Why identifiers

1
Why identifiers?

• To access resources
• To cite resources
• To unambiguously identify a resource
• To register it as intellectual property
• To record changes in ownership

2
How we identify objects now

• http//www.library.cornell.edu/dlit/systems.htm
• ftp//ftp.is.co.za/rfc/rfc1808.txt
• gopher//spinaltap.micro.umn.edu/00/Weather/New20
  York/Ithaca

3
URLs

• http//www.library.cornell.edu/dlit/systems.htm
  
• Service instructions for accessing an object
• Oriented toward access/location vs.
  identification
• Essentially, an address
• Most likely ephemeral

service, protocol
machine/host name
local file location
4
Good Identifiers

• Globally unique
• Persistent permanent
• Even when the identified resource ceases to exist
  or is no longer available
• Independent of location

5
Good Identifier Schemes

• Scalable
• Extensible
• Support for legacy schemes
• ISBNs, ISSNs, etc.

6
Uniform Resource Identifiers
URL
URN
7
URNs

• urnltNamespace IdentifiergtltNamespace Specific
  String
• Identifies an object
• Nothing about access mechanism, little about
  location
• Requires resolution services and registries
• Indirection

8
URNs

• urnissn0003-486X
• urndwrmyOrganizationstheirMembers

9
Naming schemes

• Handle System
• cnri.dlib/june98-powell
• DOI
• 10.1045/october2002-hitchcock
• ARK
• ark//some.host/1234/98765432
• ark1234/98765432

10
As (potential) URNs

• urnhdlcnri.dlib/june98-powell
• urndoi10.1045/october2002-hitchcock
• urnark//some.host 1234/98765432

11
PURLs

• A persistent URL
• Using indirection
• Temporary solution?
• Still identify a service, not an object
• Not an actual name
• Dependent on that service (HTTP)

12
In the interim

• Likely depend on proxy services for resolution to
  Cornells doorstep
• urncul...
• urnhdlcul.rmc...
• But inside our doors?
• How do we manage identifiers within Cornell
  and/or Cornell University Library?

13
Identifiers and semantics

• Semantic identifiers
• Perhaps intelligent to people
• cnri.dlib/june98-powell
• Non-semantic identifiers
• Not meaningful to people
• 10.1045/abc123456