Identification of Electronic Resources: identifiers and resolution services

1
Identification of Electronic Resources
identifiers and resolution services

• Juha Hakala
• Helsinki University Library
• 2003-01-29

2
Background

• Rapid growth of electronic publishing has
  revealed fundamental problems in our existing
  identifier systems
• Rules of implementation and/or syntax must be
  changed, new systems developed, and resolution
  systems utilising the identifiers need to be
  built
• Every resource must have an identifier!

3
Scope of the work

• Identifiers for authors (ISADN)
• Identifiers for works (ISTC, ISAN, )
• Identifiers for manifestations (ISBN,)
• Identifiers for component parts (SICI, BICI)
• Resolution services (DOI, URN)
• Always incorporate an identifier

4
ISSN

• Capacity of the system is sufficient
• 1.5 million out of 10 million IDs in use
• Many open issues
• How to implement ISBD(CR) in practice?
• Staffing nat. ISSN centres need more cataloguers
  
• Web journals are not stable 856 versus OpenURL
• Data utilisation problem the global ISSN
  database system requires modernisation
• Faster updates from nat. centres, Z39.50 access
  on-line

5
ISBN

• Capacity is an issue
• We will run out of ISBNs within a decade
• Rule problem publishers want to use ISBN also to
  component parts (not BICI) in order to simplify
  their systems
• Utilisation problem there is no global ISBN
  database (and may never be)

6
The New ISBN current plans

• ISO/CD 2108, dated 2003-01-17
• For ISO/TC 46 SC 9 /WG 4 meeting 30.-31.1.2003
• Bookland EAN prefix (978) will be added
  otherwise the structure remains the same
• ISBN-13 978-90-70002-34-3
• -gt every old number can be re-used
• Check digit calculated using Modulus 10 algorithm
  (Mod 11 in ISBN-10 0-9 X)

7
The New ISBN some cancelled ideas

• Make ISBN ISSN-like dumb number
• Enhanced capacity, but reduced usefulness
• Create the global ISBN database
• Technically, organisationally and politically
  controversial idea
• Instead, national centres will make their data
  available
• Extend ISBN to 16/25/32 digits
• Would have broken the EAN system

8
National Bibliography Number

• Traditionally the identifier for records in the
  national bibliography, if the publication did not
  have an identifier
• New scope identifier for (electronic) resources
  to which no other identifier applies
• Implemented as URNs in order to guarantee global
  uniqueness

9
National Bibliography Number examples

• All implementations based on RFC 3188
• Finnish Web Archive (11.7 million files)
• Machine generated ID based on MD-5
• urnnbnfifaltMD-5gt
• Koninklijke Bibliotheeks E-depot
• urnnbnnlkbeDepot-ltUNIX timegt

10
Uniform Resource Name

• Internet standard approved in fall 2002
• Both an identifier and resolution service
  (mechanism for linking identifier and resource in
  the Internet)
• Designed to be protocol implement the current
  version is built on top of DNS, but
  infrastructure can be changed

11
URN syntax

• Specified in RFC 2141 (1997)
• Three sections, separated by commas
• String urn
• Namespace identifier (NID)
• Namespace specific string
• urnnidnss

12
URN services

• Supply the actual document
• Deliver metadata related to the document
• Pass the list of URLs from which the resource can
  be found

13
URN namespace registration

• Each namespace must be registered as specified in
  RFC 2611
• Registration must contain the proposed NID (such
  as nbn) and an outline of how the global URN
  resolver discovery service will function within
  the namespace
• Registrations are approved as informational or
  normative RFCs

14
Administration of the NBN namespace

• Each national library is allowed to do whatever
  it wants with its own part of the NBN namespace
  (as long as the identifiers remain unique and
  persistent)
• National Library of Finland has assigned some
  organisations their own sections
• Library of Congress could do the same

15
URN resolution process

• Based on DNS there is a resource record which
  describes the location of the service which can
  resolve a URNs with given NID/NSS combination
• Complexity of the resolution process varies
• ISSN single database is enough
• ISBN databases of national centres will do
• SICI huge amount of aI-services needed

16
URN some benefits

• No assignment cost
• Trivial to create from existing identifiers
• Add a fixed prefix, such as urnnbnuscornell
• Internet standard support will gradually be
  included into the basic tools we use
• Present architecture for resolver discovery
  service is robust and scalable

17
URN some problems

• Someone must pay for the implementation of
  resolution services in e.g. ILSs
• Commercial publishers prefer DOI
• Only a handful of systems have registered
  namespaces
• E.g. ISSN, ISBN, NBN
• Dumb identifier with multiple resolution services
  does not fit into the system well (although there
  may be a cascade of resolvers)

18
URN versus DOI

• DOI system is a technology, not a standard
• Standardisation of DOI syntax is not enough
  services and the practical implementation of the
  resolution mechanism must also be fixed
• Handle system has failed to attract IETF
• In DOI system, anything can be used as an
  identifier (suffix)
• DOI requires registration of registrants
  (publishers)

19
URN versus DOI (2)

• DOI syntax is mandated by the Handle system
• Actual DOI implementations are dependent on HTTP
  protocol (which will not last forever)
• Handle system may become (but is not yet truly)
  distributed
• DOI has been widely implemented and works OK
• Only one DOI service retrieval of the resource

20
URN and ENCompass

• Endeavor has no immediate plans to develop URN
  resolution service
• I.e. mechanism for receiving URN resolution
  requests arriving via DNS
• URNs can however be stored in metadata or into
  documents themselves, and indexed
• This means that implementing URN RS should not be
  complicated (it was designed to be easy to
  develop)