourned at 1430 hours, Friday, January 30, 1998.

Wider Views