OpenURL and OpenURL Framework: Demystifying Link Resolution

1
OpenURL and OpenURL Framework Demystifying Link
Resolution

• The promise of OpenURL 1.0
• Oliver Pesch
• Chief Architect, EBSCO Publishing
• opesch_at_epnet.com

2
Registry
Referrer
ReferringEntity
By Reference
Schema
Namespace
POST
KEV
UTF-8
ContextObject
OpenURL 1.0
Profiles
Service types
Referent
URI
Metadata formats
HTTPS
OFI
Resolver
By Value
Identifiers
URI
XML
GET
HTTP
Requester
Encoding
3
Topics we will cover

• History
• Item level linking and link resolvers
• What is an OpenURL
• Why OpenURL 1.0
• A tour of the standard
• Future possibilities
• Evolving challenges
• Next steps

4
History

• Mid 1990s digital collections take hold as does
  desire for interlinking
• 1998 Van de Sompel Hoshstenbach begin work on
  linking solutions
• 1999 NISO sponsors workshops about reference
  linking. Appropriate copy problem emerged as
  the major issue.
• 1998 SFX linking server created and OpenURL (0.1)
  specification published
• 2000-today OpenURL (0.1) sees wide adoption by
  information providers
• 2001 Van de Sompel Beit-Arie approach NISO to
  make OpenURL an official standard
• 2001 NISO Committee AX formed
• 2001-2003 Committee AX works to generalize and
  allow extensibility
• 2003 OpenURL 1.0 available for trial use

5
Scholarly institutions have many resources from
many vendors
ISI
ingenta
Publishersites
Full text DBs
EhostEJS
British Library
AI service
Google
6
Some services offer links to selective resources
ISI
ingenta
Publisher sites
Full text DBs
EhostEJS
British Library
AI service
Google
User
7
Some services offer links to selective resources
ISI
ingenta
Publisher sites
Full text DBs
EhostEJS
User
British Library
AI service
Google
8
Predefined links inconsistent and their
management complex
ISI
ingenta
Publisher sites
For the library, when it is possible to control
links at these services, they must manage on each
service. This could means 100s of places to
control links, update holdings, etc.
EhostEJS
Full text DBs
For the online service it is a burden to try and
link to each other with only bi-lateral agreements
British Library
AI service
Google
Strict service-to-service linking DOES NOT SCALE
and often does not involve the library.
9
Introducing a link resolver allows
morecomprehensive and easier to manage linking
ISI
ingenta
Publisher sites
Resolver can link to the resources. Information
sources need only to know how to link to a link
resolver. Resolver controlled by library. One
place to manage holdings and links. Information
Source need not worry about bi-lateral
development. A solution that definitely scales
Link Resolver
Full text DBs
EhostEJS
British Library
AI service
Google
10
Without link resolver user may notaccess the
appropriate copy
ISI
Pay per view
ingenta
A
Publisher sites
Full text DBs
EhostEJS
Library subscribes
A
With a predefined link, the user may be presented
with a link and not have a choice where that link
goes
British Library
AI service
Google
User
11
With a link resolver the institution
candetermine which copy is appropriate
ISI
Pay per view
ingenta
A
Publisher sites
?
Link Resolver
Full text DBs
EhostEJS
Library subscribes
A
With the resolver, the library can control where
the link goes and get the user to the appropriate
copy. As long as the AI service knows the user
has a resolver and where it is, it can create the
link.
British Library
AI service
Google
User
12
What is OpenURL (0.1)

• An accepted standard syntax for creating a link
  between an information source and a link resolver
• Pre-defines sets of data elements to use in
  describing an item
• Relies on HTTP protocol for transmission

