OpenURL, MetaSearch, and Proxy Servers

1
OpenURL, Meta-Search, and Proxy Servers

• Digital Libraries II
• Shane Nackerud
• April 5, 2004

2
From Fragmentation...
3
to Integration.
4
Some Integration Technologies
5
Finding Full-Text
Starting Point
Database Citation
Maybe
Full-Text Article
Destination
6
Finding Full-Text Context-Sensitive Linking
Database Citation
Link Resolving
Full-Text Article
7
Demo Using Ex-Libriss SFX

• http//www.lib.umn.edu

8
SFX flow AI as link source
SFX Source
SFX Menu
Target SFX
9
(No Transcript)
10
(No Transcript)
11
How Does it Work?
12
Context-Sensitive Linking
Target
Link
OpenURL Metadata
Target
Source
Link Resolver
Link
Link
Target

• Accepts metadata from source
• Resolves against local database of targets and
  business rules
• Builds context-senstive URLs and presents them
  to the user.

13
Terms

• Sources. Any resource capable of generating an
  openURL. Basically, where the user starts.
• Targets. Web resources with a defined link-to
  syntax. Includes such things as full-text
  journals, OPACs, document delivery forms,
  Internet search engines and online encyclopedias.
  Basically, where the reader wants to go.
• OpenURL. A general protocol administered by NISO
  (Committee AX) which allows an information
  resource (such as a bibliographic database,
  e-print archive, or online catalog) to pass
  metadata about a citation to an institution's
  link resolver server.

14
Terms

• Digital Object Identifier (DOI). A unique
  identifier assigned to a digital object to
  uniquely identify it and provide a way to access
  the object without having to know where on the
  Internet it is located (the URL). A DOI may look
  like this 10.1000/123456.
• CrossRef. Cooperative service including nearly 80
  of the worlds leading STM publishers. CrossRef
  resolves the metadata, such as the citation
  information in a bibliographic reference, to a
  DOI. The DOI is then used to find the electronic
  version. See www.crossref.org.

15
OpenURL Metadata

• Article Citation
• Author York, Neil L.
• Title THOMAS CROWLEY'S PROPOSAL TO SEAT
  AMERICANS IN PARLIAMENT, 1765-1775.
• Citation Quaker History 2002 91(1) 1-19. ISSN
  0033-5053
• OpenURL
• http//linkresolver.lib.umn.edu?genrearticleatit
  leTHOMAS20CROWLEY27S20PROPOSAL20TO20SEAT20A
  MERICANS20IN20PARLIAMENT2C201765-1775titleQu
  aker20Historyissn0033-5053date2002volume91
  issue1spage1epage19aulastYorkaufirstNeil
  auinitL
• OpenURL Generator at http//demo.exlibrisgroup.co
  m9003/OpenURL/article.html

16
Context-Sensitve Linking
Full Text
Link
OpenURL Metadata
MNCAT
Database Citation
Link Resolver
Link
Link
ILL
17
Sources

• Commonly include
• Abstracting and Indexing Databases
• E-Journals
• OPACS
• E-print Archives
• Digital Content Repositories
• Any other information bases that are
  OpenURL-enabled

18
Targets

• E-Journals
• OPACs
• Union Catalogs
• Citation Databases
• A I Databases
• Patent Databases
• Encyclopedias
• Internet Bookstores

• Book Review Sources
• ILL/Doc Del
• Reference Services
• Local Expert Databases
• FAQs
• Other Extended Services

19
Competing with Google

• Google is easy, library searches are not
• Google provides full text all the time, libraries
  only sometime
• Google has one search box and a submit button,
  libraries
• Google is one stop searching, libraries typically
  subscribe to hundreds of different targets

20
Meta-Search

• The Holy Grail of library searching
• Searching multiple, disparate targets (databases
  and indexes) at once from a single search
• Presenting the user with
• Results from multiple databases?
• Merged results?
• De-duped results?
• Ability to refine search?

21
Challenges

• Meta-Search technology must be able to search
  multiple databases that do not
• Share a common database
• Share a common thesaurus
• Share a common interface
• Share a common set of search criteria
• Keywords? All Fields? Subjects? Descriptors?

22
Search Protocols

• Different search protocol may be necessary for
  each database
• HTTP and screen scraping
• XML
• Z39.50
• Different configuration almost always necessary
  for each database
• At least one breaks every day

23
Demo

• http//fddemo.hostedbyfdi.net/

24
(No Transcript)
25
Problems

• Librarians-
• Are we dumbing down the interface?
• Worth the bang for the buck?
• Takes away the need for a reference interaction?
• Publishers
• Can they handle the increased traffic?
• No branding at the publisher level
• How are results ranked?
• Small publishers have a good chance of being
  searched
• Users
• Is it still too complex?
• On campus vs. off campus, how do users
  authenticate to all these resources?

26
Proxy Servers

• Typical use
• Sits between the client and the desired server to
  see if it can serve the content itself
• Proxy server will either get the content, or it
  will serve content it has already retrieved
• Filters content blocks or monitors access
• Library use
• Act as a go between for off site users of
  licensed resources
• Part of library IP range given to licensed
  resources

27
Types of library proxy servers

• URL Rewriter
• Masks IP address of user
• Must configure all the domains manually
• Rewrites all the URLs of a target page to first
  go through the proxy server
• Traditional
• Masks IP range for user
• No need to configure any domains
• User must configure browser to go through the
  proxy server

28
EZProxy Demo
29
Problems with proxy

• EZProxy
• Must configure all the domains for licensed
  resources
• Time consuming and almost impossible
• Proxy servers in general
• Use IP address authentication
• Open proxy servers are becoming common
• Database vendors are getting jumpy
• Everything comes from one IP address

30
Shibboleth

• Attribute based authentication and authorization
  method
• User authenticates at origin institution and
  receives validated attributes
• User can then go to a licensed resource and tell
  them he/she is from X institution
• Licensed resource then looks at the valid
  attributes and allows or disallows user
• Based on trust between vendor and institution
• http//shibboleth.internet2.edu/shib-intro.html