Let’s Get Together

Or, How to Integrate Library Resources so Our Users Can Find Them

John Mark Ockerbloom

University of Pennsylvania

November 19, 2000

The Problem

We produce and acquire lots of digital resources….

E.g. 3000+ electronic journals, 160+ databases, 100,000 digital images, hundreds of electronic books

But, our users often don’t get matched up with the resources they need

They don’t know the existence of a resource they could use

Or, they know it exists, but can’t find it easily

Many find a few resources they’re comfortable with and stick with those-- even when better resources exist for a given task

Database usage

Although we subscribe to over 150 databases

the top 20 in 1999-2000 represented 77.5% of the logins and 88.6% of the searches

the top 6 represented over 1/2 the logins, and about 2/3 of the searches (MEDLINE alone represents 1/3 of searches)

Maybe these are just the best databases

Or maybe others aren’t being used as they should:

People don’t know about the others

People find it too much work to use the others

User surveys, reference interactions suggest underutilization of other databases

Finding resources:
One size does not fit all

Search engines: Not a panacea

Finding electronic journals: Before

Finding electronic journals: After

Electronic journals:
What changed?

We now keep our ejournal information as MARC records in our regular Franklin catalog

We defined new fields for new information not used in normal MARC records

We implemented ways for bibliographers to quickly insert new minimal ejournal records

We consulted with bibliographers and catalogers to develop an efficient workflow for manual and automatic insertion of records

We wrote tools to harvest, search, and display ejournal records

Finding databases

Finding out what’s new

Location aid:
Communities of Interest

Roughly correspond to subject areas

But also can be used for interdisciplinary work

Should be flexible enough to create new ones as needed

Now present for journals, DBs, new materials

How do we do community assignments?

Manually at small scale (e.g. databases)

Automatically at large scale (e.g. new materials) from other existing metadata

Users should be able to browse all media types by community of interest…

…and create new communities of their own

Again, manually or with automatic filtering

Location aid:
Personalized Notification (SDI)

Lets people know when new things arrive:

New books and other materials in their subject areas

New issues of journals that are of interest to them

Notification can be via email or via Web “portal”

User design can make or break SDI service:

Don’t spam them with irrelevant or uninteresting info

(Let them sign up, and specify interests narrowly)

(Make it easy to get more or less detail)

Don’t make them spend hours on setting up profiles

Let them specify things they like as they go along)

(Give them good starting points, such as community-centered notification templates)

Zooming in on resources

Placing resources where they’re needed

User-centered ways to make resources findable

Organize access to resources around user tasks

Not around particular technology, or librarian practice

But should be easily integrated with librarian workflow

Allow different ways of searching, browsing, locating, based on different user needs

Place links and controls where users will be likely to use them as they go about their work

Requires: Tools that can be composed and reconfigured for various purposes

We’ve repackaged catalog searches for many of our services

Requires: A well-structured information base

Much easier to work with than static web pages, script tricks

Once metadata is in a database, many views of it possible

User-centered ways to present resources: Flexible browsing

User-centered ways to present resources: Reference links

More information

Much of what I describe is work in progress

Experience reports, questions, comments, ideas, collaboration welcome

Mike Winkler and Delphine Khanna implemented many of the examples seen here

Useful addresses to remember:

http://www.library.upenn.edu/ (main library site)

http://digital.library.upenn.edu/ (digital library development)

ockerblo@pobox.upenn.edu (my email address)

Thanks!


	Or, How to Integrate Library Resources so Our Users Can Find Them

	John Mark Ockerbloom
	University of Pennsylvania
	November 19, 2000


	We produce and acquire lots of digital resources….
		E.g. 3000+ electronic journals, 160+ databases, 100,000 digital images, hundreds of electronic books
	But, our users often don’t get matched up with the resources they need
		They don’t know the existence of a resource they could use
		Or, they know it exists, but can’t find it easily
		Many find a few resources they’re comfortable with and stick with those-- even when better resources exist for a given task


	Although we subscribe to over 150 databases
		the top 20 in 1999-2000 represented 77.5% of the logins and 88.6% of the searches
		the top 6 represented over 1/2 the logins, and about 2/3 of the searches (MEDLINE alone represents 1/3 of searches)
	Maybe these are just the best databases
	Or maybe others aren’t being used as they should:
		People don’t know about the others
		People find it too much work to use the others
	User surveys, reference interactions suggest underutilization of other databases


	We now keep our ejournal information as MARC records in our regular Franklin catalog
		We defined new fields for new information not used in normal MARC records
		We implemented ways for bibliographers to quickly insert new minimal ejournal records
	We consulted with bibliographers and catalogers to develop an efficient workflow for manual and automatic insertion of records
	We wrote tools to harvest, search, and display ejournal records


	Roughly correspond to subject areas
		But also can be used for interdisciplinary work
		Should be flexible enough to create new ones as needed
	Now present for journals, DBs, new materials
	How do we do community assignments?
		Manually at small scale (e.g. databases)
		Automatically at large scale (e.g. new materials) from other existing metadata
	Users should be able to browse all media types by community of interest…
	…and create new communities of their own
		Again, manually or with automatic filtering


Lets people know when new things arrive:
	New books and other materials in their subject areas
	New issues of journals that are of interest to them
	Notification can be via email or via Web “portal”
User design can make or break SDI service:
	Don’t spam them with irrelevant or uninteresting info
		(Let them sign up, and specify interests narrowly)
		(Make it easy to get more or less detail)
	Don’t make them spend hours on setting up profiles
		Let them specify things they like as they go along)
		(Give them good starting points, such as community-centered notification templates)


Organize access to resources around user tasks
	Not around particular technology, or librarian practice
		But should be easily integrated with librarian workflow
	Allow different ways of searching, browsing, locating, based on different user needs
	Place links and controls where users will be likely to use them as they go about their work
Requires: Tools that can be composed and reconfigured for various purposes
	We’ve repackaged catalog searches for many of our services
Requires: A well-structured information base
	Much easier to work with than static web pages, script tricks
	Once metadata is in a database, many views of it possible


	Much of what I describe is work in progress
		Experience reports, questions, comments, ideas, collaboration welcome
	Mike Winkler and Delphine Khanna implemented many of the examples seen here
	Useful addresses to remember:
		http://www.library.upenn.edu/ (main library site)
		http://digital.library.upenn.edu/ (digital library development)
		ockerblo@pobox.upenn.edu (my email address)
	Thanks!