Metadata and Digital Libraries

1
Metadata and Digital Libraries

• The use of TEI and EAD in the University of
  Torontos Barren Lands Digital Archive

By Marlene van Ballegooie and Sian Meikle
University of Toronto Libraries
2
Background to the Barren Lands

• Project focused on two exploratory surveys
  conducted by J.B. Tyrrell in 1893 and 1894
• Approximately 5,000 images
• Full-text searchable

3
Background to the Barren Lands

• Maps

4
Background to the Barren Lands

• Letters

5
Background to the Barren Lands

• Diaries

6
Background to the Barren Lands

• Notebooks

7
Background to the Barren Lands

• Photographs

8
Background to the Barren Lands

• Newspaper clippings

9
Background to the Barren Lands

• Published works

10
Online Archives what approach?

• Question how to represent archives in an online
  environment?
• Exhibition approach low searchability but
  provides context
• Database approach high searchability, low
  context
• Neither enough for Barren Lands

11
Online Archives what standards?

• Library of Congress American Memory project,
  presenting its archival collections online, uses
  EAD
• California Digital Library, collection level MARC
  records link to EAD finding aids

12
What is SGML?

• EAD is an instance of SGML
• SGML Standard Generalized Markup Language
• device and system independent
• SGML is a metalanguage rules for defining a
  markup language
• SGML Document Type Definitions (DTD) spell out
  the allowable language

13
SGML Example

• Shows structural relationships between the
  elements of a document

ltdocumentgt lttitlegtThe story of my lifelt/titlegt
ltauthorgt ltfirstnamegtJohnlt/firstnamegt
ltsurnamegtSmithlt/surnamegt lt/authorgt
lt/documentgt
14
What is XML?

• SGML is large and complex
• XML (extensible markup language) is a subset of
  SGML
• developed by W3C
• designed for use over the web
• Many DTDs available in SGML and XML versions

15
Encoded Archival Description (EAD)

• EAD is a DTD for encoding archival finding aids
• EAD non-proprietary standard
• Available in SGML and XML versions
• Represents the multi-level nature of archives

16
EAD Structure

• lteadgt
• lteadheadergtContains info on title, author,
  creation date. etc. of the finding
  aidlt/eadheadergt
• ltarchdescgtWrapper that holds all descriptive info
  such as title, date, extent, biographical sketch,
  scope and content, administrative and arrangement
  info, notes, subject headings, etc.
• ltc01gtSeries level description e.g.
  Correspondence
• ltc02gtSub-series level description e.g. Family
  Correspondence
• ltc03gtFile level description e.g. Letters to
  William C. Tyrrell
• lt/c03gtlt/c02gtlt/c01gtlt/archdescgtlt/eadgt

17
Encoding Finding Aids With EAD

• EAD allows for description of itemsBUTnot
  sufficient for Barren Lands project. Why?
• Could not include all metadata elements
• Could not embed OCR effectively
• Could not include the necessary structural
  metadata

18
Describing Items

• What metadata standards are used to describe
  items?
• For books, TEI Lite is a popular standard
• Used by Making of America (MOA), Early Canadiana
  Online (ECO)

19
What is TEI?

• International standard for encoding electronic
  textual materials
• Developed by humanities scholars in 1987,
  guidelines published in 1994
• TEI-Lite - A subset of TEI containing the most
  commonly used tags

20
TEI Example

• The header
• ltteiHeadergt ltfileDescgt lt/fileDescgt
• Bibliographic information Author, title,
  publisher, etc
• ltencodingDescgt lt/encodingDescgt
• How the material was modified when digitized
• ltprofileDescgt lt/profileDescgt
• Non bibliographic information subject
  descriptors, etc.
• ltrevisionDescgt lt/revisionDescgt
• Revision log for digital version
• lt/teiHeadergt

21
TEI Example

• The body
• ltbodygt
• The document itself. Contains structural
  elements in turn. