Points of Pain, Peculiar

advertisement
Points of Pain,
Peculiar Possibilities,
& A Patron Paradise
or, A slightly arbitrary set
of hair-brained ideas
Roy Tennant
California Digital Library
Pilgrimage to Mecca
Life in the ‘Hood
Our users are increasingly using the Internet
for their information needs…can you say
“Google”?
As the younger generation grows to
adulthood, library funding and support may be
in jeopardy
We’re dying out here — even if it isn’t
immediately apparent
But not all gloom and doom — signs of
dissatisfaction w/Internet information offers us
a window of opportunity
We can and should trade on our reputation
What You Can Do For Us
In a nutshell: make every library
LOOK HUGE and FEEL PERSONAL
Build infrastructure and services that no single
library can build — create BUILDING
BLOCKS from which we can create services
Know when to be out front and when to let us
be out front
Offer compelling central services that drive
users to their local library
What Can We Do?
Out Google Google:
Return Google results along with a good deal
more
Build on our strengths:
Centralized metadata (WorldCat)
Dispersed service points (local libraries)
Think imaginatively
Ripoff good ideas from wherever they can be
found
The Basic Questions
What do libraries want?
What do library users want?
How can we get that for them (us)?
What Libraries Want
To provide for the information needs of a
clientele
To build useful collections and provide
effective services
To be used
What Library Users Want
To find what they want
To find as much or as little as they need
To experience as little pain as possible
To not have their time wasted
To have the option to control their
experience and make informed
decisions
To be effectively advised
Basic User Truths
Only librarians like to search, everyone
else likes to find
A cite in the hand is worth 10 in the
database
Good enough is just that
Pain avoidance is a powerful motivator
Points of Pain
Library catalogs suck as information
finding tools
There are too many possible sources to
search them separately
There is little advice about which resource
to search
There is no advice about which is better
(we know, but we’re not telling)
Why Library Catalogs Fail as
Information Finding Tools
They are unable to search the entire universe of
information
Local catalogs often lack books that can be
requested
They have too little information about items
Most are Unable to accept multiple metadata
formats
Many have hostile user interfaces (complexity is
often a sign of lazy or incompetent design)
Union catalogs often have multiple records for the
same item (which to request?)
What Better Case for FRBR?
FRBR: Functional Requirements of
Bibliographic Records, from IFLA
A recasting of bibliographic description into
levels:
Work
Expression (translations)
Manifestation (editions)
Item (copies)
Both RLG and OCLC are experimenting with it
Making the Pie
Other
A&I Dbs
OAI Repos.
DL Colls.
The Web
Integration Engine
WorldCat
Making the Pie: Metadata
Metadata: cataloging by those paid better than
librarians
Metadata: Structured information about an
object or collection of objects
We must become very, very proficient with
metadata — creating, harvesting, transforming,
serving; your Metadata Switch is very important
work
MARC is just the beginning, and unless we’re
careful, will be too limiting; we must be
proficient with Dublin Core, MODS, METS, etc.
Encoded
in TEI
XML
Full Text
Stored
File System
Search
Index
Encoded
in TEI
XML
Stored
File System
Search
Index
Full Text
Structure
Records Stored
Created
Selected
Fields
Extracted
Search
Index
METS
Repository
Project
Profile
MODS record
UC Press record
Library
Catalog
UC Press
Database
Encoded
in TEI
XML
Stored
File System
Search
Index
Full Text
Structure
Records Stored
Created
Selected
Fields
Extracted
Search
Index
METS
Repository
Project
Profile
MODS record
UC Press record
Library
Catalog
UC Press
Database
User
queries
Encoded
in TEI
XML
Stored
File System
Search
Index
Full Text
Structure
Records Stored
Created
Selected
Fields
Extracted
Search
Index
Results
in XML
METS
Repository
Project
Profile
MODS record
UC Press record
Library
Catalog
UC Press
Database
User requests
book
XSLT
Encoded
in TEI
XML
Stored
File System
Search
Index
Full Text
Java
servlet
Structure
Records Stored
Created
Selected
Fields
Extracted
Search
Index
METS
Repository
Project
Profile
MODS record
METS record in XML
UC Press record
XSLT
Library
Catalog
UC Press
Database
User requests
book segment
Encoded
in TEI
XML
Stored
File System
Search
Index
Full Text
Java
servlet
Structure
Records Stored
Created
Selected
Fields
Extracted
Search
Index
METS
Repository
XSLT
Project
Profile
MODS record
UC Press record
Library
Catalog
UC Press
Database
Book
segment
returned
Methods for Encompassing
Resources
“Ingesting” — centralized by an
individually tailored process
“Harvesting” — centralized by a process
applicable to an entire class of
resources
“Crawling” — software-based HTTP
fetching
“Dynamically Queried” — broadcast
search at the moment of user need
Making the Pie Principles
We never metadata we didn’t like
(metadata R Us)
Decentralize metadata maintenance
whenever possible
Centralize metadata searching
whenever possible — Federate, then
slice and dice
Metadata can be both mined and
enhanced
Slicing the Pie
Slicing can be pre-selected or dynamic
By:
region (e.g., Australia)
topic area
format type
ease and rapidity of access
When to Slice the Pie
Before searching:
Select general topic area
After searching:
Results clustering
Search within results
Slicing the Pie Principles
Strive to serve only that which will feed
the hunger
Few will want the whole pie; some will
want it sliced; others will want to slice it
themselves
Slicing must happen regardless of how it
was made
Serving the Pie
Provide ways for users to “drill down” in
search results
Guide the user to useful subject terms
Cluster search results
Rank by:
Numbers of holding libraries
Usage, e.g., “click through count”
Weights assigned by librarians, or
reflected in book reviews
Serving the Pie
We need ways to keep librarians happy
without enraging patrons (e.g.,
“advanced search” option)
Searching is an iterative process
A good search result is not the end, but
the beginning (e.g., provide ability to
format a bibliography, download or print
the citations)
Serving the Pie Principles
Best served by those who know the
consumer
Global services can (and should be)
locally branded to maximize service
delivery options for end users
Software “skins” are not new
Things We Must Do
No more business as usual!
Out Google Google (Google w/ added
value)
Get good at sucking things up
Be good producers and consumers of
metadata
Work together more broadly and deeply
Be user focused, but not user driven
Hire out of our ranks, read out of our
profession, get out more!
Recap
Help us LOOK HUGE and FEEL
PERSONAL
Think building blocks, extensibility,
flexibility, skins, richer and more diverse
metadata
Federate, then slice and dice
Free WorldCat!
Download