The Lay of the Land: Libraries at the Crossroads Roy Tennant California Digital Library.

The Lay of the Land: Libraries at the Crossroads

Roy Tennant

California Digital Library

GoalsRaise questionsSpark imaginationsMotivateEncourage professional self-criticism

More Specifically… I will focus on our primary and most shameful

failure: our inability to provide an easy and effective information locating tool

Remember: only librarians like to search, everyone else likes to find

However, we are failing even to do things we have explicitly tried to do

Let’s take a look at the evidence…

260 Berkley, CA : Library Solutions Press, [c]1993 300 [vii,]134 p. : ill., maps ; 28 cm 500 Includes bibliographic references (p. 32-35) and index 650 0 Internet

250 1st ed 260 Berkley, CA : Library Solutions Press, c1993 300 viii, 134 p. : ill., maps ; 29 cm 500 Includes bibliographic references (p. 32-35) and index 500 "An earlier version of this book was published as a workbook in support of hands-on Internet training workshops." 650 0 Internet

250 1st ed 260 Berkeley, CA : Library Solutions Press, c1993 300 viii, 134 p. : ill. [,maps]; 28 cm 504 Includes bibliographical references (p. 32-35) and index 650 0 Internet -- Handbooks, manuals, etc 650 2 Computer User Training 650 2 Computer Communication Networks

4

3

9

Typical SearchesKnown Item “A Few Good Things”Comprehensive

Typical Searches: Known ItemThe good: searches can be limited to a

particular field: author, title, etc.The bad: limiting to a particular field

doesn’t always act the way you expectThe ugly:

The Really Ugly

Typical Searches:“A Few Good Things”The one type of search we have so far

ignored in library system designA type of search that we can do

something about todayBring Google-style relevance to library

catalogs

Typical Searches: Comprehensive Most library catalogs hide many things

available via regional cooperative or ILL It is difficult, if not impossible, to search all

appropriate journal databases Most libraries do not provide good access to

gray literature and web sites Subject headings are often unintuitive, and

catalogs give no guidance Catalogs give no chapter-level access to

book content

Some of the Things Most Users Care AboutWhat information resources are accessible to

themWhat they have to offer, in more detail

(contents, index, cover copy, etc.)What others think about themHow much pain they must endure to get themWhat they can expect when they show upWhat they must do with them when they’re

done

Some of the Things Most Users Do Not Care AboutMany of the things we care aboutWhere the information comes fromWho is responsible for providing itQuality, if it means spending a lot of

time and effort to get itDifferences between printings of the

exact same bookThe height of a book (in centimeters!)

What Many Users ExpectA simple search boxAutomatic filters, sorts, and groupings,

and/or some that they can apply Fault-tolerant search systems (“If you

can’t give me exactly what I asked for, do your best to give me what I want”)

Let’s see how fault-tolerant we are…

Acquired Immunodeficiency Syndrome -- Africa Acquired Immunodeficiency Syndrome -- epidemiology -- Africa Acquired Immunodeficiency Syndrome -- transmission AIDS (Disease) -- Africa AIDS (Disease) -- Africa AIDS (Disease) -- Etiology AIDS (Disease) -- Public opinion AIDS (Disease) -- Social aspects AIDS (Disease) in mass media Arts and society -- History -- 20th century Culture -- Philosophy Ethnic arts Marginality, Social -- History -- 20th century Mass Media Minorities in art Prejudice Public Opinion Race Relations Racism

Recap on Library Catalogs We cannot claim to support any of the top

three main types of searches well Our systems work inconsistently and

demonstrably incoherently Other bibliographic search systems (e.g.,

Amazon) demonstrate how pitiful our systems are to our users

We have taken very few steps toward fixing our broken systems

What We Have A computerized card catalog focused on inventory

control Non-standard database records Systems that don’t interoperate In union catalogs, multiple catalog records for the

same book An A&I database Tower of Babel Haphazard attempts to provide access to web sites Limited experiments providing access to gray

literature

What We Must Do We should design our systems for 80% of our user

needs, not 20% We must design the public view of our catalogs for

searching, not inventory control We should stop worrying about things that don’t matter

(e.g., book measurements) and start worrying about things that do (e.g., our inability to use one record per book)

We must think imaginatively and critically about how to design useful search systems

We need to design systems to integrate access, not fracture it

The Road Not (Yet) TakenCreate effective methods to put users in

touch with what they need, wherever it can be found

Design fault-tolerant, multi-purpose systems

Build for interoperabilityStrive for the Holy Grail of Librarianship:

one-stop searching for everything

What Most Users Want

How We Can Give it To Them

OAI-CompliantArchives

GoogleWorldCat on

SteroidsSerial

Databases

DigitalLibrary

Collections

The Integration Engine

The UserInterface

OnlineReference

Local Circulation Systems

http://searchlight.cdlib.org/cgi-bin/searchlight

Source: ARL Statistics

The Integration Engine Requirements:

Parse the query for each database Sort, organize, and de-dup the results Rank according to perceived relevance Be fault-tolerant (do the best it can with what it’s

given)

Targeted search engines may be better: Specific topic areas “A few good things” vs. Comprehensive

Concluding Thoughts We’re failing at our own goals We need to think imaginatively about our

challenges No library can do this alone Regional cooperatives are the smallest unit

for tackling this problem A regional cooperative with vision and guts

could lead the way for the rest of us