Meeting summary prepared by John James

Agenda

Introduction: Thomas Krichel (LIU)

Welcome: Michael Koenig (LIU), Pat Harris (NISO)

David Seaman (DLF): Business Models for Libraries in the Digital Age: the Roles of Linking.

Addressed core problems facing libraries: the data-silo problem. Advances in linking technology are a critical part of the solution to this pervasive problem of data isolation. This in not an anti-publisher viewpoint, but it is an anti-silo one. Publisher and libraries both have a role here.

Provided an overview of the DLF, emphasizing areas relevant to OpenURL

Data silos: what is the problem?

• Every publisher is an island producing silos of data that play badly with other.
• Publishers (including library publishers) build product that can only appear on their terms, in their interface
• Production of self contained units (the failure of the book).

The imagined user is...

• (a) Either someone so specialized that they only need access to one or two silos and they know what they are.
• (b) Someone with so much time that they can of in and out of many sites and learn multiple interfaces.

• (c) A librarian more than the scholar.

• This also images a user who wants to look at but no to gather, reshape, analyze, cross search, annotate, repackage and richly link within the article. A browser not a hunter/gatherer. A spectator not a participant.

What are users telling us?

Electronic content is increasingly important, but print predominates

Unmet needs and challenges: lack of bulk, time, integration, training hinder use.

• Content--not enough
• Time
• Retrieval--finding information
• Training

So many silos, so little time

• what do they do when faced with all the library-provided silos? They go to google.
• the library is absent from many pf the places users are going, so library-licensed material is often not found
• we ignore convenience at our peril
• we ignore low-level user dissatisfaction at our peril.

Digital impact on ILL, ereserves, etc. are all recognized good, but there is a low level grumbling, a dissatisfaction.

The looming force of user expectations

• Users at present have fairly low expectations beyind availability
• Expect a medium to behave according to their daily sense of its basic capacities
• Users don't work in physical libraries whose books are shelved by publishers - but our digital libraries are organized this way.
• Users turn to libraries for hel[--and libraries face a chronic inability to repackage content for local use--we are failing
• in our service mission. Services such as OpenURL are mission critical. Users face a chronic inability to repackage content and create linking.

What do we need to move forward? Linking+

• Malleability -- customer reshapeable data
• Multiplicity - the appropriate copy in the appropriate format
• Management - ability for libraries to build local services that allow users to interact richly across vendors. Pulbishers need to help libraries be data aggregation services for their users (OAI; OpenURL; DOI; CrossRef)

Closing thoughts

• Data that play well with others
• Innovative users need innovative data
• Tools exist, increasingly the will to use them
• The centrality of library skills - core traditional library skills turn out to be just those skills of enduring value.

Question and Answer session

DLF very recently fessed up to the fact libraries are publishers of sorts. Libraries are just as guilty as publishers in the data-silo issue. We compete on content just like publishers.

Herbert von de Sompel

The OpenURL Framework: Origins, Evolution, Concepts

The context: Library automation environment around 1998

• distributed information environment
• local and remote A&I databases
• rapidly growing e-journal collections
• need to interlink the available information

The Problem:

• links delivered by infor providers
• links are not sensitive to the users' context appripriate copy problems
• links dependent on business agreements between information vendors
• links don't cover the complete collections

The Real Problem:

• libraries have no say in linking
• libraries are loosing part of their core services

The Solution:

• DO NOT provide a link shich is an actual service related to a referenced item (e.g., a lin from a record in an A&I database to the corresponding full-text)
• BUT rather provide
  o a link that transports metadata about the referenced item is the Open URL to
  o others that better placed to provide the service

In other words the vendor provides metadata (OpenURL) to the library which provides the resolution of the metadata and identifiers into services (context sensitive)

Evolution ~ 1998

• Nature of solution determined
• Experiment with local databases at Ghent Univ

Feasibility tested in 2 complex environments in 1999: Ghent and UPS

OpenURL 0.1 released in 2000 and SFX linking server goes beta.

