Metadata crosswalks

Crosswalks

March 25, 2013
Richard Sapon-White

1

Overview
 Crosswalk definition and description
 Issues

2

Interoperability

Search interoperability
 The ability to perform a search over
diverse sets of metadata records to
obtain meaningful results
Today’s session focuses on sets of
records using different metadata
schemes

3

Definition
 An authoritative mapping from the
metadata elements of one scheme to
the elements of another
 Example:

Dublin Core to MARC Crosswalk

4

Reciprocal Crosswalks
 Two crosswalks are needed to map
from metadata scheme A to scheme B
AND
from scheme B to scheme A
 With two crosswalks, “round-trip”
mapping results in loss or distortion of
information
5

More Examples
 Library
of Congress has crosswalks for
MARC21 to/from
– DC (Dublin Core)
– FGDC Content Standards for Geospatial
Metadata (Federal Geographic Data
Committee)
– GILS (Global Information Locator Service)
– ONIX ((ONline Information eXchange)
6

Uses of Crosswalks
 Record exchange
 Union catalogs
 Metadata harvesting
 Search engines: query fields with
similar content in different databases
 Aid to understanding unfamiliar
schemes
7

Complexities of Crosswalk
Creation
 No standard format for metadata schemes
– Different properties of elements are specified
– Same properties may employ different terms
 Some elements may map to multiple
elements in a second scheme, or vice versa
 Elements may be repeatable in one scheme,
non-repeatable in another

8

Complexities of Crosswalk
Creation (cont.)
 Source scheme may specify an element
for which there is no comparable
element in the target scheme
 Differences in content rules (e.g., use of
a controlled vocabulary) or data
representation (e.g., Michał Kowalski
vs. Kowalski, Michał)

9

Issues in Crosswalking Content
Metadata Standards
Barriers to creating crosswalks
1. Lack of common terminology between
metadata schemes
2. Metadata standards are not organized in
the same way

Margaret St. Pierre and William LaPlant
http://www.niso.org/publications/white_papers/crosswalk/ (1998)

10

St. Pierre and LaPlant (cont.)
Barriers to mapping
 One-to-many mapping: source field contains
multiple keywords while target field is
repeatable with one keyword per field
 Many-to-one mapping: results in loss of
information
 Source element does not map to any element
in target
 Mandatory element in target without any
element in source
11

Example
 Dublin Core element “Creator” – an
uncontrolled name
 Creator did not map to MARC
 MARC name fields defined as main or
added entries (1xx, 7xx) - content
defined by AACR2
 To develop a crosswalk, a new 720 field
was added to MARC
12

Mapping DC Subject to MARC
 DC Subject
– the topic addressed by the work
– Can be qualified by the scheme (e.g., LCSH)
 MARC fields 600, 630, 650, 651, 653
– 600, 630, 650, 651 are controlled vocabulary with
indicator for the scheme used
– 653 is uncontrolled vocabulary
 If map to 653, then lose identification of
controlled vocabulary
13

Mapping DC Subject to MARC
(cont.)
 Cannot map to other subject fields since DC
doesn’t distinguish between them
 Suggestion: create new MARC field for
generic subject field (not done)
Unqualified:
653 ##$a (Index Term--Uncontrolled)
Qualified:
Scheme=LCSH: 650 #0$a (Subject added entry--Topical term)
Scheme=MeSH: 650 #2$a (Subject added entry--Topical term)
Scheme=LCC: 050 ##$a (Library of Congress Call Number/Classification
number)
Scheme=DDC: 082 ##$a (Dewey Decimal Call Number/Classification
number)
Scheme=UDC: 080 ##$a (Universal Decimal Classification Number)
Scheme=(other): 650 #7$a with $2=code from MARC Code List for 14
Relators, Sources, Description Conventions

Mapping DC Title to MARC
 DC Title does not distinguish between
title (245 $a) and subtitle (245 $b) or
any other kinds of titles
Unqualified:
– 245 00$a (Title Statement/Title proper)
– If repeated, all titles after the first: 246 33$a (Varying Form
of Title/Title proper)

Qualified:
– Alternative: 246 33$a (Varying Form of Title/Title proper)

15

Mapping DC Publisher to MARC
 One-to-one relationship between DC
Publisher and MARC 260 $b
 EASY!

16

Mapping DC Date to MARC
 Publication date in DC element Date best
maps to MARC21 260 $c
 Other dates exist in MARC21:
– 008/07-10: date in standardized form
– 260 $c can also include copyright or printing dates
Unqualified:
 260 ##$c (Date of publication, distribution,
etc.)