13
OpenURL
Link Resolver
INFORMATION SOURCE
http//lib.example.edu/resolver.cgi
OpenURL is the connection between the Information
Source and the link resolver. OpenURL is
transferring the information about an item
(article) from the Information Source to the
resolver.
14
OpenURL 0.1
genrearticletitleJournal of Banking and
Financeissn0378-4266date2000-03volume24
issue3spage395aulastNarayananaufirstRang
aatitleInsider Trading and the Voluntary
Disclosure of Information by Firms
sidInfoSource
Journal Journal of Banking and FinanceISSN
0378-4266Date 2000-03Volume 24Issue
3Page 395 Author Narayanan, RangaTitle
Insider Trading and the Voluntary
Disclosure of Information by Firms
Article metadata is tagged using predefined
element set.
Genre is added to give the resolver information
on the nature of the item being referenced.
The source of the link is added as an SID.
http//lib.example.edu/resolver.cgi
TitleJournal of Banking and Financeissn0378-42
66date2000-03volume24issue3spage395a
ulastNarayananaufirstRangaatitleInsider
Trading and the Voluntary Disclosure
of Information by FirmssidInfoSource
Add the address and now the link is ready.
Address
The OpenURL is a self-contained link very much
tied to HTTP protocol.
15
OpenURL 0.1 - limitations
genrearticletitleJournal of Banking and
Financeissn0378-4266date2000-03volume24
issue3spage395aulastNarayananaufirstRang
aatitleInsider Trading and the Voluntary
Disclosure of Information by
FirmssidInfoSource
Allowable metadata genres and elements
pre-defined with no means to define new ones
Only provides for key-value pair (HTTP GET or
POST) representation of metadata.
OpenURL 0.1 is tied to HTTP transport
16
Why OpenURL as a NISO standard?

• Ensure wide acceptance
• Address specific shortcomings
• Support additional genres
• Support richer data formats
• Provide more complete context description
• Allow ability to send request by reference
• Support transports other than HTTP
• Provide environment for OpenURL to evolve in a
  controlled way

17
New genres

• OpenURL 1.0 (format/genres)
• Journal
• Journal
• Issue
• Article
• Conference
• Proceeding
• Preprint
• Book
• Book
• Book item
• Report
• Document
• Dissertation
• Patent
• Dublin CORE
• Others can be registered

• OpenURL 0.1 (genres)
• Journal
• Article
• Preprint
• Book
• Book item
• Conference
• Proceeding

18
Richer data formats

• OpenURL 0.1
• Physical representation
• Key encoded values

• OpenURL 1.0
• Physical representation
• Key encoded values
• XML
• Others can be registered
• Constraint languages
• Matrix
• XML Schema
• Others can be registered

19
Goals of the standard

• Create an OpenURL Framework
• Separate the information package from the
  transport
• Formalize the framework for the information
  package and for OpenURL itself
• Provide a means to extend
• Allow practical application of a generalized
  framework

20
Deliverables from NISO Committee AX

• 2 part standard
• Part 1 General Framework
• Part 2 Initial Registry content
• Registry
• Community profiles
• Implementation guidelines

21
Part 1 Context Object and Transport Mechanisms

• Defines the general framework for bundling
  contextual metadata
• Introduces the ContextObject
• Defines Formats and other elements
• Defines the registry
• Defines the notion of Community Profiles

22
OpenURL 0.1 example
In original OpenURL, you can see the information
package can be separated. It contains the genre,
the item description and the referrer.
http//lib.example.edu/resolver.cgi?
genrearticletitleJournal of Banking and
Financeissn0378-4266date2000-03volume24
issue3spage395aulastNarayananaufirstRang
aatitleInsider Trading and the Voluntary
Disclosure of Information by
FirmssidInfoSource
However, nothing within this package describes
the nature what is in the package. It is just
assumed to be key-value pair and based on genre
only one element set. It does not accommodate
different character encodings.
To offer ability to extend to different element
sets, and representations (e.g. XML), this
information package must contain information
about its make up.
23
The ContextObject

• The information package
• At the heart of the standard
• Describes a referenced item and the context
  within which it is being referenced
• Allows independence from transport method
• Formalizes expression of context

