The Future of Cataloging and Catalogers

Part 1: Whither Cataloging?

Libraries are no longer the first place people come for information

The Internet has changed the way people (including us) behave when seeking information

Our former “granularity consensus” is coming apart

To compete effectively for user attention, we must:

Join the larger world of information, where our users are

Learn how the competition attracts users, draws them in, and takes good advantage of their interest in participating

Find a better balance between protecting privacy and capturing usage behavior

12/5/08 NELINET Seminar

And Why Must We Do This?

The comfortable certainties we know are coming undone, whether we’re ready or not

We have much experience and insight to offer the larger information world (but not everything we’ve learned is relevant)

We are collectively about the size of the Queen Mary, unable to turn on a dime—this change will take time

Resistance is futile—we are not in charge of this new world, and our options are two: adapt or retire

The Map of Change Charting Our Course 12/5/08 NELINET Seminar

What We Must Leave Behind

A view of metadata based on catalog cards

Library software that can’t sort search results better than “random” or “alphabetic”

Search interfaces even Librarians hate (and we know the data)

Clunky static HTML pages that don’t attract our user’s interest, or guide them well

One silo for books, others for journal articles, images, digitized books, etc. (explain that to a user!)

Starting to Move Forward

A Starting Point: The Working Group on the Future of Bibliographic Control (Library of Congress)

“On the Record”—final report, January 2008 http://www.loc.gov/bibliographic-future/

A good, comprehensive overview of our new world and what we need to do

Recommendations for LC, OCLC, ALA, library educators and all of us

Extensively discussed at the Library of Congress and within the profession at large

“The Web is our platform”

1.2.4.2 All: Explore tools and techniques for sharing bibliographic data at the network level using both centralized and non-centralized techniques (e.g., OAI-PMH).

3.1.2.1 All: Express library standards in machine-readable and machine-actionable formats, in particular those developed for use on the Web.

3.1.2.2 All: Provide access to standards through registries or Web sites so that the standards can be used by any and all Web applications.

A New Look at Library Systems

4.1.1.1 All: Encourage and support development of systems capable of relating evaluative data, such as reviews and ratings, to bibliographic records.

4.1.1.2 All: Encourage the enhancement of library systems to provide the capability to link to appropriate user-added data available via the Internet (e.g., Amazon.com, LibraryThing, Wikipedia). At the same time, explore opportunities for developing mutually beneficial partnerships with commercial entities that would stand to benefit from these arrangements.

Enriching Library Data

4.1.2.1 All: Develop library systems that can accept user input and other non-library data without interfering with the integrity of library-created data.

4.1.2.2 All: Investigate methods of categorizing creators of added data in order to enable informed use of user-contributed data without violating the privacy obligations of libraries.

4.1.2.3 All: Develop methods to guide user tagging through techniques that suggest entry vocabulary (e.g., term completion, tag clouds).

Exploring Our New World Avoiding the Traps of Wrongovia 12/5/08 NELINET Seminar

Taking a Look Around

What’s this Semantic Web thingy all about, and why do we care?

Is RDA really going to happen?

Is it that different from AACR2?

Why can’t we use RDA with MARC?

How will RDA implementation affect cataloging?

How can we best prepare for all this?

Acronymia, We Are Here

RDA: Resource Description and Access

FRBR: Functional Requirements for Bibliographic Records

FRBRoo: Object Oriented FRBR (harmonized with CIDOC CRM)

FRAD: Functional Requirements for Authority Data

FRASAR: Functional Requirements for Subject Authority Records

Standards Upgrade! 12/5/08 NELINET Seminar Type of Standard Old Standard New Standard(s)? Bibliographic Model None FRBR, FRBRoo Metadata Content AACR2 RDA Metadata Structure MARC21 Bibliographic RDAVocab Name Authority MARC21 Authority FRAD Subject Authority MARC21 Authority FRASAR, SKOS Encoding MARC21 XML, XML/RDF

The RDA You’ve Heard About …

4 th quarter calendar 2008 – Full draft of RDA available for constituency review (ending in early February 2009)

http://www.collectionscanada.ca/jsc/rdafulldraft.html

2 nd quarter calendar 2009 – RDA content is finalized

3 rd quarter calendar 2009 – RDA is released

3 rd and 4 th quarters calendar 2009, possibly into 1 st quarter calendar 2010 – Testing by national libraries

1 st and 2 nd quarters calendar 2010 – Analysis and evaluation of testing by national libraries

3 rd -4 th quarters calendar 2010 – RDA implementation ?

