European Endeavor Users Group MeetingHelsinki, Sept 1-6 2003Esa-Pekka Keskitalo, System AnalystHelsinki University Library
OpenURL 1.0
Outline
• Background• The Draft• OpenURL in Libraries
Background
• Appropriate Copy Problem• Broken Link Problem
• Appropriate Copy – Linkking should be user sensitive
• Broken Links should be fixed centrally (need to be fixed somewhere)
OpenURL is
”a methodology for describing • resources that that are referenced in
a networked environment • context of the reference”• Neutrally in respect to applications
and user communities
From 0.1 to 1.0
• Much extended• ”Registry”• From ”scholarly” to a more general
approach
Standard in two parts
1. ContextObjecta) A description of a resource
that is referenced (target)b) A description of the context
of the reference (source)c) A definition of a set of Namespaces
2. Registrya) Detailed lists of options – will expand
A Context Object is a Set of Entities
• Context Object is an information construct with six Entities.
• Referent is the core entity sine qua non
• Referent = what is being referenced= what we e.g. want to read in fulltext
ContextObject
Entities
Resolver
Referent
Referrer
Service type
RequesterReferring Entity
ContextObject
Entities
1. Referent: a referenced resource, e.g. an article, a person etc.
2. Referring entity: where the reference is, e.g. an article
3. Requester: who is using the system4. Service Type: what is wanted, e.g.
”fulltext”5. Resolver: the target resolving service6. Referrer: The system that generates the
ContextObject.
ContextObjectEntities
Descriptors
Detailed information about an Entity.1. Identifiers2. Metadata-by-Value3. Metadata-by-Reference4. Private Data
ContextObjectEntitiesDescriptors 1) Identifiers
Defines1. Namespace that is being used2. Identifier, unique in that Namespace
Namespace: DOIIdentifier: 10.1126/science.275.5304.1320
Namespace: MailtoIdentifier: [email protected]
Namespace: ISSNIdentifier: 1234-5678
ContextObjectEntitiesDescriptors 2) By-Value Metadata
Defines• the metadata format used• metadata expressed in that format
Metadata: MARC21 = MARCXML Schemarft_val_fmt = ori:fmt:xml:xsd:MARC21
<?xml version=”1.0” encoding=”UTF-8”?>---<varfield id="245" i1="1" i2="0"> <subfield label="a">Policing the new world disorder /</subfield> <subfield label="c">by Robert Oakley and Michael Dziedzic. </subfield> </varfield>
ContextObjectEntitiesDescriptors 3) By-Reference-Metadata
Defines• the Metadata Format used• the network location of the metadata
Metadata: MARC21 = MARCXML Schema
rft_ref_fmt = http://www.loc.gov/standards/marcxml/schema/MARC21slim.xsd
rft_ref = http://www.loc.gov/standards/marcxml/Sandburg/sandburg.xml
ContextObjectEntitiesDescriptors 4) Private Data
• Not defined in the Standard• Specific to the Referrer (the generator
of the ContextObject)• Must be understood by the Resolver
ContextObjectEntities Constraints
• Exactly 1 Referent in one ContextObject• No more than 1 Referring Entity,
Requester, Referrer• Zero or more Service Types and
Resolvers.
• All Descriptors may describe all Entities• Multiple Descriptors possible
REGISTRY
• description of network-referenced resources
• description of the context of the references
• support of transportation of the descriptions over the network
Registry
Components
1. Namespaces– 3 Naming Environments
2. Character Encodings3. Physical Representations4. Constraint Languages5. Context Object Formats6. Metadata Formats7. Transports8. Community Profiles
Registry Components
1) Naming Environments
a) URI – Uniform Resource Identifierb) ORI – Open Resource Identifierc) XRI – Private Resource Identifier
Registry Components Naming Environments
a) Uniform Resource Identifier
– Internet Assigned Naming Authority (IANA)
– http, ftp, mailto– ldap– urn, urn:ISBN, urn:ISSN, urn:NBN
Registry Components Naming Environments
b) Open Resource Identifier
• bibcode, doi, isbn, issn, oai, sici• Id for Referrers:
– domain name + string
• Id for Service Types: – abstract, citation, fulltext, holdings, ILL
Registry Components Naming Environments
c) Private Resource Identifier
• Not defined any further
Registry Components
2) Encodings
• Character Encodings– UTF-8, ISO-8859-1, Big5– IANA List
Registry Components 3) Physical Representations
• Physical Representations– XML and KEV (Key/Encoded Value)– KEV specified in the Standard
Registry Components 4) Constraint Languages
• Constraint Languages– Z39.88-2003 MTX– W3C XML Schema
Registry Components
5) Context Object Formats
• Format– a concrete method to express a class of
information constructs– Physical representation + constraint language+
constraints
• A definite choice set within the Format triple– KEV + MTX + Matrix of Section 8.1. of Part 2– XML + XSD + XML Schema of Section 8.2 of Part
2
Registry Components
6) Metadata Formats
• MARC 21• OAI_dc• XML and KEV metadata formats for
books, dissertations, journals, patents specially defined
Registry Components
7) Transports
• Method of tarnsporting over the network• OpenURL 0.1 for backwards compatibility• 6 combinations: OpenURL http or https
– inline– by reference GET / POST– by value GET / POST
Registry Components
8) Community Profiles
• A set of choices from Registry entries for an application domain– 1 Context Object Format– Metadata Formats– URI and ORI Namespaces– Transports– Identifiers– Expressed in XML
Draft Standard 1.0
• Released in March• Time for testing and comments ends
in November, 2003• Big changes to version 0.1
OpenURL resolving server
• Resolving server needs constat attention
• Centralised update services emerging, but -
• Central knowledge needs adjusting• Local knowledge needed in addition • Growing demand for reliable identifiers• A standard for knowledge database
format would be useful
Library Systems
• Need to be ”OpenURL Enabled”• 0.1 ”pretty easy”; how about 1.0?
OpenURL and cataloguing
• What should be catalogued into the records?
• Every static URL in records will be a problem
• Static URLs still useful as long as they work
• Cataloguing rules revision?
OpenURL and the others
• DOI and URN are for persistent linking, OpenURL for context-sensitive linking Work well togethet
• OpenURL is not a new Z39.50. With OpenURL, you have to know what you are looking for.
• OpenURLs are not for human eyes
OpenURL in the Future
• OpenURL is gaining momentum
• Standardisation• Sharing of knowledge• Identification• Software development