Persistent Resolvable Identifiers in the Online Environment

1
Persistent Resolvable Identifiers in the Online
Environment

• IDEA 2006 Open e-Learning Forum
• Matthew Walker
• Director, IT Collection Infrastructure
• National Library of Australia

2
Outline

• What is a Persistent Identifier (PI)?
• Why do we need PIs?
• Real world examples
• NLAs role in relation to PIs
• Meaningful vs. Arbitrary PIs
• NLAs PI scheme
• Lessons learned
• Futures
• References

3
What is a Persistent Identifier (PI)?

• Definitions
• Identifier
• Linguistics. A linguistic element that has the
  function of identifying. (Oxford English
  Dictionary Online, http//dictionary.oed.com)
• the identity of things through time is due to
  the presence in them of an identifying component
  or substance, an identifier (A. Quinton, Nature
  of Things)

4
What is a Persistent Identifier (PI)?

• Definitions
• Persistent
• Continuous, continuing to exist enduring,
  lasting chronic. (Oxford English Dictionary
  Online, http//dictionary.oed.com)
• Persistent Identifier
• An element that identifies a resource on a
  continuing, enduring basis.
• Implies that there is a service somewhere that
  allows you to resolve that persistent identifier.

5
What is a Persistent Identifier (PI)?

• Definitions
• Persistent Resolvable Identifier
• An identifier that can be used to refer to an
  item or resource on an ongoing basis, to find
  information about it, and to locate it.

6
Resolvable Persistent Identifier

• Resolver Service at the NLA

The Web
Web Browser
1. PI request
4. Resource contents
Resolverhttp//nla.gov.au
NLA WebInfrastructure
2. URI or file request
3. Resource contents
Resource
7
Why do we need PIs?

• To ensure that access to a cited object is
  ongoing.
• Example policy document on a website moves to
  another folder. Location metadata should be
  updated so that the PI still refers to the
  document in its new location.

8
Why do we need PIs?

• Persistent association with an item.
• Digital object
• Metadata
• Description
• Information on how to get the item
• Other issues
• Rights restrictions
• Item no longer exists

9
Real world examples

• Social Security Numbers Tax File Numbers
• Yes, assuming that your number doesnt get
  re-used after you die.
• ISBN
• Yes, if it is representing a particular
  publication of the book (a manifestation), but
  you cant use it to identify an individual copy
  of a book.

10
Real world examples

• Street address
• No these can change.
• Longitude/latitude combination
• Depends what object it is supposed to represent
  on an ongoing basis. Yes, if it represents a
  particular point on the globe. No, if it
  pinpoints your house, which may not exist in 100
  years.

11
Real world examples

• Web domain name
• No what happens when you forget to renew your
  domain name subscription?
• Date/time
• Maybe, if it is used to refer to a particular
  point in time.

12
NLAs role in relation to PIs

• Need for own collection
• Participate in development of standards for
  recording PIs
• Advise other institutions on PIs and digital
  collection management
• Emerging role of providing resolution services
  for other peoples PIs
• Handle DOI
• info URI - registration options
• Every individual items PI or
• Rely on central repository to resolve to the
  remote item.
• Each agency/institution registers.
• Assumes agency/institution has infrastructure to
  resolve to its own items.

13
Meaningful vs. Arbitrary PIs- Advantages

• Meaningful PIs
• Content managers users love the fact that the
  meaning of the PI is intrinsic.(heavily weighted)

• Arbitrary PIs
• Allocation of PI is automatic saves content
  managers time.
• Dont have to maintain PI scheme saves content
  managers time.
• Dont have to implement changes to the PI scheme
  in underlying technology saves IT time.

14
Meaningful vs. Arbitrary PIs- Disadvantages

• Meaningful PIs
• Assignment of the PI has to be done manually by
  the content manager/s takes time introduces
  possibility for error.
• Have to create/maintain policies for the
  structure of the PIs.
• What happens if a content manager assigns the
  wrong PI?

• Arbitrary PIs
• Content managers and users cant interpret what
  the PI represents without a resolution
  service.(heavily weighted)

15
NLAs PI scheme

• Meaningful PIs
• http//nla.gov.au/nla.aus-an3281107
• http//nla.gov.au/nla.pic-an21621679
• http//nla.gov.au/nla.ms-ms51
• http//nla.gov.au/nla.arc-13467

16
NLAs PI scheme

• PI extensions to identify levels in the hierarchy
  of bibliographic metadata
• http//nla.gov.au/nla.aus-an3281107-0-s3
• http//nla.gov.au/nla.pic-an23323328
• http//nla.gov.au/nla.ms-ms51-1
• http//nla.gov.au/nla.ms-ms51-1-2
• http//nla.gov.au/nla.arc-13467-20000911-http//ww
  w.acfonline.org.au/campaigns/international/trade/t
  akeaction.htm

17
NLAs PI scheme

• Further extensions to identify different
  representations/copies of an item. For example
• original copy (-o)
• master copy (-m)
• http//nla.gov.au/nla.aus-an3281107-0-s3-v
• http//nla.gov.au/nla.pic-an23323328-t
• http//nla.gov.au/nla.map-lfsp1958-e
• http//nla.gov.au/nla.ms-ms9-fa

18
Lessons Learned

• Dos
• Allocate PIs for description of the resource and
  its sub-units.
• Register your PI scheme with InfoURI or use a
  URN.
• Registration with InfoURI is simpler less
  costly than setting up with URN.

• Donts
• If using meaningful PIs, dont develop systems
  that are dependent on the content of the PI for
  ingest and dissemination workflows.
• need a standard SIP/DIP (METS is our choice)
• Dont link meaningful PIs directly to file
  system. Need a database to deal with issues such
  as multiple PIs for the same object, best copy,
  access control.

19
Lessons Learned

• Need a syntax for requesting copies of an object
  with different roles (e.g. thumbnail, contextual
  display, structural map).
• We encoded this in the PI but a protocol would be
  better
• Example
• see this catalogue record with stacks of URIs
  http//nla.gov.au/anbd.bib-an40416514

20
Futures

• PI solution for whole domain web archive.
• PIs for metadata.
• PIs for entities authority records, people,
  places, topics.
• PIs for works, expressions, manifestations, items
  as defined by FRBR standard.
• Handle or DOI?

21
References

• A practical approach to ensuring the persistence
  of digital collections at the National Library of
  Australia, Tony Boston Ninh Nguyenhttp//www.nl
  a.gov.au/nla/staffpaper/2002/boston2.html
• Persistence is a Virtue, Diana Dackhttp//www.nla
  .gov.au/nla/staffpaper/2001/dack.html
• What is FRBR?, Barbara Tilletthttp//www.loc.gov/
  cds/downloads/FRBR.PDF
• info URIhttp//info-uri.info/registry/docs/misc
  /faq.html

22
References

• METShttp//www.loc.gov/standards/mets/METSOvervie
  w.v2.html
• RDAhttp//www.collectionscanada.ca/jsc/rdaprospec
  tus.html
• PURLhttp//purl.org/
• DOIhttp//www.doi.org/
• OAIS (SIP/AIP/DIP)http//public.ccsds.org/publica
  tions/archive/650x0b1.pdf