What You Might Not Have Heard

JSC has gradually backed away from their original stance that RDA could be expressed easily in MARC21

Full integration of FRBR entities into RDA has made that problematic

RDA has been developed explicitly to take advantage of the Semantic Web (although there are still residues of past practice)

Well supported rumors indicate that LC is considering discontinuing update of MARC21 sometime in 2010

Under the RDA Hood

A FRBR-based approach to structuring bibliographic data

More explicitly machine-friendly linkages (preferably with URIs)

More emphasis on relationships and roles

Less reliance on cataloger-created notes and text strings (particularly for identification)

Less reliance on transcription

JSC Scenarios

Scenario 1: separate records for all FRBR entities with linked identifiers

Scenario 2: composite bibliographic records (with authority records representing each entity)

Scenario 3: one flat record, with all Group 1 entities on a single record

This is the only scenario that MARC can handle

Not really a viable option, and as far as I know, no one is explicitly planning for it

The Rest of the Story

RDA elements, roles and vocabularies have been provisionally registered

The vocabularies and the text will be tied together in the RDA online tool

Some efforts have begun to consider how MARC21 data can be parsed into FRBR entities and RDA

eXtensible Catalog Project moving strongly in this direction

Unfortunately, we don’t know what OCLC is planning

Discussions about long term maintenance of both RDA and the vocabularies have yet to occur

The push is already on for a multi-language RDA Vocabulary

RDA & FRBR: Registered!

RDA Elements:

http://metadataregistry.org/schema/show/id/1.html

RDA Roles:

http://metadataregistry.org/schema/show/id/4.html

RDA Vocabulary: Base Material

http://metadataregistry.org/vocabulary/show/id/35.html

FRBR Relationships (Sandbox version)

http://sandbox.metadataregistry.org/vocabulary/show/id/90.html

Who’s Doing This?

DCMI/RDA Task Group

See: http://dublincore.org/dcmirdataskgroup/

Set up during the London meeting between JSC and DCMI

Gordon Dunsire and Diane Hillmann, co-chairs

Karen Coyle & Alistair Miles, consultants

IFLA Classification and Indexing Section

Gordon Dunsire, Centre for Digital Library Research, University of Strathclyde, will be registering FRBR entities and relationships

Possible inclusion of ISBDs, FRAD, etc., in future

How Soon Will All This Happen?

The bad news: This isn’t like 1981, when there was a “start date” and we knew exactly when to change gears

More bad news: This transition is likely to be a pretty messy one, and last longer than we would like

One unknown is OCLC’s role—at present they seem to be focused on consolidating control over library data and promoting WorldCat

The good news: library vendors are starting to wake up and smell the coffee!

What Are the Challenges?

Coordination with JSC (or it’s successor, given the need to move beyond “Anglo-American”) on long-term maintenance planning

Need for lightweight process, where change is not a multi-year marathon

Continuing development towards a more Semantic web-friendly RDA (less reliance on transcription, for instance)

Tool development (at all levels, including ILS vendors)

Yet More Challenges

Application profiles that express more than one notion of “Work” and more than one point of view

JSC still seeing the process through the lens of a text cataloger

Their “core elements” only make sense for traditional books, serials, and other text-based objects

Moving the MARC legacy data into RDA

OCLC’s silence is worrisome, makes planning difficult

Multi-lingual and specialized extensions

Non-Anglo-American communities eager to participate

Multi-lingual RDA

The DCMI Registry approach:

Translations of labels, definitions and comments

URIs stay the same, as do relationships

Responsibility for updating translations rests with translation “owner”

Disadvantages

Translations tend to become outdated over time without sophisticated notification services to flag new areas needing attention

Communication with translation “owners” is managed loosely by a committee—support needs still unknown

Part 2: Whither Catalogers What Happens When The Revolution Comes? 12/5/08 NELINET Seminar

Focus on Catalogers

What do we anticipate will be different about our changed working environment?

How will workflow change?

How will the data look?

What will the library vendor systems do with it?

How will we integrate user data? What kinds of user data?

What do we need to know to operate in this new environment?

