Once and Future Metasearching

Once and Future Once and Future MetasearchingMetasearching

Once and Future Once and Future MetasearchingMetasearching

Roy TennantCalifornia Digital Library

Roy TennantCalifornia Digital Library

The ProblemThe Problem• Most users want to find everything

useful with one search• But our systems require them to know:

– How to select one or more databases– How to get to them– How to use the unique search options for

each

• How can we create systems that minimize what the user needs to know to get what they want?

• Most users want to find everything useful with one search

• But our systems require them to know:– How to select one or more databases– How to get to them– How to use the unique search options for

each

• How can we create systems that minimize what the user needs to know to get what they want?

The Metasearch SolutionThe Metasearch Solution• Just-in-time searching of multiple

sources• Also called cross-database searching• Federated searching: just-in-case • Metasearching is far from perfect• It is only one tool among many• It does not replace searching

individual databases for some users and purposes

• Just-in-time searching of multiple sources

• Also called cross-database searching• Federated searching: just-in-case • Metasearching is far from perfect• It is only one tool among many• It does not replace searching

individual databases for some users and purposes

Content Discovery Principles*Content Discovery Principles*• Only librarians like to search, everyone else

prefers to find• One place to search is better than two or

more• “Good enough” is just that• What is not searched is as important as what

is• If you can’t centralize metadata, centralize

searching

• Only librarians like to search, everyone else prefers to find

• One place to search is better than two or more

• “Good enough” is just that• What is not searched is as important as what

is• If you can’t centralize metadata, centralize

searching

* First hit of Google search: “digital library integration principles”

The State of Metasearch: Overview

The State of Metasearch: Overview

• Many database vendors still do not support a search protocol

• Software still in early stages

• Market still in upheaval

• Most libraries still learning from early implementations

• Google, Microsoft now big players

• Many database vendors still do not support a search protocol

• Software still in early stages

• Market still in upheaval

• Most libraries still learning from early implementations

• Google, Microsoft now big players

The State of Metasearch: The Landscape

The State of Metasearch: The Landscape

Publishers

Portal Providers

Indexers/Aggregators

Software VendorsLibraries

Users

Elsevier

CaliforniaDigitalLibrary

Google

The State of Metasearch:The Tyranny of NumbersThe State of Metasearch:The Tyranny of Numbers

• Number of resources being searched

• Number of results being returned by each resource

• Number of simultaneous users

• Number of resources being searched

• Number of results being returned by each resource

• Number of simultaneous users

X * Y * Z = potential disaster!X * Y * Z = potential disaster!

First GenerationFirst Generation

• Out-of-the-box interface with minimal customization

• “One-stop-shopping” model

• “Shotgun” searching

• No added-value services for specific clientele or purposes (e.g., UMNs Assignment Calculator)

• Out-of-the-box interface with minimal customization

• “One-stop-shopping” model

• “Shotgun” searching

• No added-value services for specific clientele or purposes (e.g., UMNs Assignment Calculator)

/V interface/V interface

2nd Generation2nd Generation• Using the metasearch application as a

service, not a destination• Creating search tools tailored to

specific audiences and/or purposes• Highly tailored interface• Better integration with

other systems and services

• Using the metasearch application as a service, not a destination

• Creating search tools tailored to specific audiences and/or purposes

• Highly tailored interface• Better integration with

other systems and services

Your Application

Metalib

DatabasesXMLhttp

* Slide courtesy of David Walker, CSU San Marcos

* Slide courtesy of David Walker, CSU San Marcos

* Slide courtesy of David Walker, CSU San Marcos

* Slide courtesy of David Walker, CSU San Marcos

* Slide courtesy of David Walker, CSU San Marcos

MetalibFederated Search Engine, hosted at the Chancellor’s Office.

Library CatalogDatabase of local book, media, and journal holdings, as well as print and media reserves.

CircuitDatabase of San Diego area academic library collections.

E-ResDatabase of electronic reserves.

SFXOpenURL link server, hosted at the Chancellor’s Office

MA

RC

-XM

L

OpenURL-XML

Custom XML

Custo

m X

ML

HTML

RSS

Reserves HarvesterGathers and combines reserves from catalog and Eres.

XML

Z39.50

MARC

Xerxes

Third GenerationThird Generation

• As many resources locally integrated as possible

• More resources available via API

• One-stop shopping — system organizes itself around user’s query

• Useful ranking

• Post-search clustering and enhanced browsing

• As many resources locally integrated as possible

• More resources available via API

• One-stop shopping — system organizes itself around user’s query

• Useful ranking

• Post-search clustering and enhanced browsing

Windows Live AcademicWindows Live Academic

Google ScholarGoogle Scholar

$30 from SpringerLink

$30 from SpringerLink

$40 from Ingenta

Not available except by person account or site license

FREE

A Brief History of Finding InfoA Brief History of Finding Info• 30 years ago: only print, very difficult and

time consuming• 20 years ago: arcane, mediated online

searching of many commercial DBs, CD-ROMs

• 10 years ago: unmediated online searching of many commercial DBs

• 1 year ago: easy, unmediated online searching of one free source

• Tomorrow?

• 30 years ago: only print, very difficult and time consuming

• 20 years ago: arcane, mediated online searching of many commercial DBs, CD-ROMs

• 10 years ago: unmediated online searching of many commercial DBs

• 1 year ago: easy, unmediated online searching of one free source

• Tomorrow?

In SummaryIn Summary• Metasearching (both library and

commercial) still primitive• But… quickly improving• Soon (if not already) databases will be

irrelevant from the user’s perspective• We no longer control access to information

(there are huge players in this space)• Either we get good in the finding game, or

we get gone

• Metasearching (both library and commercial) still primitive

• But… quickly improving• Soon (if not already) databases will be

irrelevant from the user’s perspective• We no longer control access to information

(there are huge players in this space)• Either we get good in the finding game, or

we get gone

Implications for InstructionImplications for Instruction

• We live in an age where even idiots can find good stuff!

• So on what should we focus?

• We live in an age where even idiots can find good stuff!

• So on what should we focus?

Implications for InstructionImplications for Instruction• Stop teaching mechanics and databases

as soon as possible • Work on building systems and tools that

don’t require instruction to use• Focus on fostering critical information

skills • Craft key messages and deliver them

consistently and ubiquitously across the library

• Stop teaching mechanics and databases as soon as possible

• Work on building systems and tools that don’t require instruction to use

• Focus on fostering critical information skills

• Craft key messages and deliver them consistently and ubiquitously across the library