Federated Searching

1
Federated Searching

• An Overview of the 2002-2003 ENCompass
  Implementation at Loyola-Notre Dame Library

John W. McGinty, Director Charles Lockwood,
Digital Librarian
2
Loyola/Notre Dame Library

• Only single library organization and building
  supporting two independent colleges, opened 1973
• Two colleges enroll 9,200 undergraduate and
  graduate students, full and part-time
• Library supports two doctoral programs, twelve
  masters programs and 26 majors
• Collection priorities business, education,
  English, history, pastoral counseling, psychology
  and the sciences
• Blackboard course management system at both
  colleges, portal for library resources

3
Context Consortium
4
History of Our Implementation

• Fall 2001 purchase ENCompass, create digital
  team
• January 2002 install ENCompass 1.0 software
  (no federated search available)
• February design interface
• March add first local collection - dissertations
  
• April load Encompass 2.0 for Resource Access,
  add twenty working targets including Voyager OPAC
• Summer conduct focus groups with select faculty
• September go live to LNDL user community
• Jan 2003 upgrade to ENCompass 3.0 Beta

5
Our Digital Library Goals

• Leverage new technology to host local
  intellectual property or academic content
  previously neglected in collection development
  (e.g. from Blackboard)
• Present valuable resources within Blackboard and
  other campus points of access.
• Develop local digital collections, special
  projects (e.g. foreign language journals)
• Build a search interface for easy access to all
  these vital resources.
• Demonstrate non-research libraries can build a
  comprehensive digital library

6
Voyager and ENCompass Structures
Integrated Query and Display (XML/XSL)
Parallel Search Router/Handler
DTD Aware Search Engine
Voyager API
TEI
Library Collection
Full-Text Content
Courtesy Endeavor
7
Whats ENCompass Made Of? A Systems View

• Oracle is the Database Manager. All local
  objects are stored in Oracle
• Apache to display the web pages
• Tomcat - implementation of Java Servlet and Java
  Server pages

8
Architectures in ENCompass

• Interface(s)
• Database of Patrons from Voyager for access
  levels
• Content architectures remote databases, multiple
  formats, full-text collections
• Rights management

9
Federated Searching in ENCompass

• Proprietary databases via Z39.50, http, and xml
• Local digitized collections
• Other library catalogs
• Websites

10
Z39.50 Connections

• Pros
• Reliable
• Standardized
• Mature technology
• Plenty of documentation
• Easy to set up in ENCompass
• Major vendors Gale, OCLC, Ebsco supported
• Cons
• Better suited for citation and bib db than
  custom datasets with full text, images, etc.
• Problem with SUTRS records as opposed to MARC
  (incompatible with CSA)

11
HTTP Connections

• Pros
• Easy to set up (prepackaged)
• Can integrate web sites that do not support
• XML, Z39.50, theoretically, any site.
• Cons
• Each has to be custom coded to recognize text
  patterns of web pages. Secret recipes.
• Labor intensive, need programmers.
• Not guaranteed stable over long run.
• No documentation

12
XML Gateways

• Pros
• XML Perceived as the future of data exchange.
  Flexible and powerful.
• Now you can have Pub Med, Elsevier Science
  Direct
• Coming Lexis
• Cons
• Expertise needed to set up and administer.

13
Searching Structure

• We chose to make top level collections by subject
  (but could be by format, or anything you choose)
• Any new database added can be associated with any
  one or may top level subject collections.
• We use only general keyword and title options to
  quickly integrate Z39.50. Soon will add author,
  subject. (there are many more to choose from but
  not all databases support each)
• If you add a database not set up for a certain
  search attribute it will fail when searched on
  that attribute.
• To make searches finish fast we capped total hits
  at 50. There are advantages and disadvantages to
  this number.
• You can change database weighting and field
  weighting in System Admin, we have kept default
  values.