Approaching Change

Catalogers will need to separate what they know about information based on their current systems

Much of the knowledge is portable, but needs updating

The new environment is not as well organized (yet), so much learning will need to be self-directed

Catalogers’ role may become closer to that of Metadata Librarian

Managing data at a more abstract level (not yet a stable structure to fit data into)

Understanding of the goals of changes anticipated and new requirements will be essential

Walking through a concrete example … From the Cataloger Scenarios 12/5/08 NELINET Seminar

A Cataloger Scenario Jane Cataloger is assigned to work on a gift collection. Her first selection is a Latvian translation of Kurt Vonnegut's "Bluebeard: a novel." She searches the library database for the original work, and finds: * Author : Kurt Vonnegut * Title of the Work : Bluebeard: a novel * Form of Work : Novel * Original Language : English 12/5/08 NELINET Seminar

with links to the following expression information : * Language of Expression : English * Content Type : Text and one manifestation : * Edition : 1st trade edition * Place of Production : New York * Publisher’s Name : Delacorte Press * Date of Production : 1987 * Number of Units : 300 pages * Resource Identifier : [ISBN]0385295901 12/5/08 NELINET Seminar

Jane begins her description by linking to the existing Work entity. She then creates an expression description: * Language of Expression : Latvian * Translator : Arvida Grigulis She creates an authority record for the translator since none yet existed. She continues by creating a fuller description for the new manifestation, linking to the authority record for the Latvian publisher (what luck, it already existed!). * Title : [in Latvian] * Place of Production : Riga * Publisher’s Name : Liesma * Date of Production : 1997 12/5/08 NELINET Seminar

A Cataloger Scenario: Updating Jane Cataloger is assigned to work on a gift collection. Her first selection is a Latvian translation of Kurt Vonnegut's "Bluebeard: a novel." She searches the library database for the original work, and finds: * Author : Kurt Vonnegut * Title of the Work : Bluebeard: a novel * Form of Work : Novel * Original Language : English 12/5/08 NELINET Seminar

A Cataloger Scenario: Updated Jane Cataloger is assigned to work on a gift collection. Her first selection is a Latvian translation of Kurt Vonnegut's "Bluebeard: a novel." She searches the library database for the original work, and finds: * Author : http://lcnaf.info/79062641 * Title of the Work : Bluebeard: a novel * Form of Work : http://RDVocab.info/genre/1008 * Original Language : http://marclang.info/eng 12/5/08 NELINET Seminar

12/5/08 NELINET Seminar with links to the following expression information : * Language of Expression : English * Content Type : Text and one manifestation : * Edition : 1st trade edition * Place of Production : New York * Publisher’s Name : Delacorte Press * Date of Production : 1987 * Number of Units : 300 pages * Resource Identifier : [ISBN]0385295901

12/5/08 NELINET Seminar with links to the following expression information : * Language of Expression : http://marclang.info/eng * Content Type : http://purl.org/dc/dcmitype/Text and one manifestation : * Edition : 1st trade edition * Place of Production : http://www.getty.edu/tgn/7007567 * Publisher’s Name : http://onixpub.info/2039987 * Date of Production : 1987 * Number of Units : 300 pages * Resource Identifier : urn:ISBN:0385295901

12/5/08 NELINET Seminar Jane begins her description by linking to the existing Work entity. She then creates an expression description: * Language of Expression : Latvian * Translator : Arvida Grigulis She creates an authority record for the translator since none yet existed. She continues by creating a fuller description for the new manifestation, linking to the authority record for the Latvian publisher (what luck, it already existed!). * Title : [in Latvian] * Place of Production : Riga * Publisher’s Name : Liesma * Date of Production : 1997

12/5/08 NELINET Seminar Jane begins her description by linking to the existing Work entity. She then creates an expression description: * Language of Expression : http://marclang.info/lat * Translator : http://lcnaf.info/88007685 She creates an authority record for the translator since none yet existed. She continues by creating a fuller description for the new manifestation, linking to the authority record for the Latvian publisher (what luck, it already existed!). * Title : [in Latvian] * Place of Production : http://www.getty.edu/tgn/7006484 * Publisher’s Name : http://onixpub.info/6770094 * Date of Production : 1997

12/5/08 NELINET Seminar A Dublin Core View of the World DCMI Abstract Model: http://dublincore.org/documents/abstract-model/

12/5/08 NELINET Seminar Anatomy of a Statement Place of Production : New York Property Value Value String

12/5/08 NELINET Seminar Anatomy of a Statement Place of Production : http://www.getty.edu/tgn/7007567 Property Value Related Description

A Related Description 12/5/08 NELINET Seminar

12/5/08 NELINET Seminar Description Sets a Key Concept!