17

Mapping DC Date to MARC
(cont.)
Qualified DC:
Available: 307 ##$a (Hours, Etc.)
Created: 260 ##$g (Date of manufacture)
Issued: 260 ##$c (Date of publication,
distribution, etc.)
Modified: 583 ##$d with $a=modified
Valid: 518 ##$a (Date/Time and Place of
an Event Note). Text may be generated in $3
to include qualifier name.
18

Mapping DC Identifier to MARC
 DC Identifier is any string or number
used to uniquely identify an object
 Could be ISBN, ISSN, LCCN, URL
– Each coded differently in MARC21
 MARC 024 (other standard identifier)
could be used if type of identifier not
specified

19

Mapping DC Identifier to MARC
(cont.)
Unqualified:
 024 8#$a (Other Standard Identifier/Standard number or code)

Qualified:
 Scheme=URI: 856 40$u (Electronic Location and
Access/Uniform Resource Locator)
 Scheme=ISBN: 020 ##$a (International Standard Book
Number)

 Scheme=ISSN: 022 ##$a (International Standard Serial
Number)

 Scheme=(other): 024 8#$a (Other Standard Identifier/Standard
number or code) with $2=scheme value
20

Resolving Difficulties in
Crosswalk Creation: A Summary
 Create a new field in MARC
 Use qualifiers (Qualified DC) to map to
specific MARC fields
 If using unqualified DC, then map to
closest matching field (with loss of
some information)
– Some information maps to a “wrong” field
– Map to an “other” or “uncontrolled” field

21

Introduction to MarcEdit, from first run to philosophy

Terry Reese
Gray Family Chair for Innovative
Library Services
Oregon State University
Email: terry.reese@oregonstate.edu

Getting Started
1. Sample Data Files
– Sample MARC records need to be downloaded.
– Get them from:
http://oregonstate.edu/~reeset/marcedit/examples/session_
data.zip (~5 MB)
– Unzip the data to the Desktop
• Right click, Extract all to Desktop.
– Worksheet File
• Includes the examples that I’ll be working from:
– http://oregonstate.edu/~reeset/marcedit/examples/marc_worksheet.docx
– When you start MarcEdit for the first time, it will ask you to
update. Don’t. Tell it no – then we’ll turn off the automated
update checker.
– We’ll use this information later.

Keypoints
 What is MarcEdit?
– Background
– System Requirements
 Installation Notes
– First Run
 Understanding the Application Settings
– Editor Settings
– Language settings
 Accessing Application Data
 MarcEdit Infrastructure
 Getting Help
 Questions

What is MarcEdit?
 Started development in 1999
– Originally coded in 3 programming
languages: Assembler (libraries), Visual
Basic (UI) and Delphi (COM).
– Initially designed as a replacement for LC’s
DOS-based MARCBreakr/MARCMakr
software

What is MarcEdit?
 Today:
– Written in C#
– Continues to be freely available
– Supports both UTF/MARC8 charactersets
– MARC Neutral
– XML aware

Installing MarcEdit
 Windows:
– Installing from the Windows Installer
• 32-bit version:
http://people.oregonstate.edu/~reeset/marcedit/
software/development/MarcEdit_Setup.msi
• 64-bit version:
http://people.oregonstate.edu/~reeset/marcedit/
software/development/MarcEdit_Setup64.msi

– Installing using a Zip file:
• http://oregonstate.edu/~reeset/marcedit/softwar

Setting up MarcEdit
 Onfirst run, MarcEdit will ask you to
confirm some settings. These are
broken down into 5 areas
– MarcEditor
– Language
– Export
– MARCEngine
– Other

MarcEdit Export Properties
 Defines MARC
import
 Can capture port
output from record
input (much in the
same way OCLC’s
Connexion can)

MarcEdit: crosswalking design

 MarcEdit model:
– So long as a schema has been
mapped to MARCXML, any
metadata combination could be
utilized. This means that no more
than two tranformations will ever
take place. Example: MODS 
MARCXML  EAD

MarcEdit: crosswalking design

 MarcEdit Crosswalk model
– Pro
• Crosswalks need not be directly related
to each other
• Requires crosswalker to know specific
knowledge of only one schema
– Con
• each known crosswalk must be mapped
to MARCXML.

MarcEdit Crosswalking model

EAD

Dublin Core FGDC

MARC21XML

MARC MODS

MarcEdit: Crosswalks for everyone


 Example Crosswalks:
– MODS => MARC
– MODS => FGDC
– MODS => Dublin Core
– EAD => MODS
– EAD=>HTML


 What’s MarcEdit doing?
– Facilitates the crosswalk by:
1. Performing character translations
(MARC8-UTF8)
2. Facilitates interaction between binary
and XML formats.

Examples
 Project
Gutenburg RDF => MARC
 EAD=>MARC

MarcEdit Demo
 http://people.oregonstate.edu/~reeset/m
arcedit/html/index.php

38

Metadata crosswalks

More Related Content

What's hot (20)

Similar to Metadata crosswalks (20)

More from Richard.Sapon-White (20)

Recently uploaded (20)

Metadata crosswalks

Editor's Notes