14
A Frequently Asked Question

• Q. Looking at all the hits coming back from each
  targets Id like to know, What is controlling
  the way hits are ranked by relevancy and sorted?
• A. It brings back the first x number of records.
  Those records are sorted by the remote source.
  Once records are retrieved they are sorted based
  on the choices the library makes in the sys admin
  client usually relevancy ortitle.

15
Scenario you have 2 weeks toset up federated
searching

• To Connect OPAC
• You need only find Web address and port. (10
  min)
• To Connect a Content Provider/Database
• Find connection information (10 min to 2 days)
• Database code name
• Library pass code
• IP port
• Search attribute numbers Title,Keyword,
  Author, etc.
• Test in Windows client (optional) (1 hour)
• Put in ENCompass Sys Admin as a Virtual
  Repository (5 min)
• Create Collection record with some metadata in
  (20 min) link repository
• Test in ENCompass (2 days)
• Done!

16
Staff Admin Screen Adding a Voyager OPAC
17
Adding a Virtual Z39.50 Repository in ENCompass
18
Attribute Selections in ENCompass Sys Admin
19
(No Transcript)
20
Federated Searching in a Consortial Context

• Do licensing agreements allow your different
  libraries to share database connections and
  passwords?
• Do your libraries want different interfaces?
• Do your libraries share a unified patron database
  from which to log in?
• Is academic work/curricular materials to be
  shared?

21
Digital Collection Development in aConsortium

•  
• Inventory electronic databases among MIC
  Libraries
• Identify commonly subscribed databases among MIC
• Prioritize integration what are MIC libraries
  priority or best databases?
• Investigate shared connections via one Z39.50
  connection (contractual, not technical issue) 
• Inventory possible digitization
  projects/collections (local) among MIC libraries
  and prioritize. 

22
Authentication Access

• ENCompass can authenticate to Voyager or LDAP
• Using second ENCompass Collection Manager or Test
  server, model individualized interfaces in
  combination with rights management software
  (LNDL, Endeavor, a MIC partner) 
• Inventory supplemental authentication methods
  specific to MIC libraries (e.g. Obvia, Ezproxy)
• Inventory broader authentication options in place
  in campus IT shops (e.g. LDAP services or Virtual
  Private Networks)

23
Consortial Planning Document
24
Ongoing Considerations

• Reality! Only 60-70 of all the resources we
  list on our web site are able to be used in
  federated search.
• What to do with remainder? Leave them out or co-
  list them in ENCompass

25
In this interface Federated Search resources have
the checkbox. Ones with no checkbox can not be
searched this way. They are just links to the
sites.
26
Our aim at the momentA. Only include federated
searchableB. Give them full text!
27

• Results set to 50 records

28

• Results with 10 per page listed (could choose up
  to 50)
• Metadata display for results basic Title in
  alpha order. You could add more fields or sort by
  date or other fields.

29
2003 Work Plan

• Jan 2003 Get bugs fixed in Beta, get a new
  build, customize interface.
• Feb 2003 Add Linkfinder Plus knowledge base
  for citation linking
• Mar-April 2003 second interface development for
  Mount St. Marys College
• May-Jul 2003 revisit Blackboard relationship,
  more rights management

30
Staff Organization

• Director theory, planning, resource allocation
• Digital Librarian project manager, software
  integration 
• Digital Assistant interface design, digitization
• Systems Librarian Voyager integration
• Network Specialist hardware, server, network
• Head of Cataloging/Metadata metadata structures

31

• Short Staffed?
• No Excuses
• if we can do it, you can do it
• John McGinty, Director Loyola Notre Dame Library
  jmcginty_at_loyola.edu
• Digital Access Support Phillip Fryer, Systems
  Librarian, pdf_at_loyola.edu, Charles Lockwood,
  Digital Librarian clockwood_at_loyola.edu, Youlanda
  Halterman, Digital Assistant, yhalterman_at_loyola.ed
  u, Steve Watson, Library Network Specialist
  swatson_at_loyola.edu,