Description Set= “A set of one or more descriptions , each of which describes a single resource .”* *DCAM Definition 12/5/08 NELINET Seminar

12/5/08 NELINET Seminar A Description Set “Package” Work Manifestation Expression

New Tools, New Knowledge Getting There From Here 12/5/08 NELINET Seminar

What’s This Semantic Web?

RDF: Resource Description Framework

Statements about Web resources in the form of subject-predicate-object expressions, called triples

E.g. “This presentation” –“has creator” –“Diane Hillmann”

RDF Schema

Vocabulary description language of RDF

SKOS: Simple Knowledge Organisation System

Expresses the basic structure and content of concept schemes such as thesauri and other types of controlled vocabularies

An RDF application

OWL (Web Ontology Language)

Explicitly represents the meaning of terms in vocabularies and the relationships between them

Semantic Web Building Blocks

Each component of an RDF statement (triple) is a “resource”

RDF is about making machine-processable statements, requiring

A machine-processable language for representing RDF statements

Extensible Markup Language (XML)

A system of machine-processable identifiers for resources (subjects, predicates, objects)

Uniform Resource Identifier (URI)

For full machine-processing potential, an RDF statement is a set of three URIs

Things Requiring Identification

Object “This presentation”

e.g. its electronic location (URL)

Predicate “has creator”

e.g. http://purl.org/dc/terms/creator

Object “Diane Hillmann”

e.g. URI of entry in Library of Congress Name Authority File (real soon now?)

NAF: nr2001015786

Declaring vocabularies/values in SKOS and OWL provides URIs—essential for the Semantic Web

What Happened to XML?

Nothing: XML (eXtensible Markup Language) is most likely how library systems will evolve after MARC

It makes sense to use XML to exchange data between libraries, and some external services

But RDF is gaining ground, and libraries will need to be able to accommodate it, and understand it

An XML record is essentially an aggregation of property = value statements about the same resource

RDF triples can also be aggregated using XML, but this isn’t necessarily the best way to realize the potential of RDF

User Participation Bringing Users (and Usage) Into the Circle 12/5/08 NELINET Seminar

User Data “R” Us

Sources of ‘active’ user data

Tagging, etc.

Review and rating systems

Courseware systems

Sources of ‘passive’ user data

Logs of user activity

Circulation or download data

“ Making data work harder …” –Lorcan Dempsey

Collaborative filtering

Data mining

Active User Data

User tagging and description

Ex.: The LC Flickr Project

Ex.: LibraryThing

Review and rating systems

Ex.: Penn Tags

Ex.: Amazon

Courseware Systems

Making connections so that courseware can reuse catalog information; catalogs can know what has been used in courses, when, and who assigned it

LC-Flickr Project

Library of Congress and Flickr--“In a very elegant way, Flickr solves the authority conundrum of exposing collections content to social process. No need to worry if some comments or tags are misleading, arbitrary or incorrect - it’s not happening on your site, but in a space where people know and expect a wide variety of contributions. On the other hand, LC selectively reaps the benefit of these contributions.”

( http://hangingtogether.org/?p=401 )

Passive User Data

Logs of user activity

Usually locally maintained and analyzed

Services like Google Analytics can provide important aggregate information

Circulation or download data

Tricky in library settings, where user privacy an important value

Anonymized data can be stored and used for relevance ranking

Hard Working Data

Collaborative filtering

Wikipedia: “ … the process of filtering for information or patterns using techniques involving collaboration among multiple agents, viewpoints, data sources, etc.”

Ex.: Amazon (people who bought “X” also bought “Y”)

Data mining

Wikipedia: “ … statistical and logical analysis of large sets of transaction data, looking for patterns that can aid decision making.”

Ex.: LibraryThing Zeitgeist

User Data Issues

Privacy

Being able to use information about a contributing user without violating personal privacy

Complicated by differences in generational ideas about what privacy is

Authority (who said?)

Librarians have traditionally valued “objectivity,” but there’s no evidence that users see this as a value

Management

Keeping spammers out

Filtering language and malicious intent

Sharing User Contributions

Note how LibraryThing pulls Amazon descriptions

Amazon has an API that allows other services to use its data

Positioning Amazon data in other sites drives users back to Amazon

As libraries move more of their unique data to the Web, they need to be aware of the marketing value of sharing data and allowing other services to combine it in new ways