In 2001 integration of OpenURL framework and DOI/CrossRef framework

• Experiment involving CNRI, LANLm OhioLink, Academic Press, Ex Libris

In 2001 Generalization of OpenURL framework concepts beyond the scholarly information community (described in Van de Sompel in July/August 2001 D-Lib Magazine)

NISO OpenURL Standardization Charge

• Use existing OpenURL framework as a starting point
  o notion of context sensitive services
  o notrion of transporting contextual metadata packages to obtain context sensitive services

Must be extensible

Must address:

• OpenURL framework beyond scholarly resources

• contextual metatdata packages

OpenURL 0.1: scholarly metadata format; scholarly information identifiers; key/value formatted payload; http transport

NISO OpenURL stnadadization approach:

• not by specifying an OpenURL per cpmmunithe
• but by defining how communities can specify their own OpenURL frame work (Part I of Daraft Standard)

Giving an example of a specification (a la xml standard)

• Defines some core concepts:
• Context object
• Naming environment
• Formats
• Specifies the core components required to define an OpenURL framework for a community.

How do we manage it?

• introduces a registry to contain different instantiations for each of the core components
• introduces the community profile which is a selection of registry entries made by a communithy to implement its OpenURL framework.

Registry - Concrete:

Character encodings Physical representation Constraint languages Context object formats Metadata formats Namespaces Transports

OpenURL 0.1: core concepts

• contextual metadata - Referent is the core of the OpenURL

6 entities

• Entities explicitly present: referent, resolver, requestor, referring entity, referrer, service type

Pat Harris spoke briefly about Open Informatics filing of a patent. NISO does not perceive a conflict or threat to the development, implementation or use of the standard at this point as a result of the patent filing should the patent be issued.

Need resources to support standards: specifically need registry in this case.

Nettie Lagace, Ex Libris

Smart Linking with SFX

SFX is an independent link server.

Example: SFX server in your library contains everything your users have access to.

Examples of displays and flows of user searches in OpenURL aware databases

SFX knowledgebase

• comprehensive
• set of rules for linking
• library localizes these to match local subscription and coniditons using SFX tools
• regularly updated and distributed by Ex Libris

(PowerPoint slide examples attached)

{make catalog open url aware and send metadata to link server instead of a hard-coded link in the catalog}

Marrianne Parkhill, Endeavor Information Systems

OpenURL and What it Means for the Digital Library

Described Endeavor open linking product; talked about user behavior and OpenURL helps.

David Stearn, Yale University

Implementing Seamless Linking and Enhancing the Domain: Present Status, Investigations and future scenarios.

i. Put the OpenURL idea into context
ii. What Yale has discovered in sources and targets during implementation

Searching issues

• key words
• controlled vocabulary
• citation tracking
• cross-database searching

Linking issues

• DOI-based links/ CrossRef
• appropriate copy issue
  o local resolvers
• metadata driven (enhanced) links
  o imbedded links

Simultaneous searching across databases

OpenURL gives you enough data to go out and create a much richer set of links.

Searching

• Broadcast and Federated
  o Broadcast is real time
  o Federated is searching across pre-harvested data
• Integrated
  o Media type (metadata)
  o Normalize indexing
  
  • Vocabularies
  • Hierarchies

Linking

• Standard concerns
  o Multiple resources (html, pdf)
  o Local resolvers (subscriptions)
  o Local subscriptions (aggregators)
  o Customized link 'brand' (logo)
• Enhanced possibilities
  o Customized to local options
  
  • Local resources
  • Selected added value tools
• Metadata re-searching
  o Obtain additional descriptors
  o Search across additional engines

Problem

• Information overload

Solution

• Additional sources search with
  o Single search syntax
  
  • Immediate limit options before search
  • Navigational assistance
    # Local resource identified/linked
    # Relevant resources highlighted