24
The ContextObject contains
Version control, character encodings, timestamp,
etc.
Administrative
Referent
Item being referenced. E.G. bibliographic
reference in full text article.
User making the request
Service creating the link. E.G. where the user
found the reference.
Requester
Referrer
The item which contains the reference. E.G. the
article in which the bibliographic reference was
found.
ReferringEntity
Resolver
ServiceType
The target of the link. E.G. the link server of
the users institution
The desired services from the Resolver. E.G.
Full text, ILL, Abstract, etc.
25
KEV ContextObject example
ContextObject
ctx_verZ39.88-2004ctx_tim2003-10-26
Administrative
ctx_verZ39.88-2004ctx_tim2003-10-26
rft_val_fmtinfoofi/fmtkevmtxjournal
rft.genrearticlerft.titleJournal of Banking
and Financerft.issn0378-4266rft.date2000-03
rft.volume24rft.issue3rft.spage395rft.au
lastNarayananrft.aufirstRangarft.atitleInsi
der Trading and the Voluntary
Disclosure of Information by Firms
rft_val_fmtinfoofi/fmtkevmtxjournalrft.genr
earticlerft.titleJournal of Banking and
Financerft.issn0378-4266rft.date2000-03rft
.volume24rft.issue3rft.spage395rft.aulast
Narayananrft.aufirstRangarft.atitleInsider
Trading and the Voluntary Disclosure
of Information by Firms
Referent
rfr_idinfosid/InfoSource.com
res_idhttp//lib.example.edu/resolver.cgi
Referrer
Resolver
rfr_idinfosid/InfoSource.com
res_idhttp//lib.example.edu/resolver.cgi
26
KEV ContextObject example
To make OpenURL we add the URL of the resolver.
Note the additional descriptive information.
url_ctx_fmt tag says that the ContextObject in
the back of the transport is formatted using
key/encoded values (KEV) as describe by the
matrix (MTX) for the ContextObject (CTX) that is
found on the registry using the
identifierinfoofi/fmtkevmtxctx
Load ContextObject on transport
http//lib.example.edu/resolver.cgi?url_verZ39.8
8-2004url_tim2003-10-25url_ctx_fmtinfoofi/f
mtkevmtxctx
Address and packing slip to driver
27
KEV ContextObject example
Link is ready for clicking. The payload of the
link, the ContextObject is on the back, the
driver has the address (baseURL) and there is the
packing slip that describes to the recipient what
is in the back and how it can be interpreted
28
Format

• A concrete method of expression for a class of
  information constructs. It is a triple
  comprising
• (1) a Physical Representation
• (2) a Constraint Language and
• (3) a set of constraints expressed in that
  Constraint Language

29
Example of metadata Format

• infoofi/fmtkevmtxjournal

The actual set of elements that have been defined.
Namespace to do with OpenURL
Constraint language is a table (matrix). This is
the form used to describe the list of possible
elements. Other examples are DTD or XML Schema.
Describing a format
Physical element representation is key-encoded-
values. E.G. aulastSmith
30
Matrix for infoofi/fmtkevmtxjournal
31
Using Format in a sentence

• http//example.org/myResolver?url_verz39.88-2004
  url_ctx_fmtinfoofi/fmtkevmtxctxrft_val_fm
  tinfoofi/fmtkevmtxjournalrfr_idinfosid/my
  id.commydbrft_idinfodoi/10.1126/science.275.5
  304.1320rft_idinfopmid/9036860rft.genrearti
  cle rft.atitleIsolation of a common receptor
  for coxsackie Brft.titleSciencerft.aulastBer
  gelsonrft.auinitJrft.date1997

32
Elements of the general Framework
ContextObject

• Entities
• ContextObject Formats
• Metadata Formats
• Identifiers
• Namespaces
• Character encodings
• Transports
• Profiles

33
Part 2 Initial Registry Content

• KEV and XML OpenURL instances
• Bootstrap standard for initial use by scholarly
  information community
• Defines HTTP-based transport methods

34
Registry Entries

• Namespaces
• Character encodings
• Physical representations
• Constraint languages
• ContextObject Formats
• Metadata Formats
• Transports
• Community Profiles

35
http//www.openurl.info/registry
36
ContextObject Formats
37
KEV ContextObject Format
38
KEV ContextObject Format
39
Metadata formats in registry
40
Journal KEV Metadata Format
41
Community Profiles

• San Antonio Level 1
• Built around Key/Encoded Value
• Provide close compatibility with original OpenURL
• Metadata Formats support new genres
• San Antonio Level 2
• Built around XML representations
• Similar Metadata Formats as SAP1
• Dublin Core
• XML based
• Dublin Core Metadata Format

42
Implementation Guidelines

• Centered on SAP1
• For implementers
• Describe necessary bits of the standard
• How to create KEV OpenURLs
• Demonstrates the upgrade path from OpenURL 0.1 to
  OpenURL 1.0
• Includes hybrid OpenURLs

43
The future for OpenURL

• Adoption of OpenURL 0.1 very solid
• OpenURL 1.0 can be extended to meet evolving
  needs
• Provides focal point standardizing link-to
  syntaxes
• Potential use in server-to-server applications

44
Evolving Challenges

• Adoption of the standard
• Data quality issues
• Knowledgebase
• Identifiers
• Authentication
• Extending interest within scholarly information
  communities and beyond

45
Next steps though NISO

• Standard will go to ballot
• Assign Maintenance Agency
• Advisory Committee
• Registration policies
• Assign Registration Agency
• Facilitate community interest groups
• Develop new uses for OpenURL or ContextObject
• Create additional profiles
• Register additional metadata formats and
  identifier namespaces

46
Thank you

• Oliver Pesch
• opesch_at_epnet.com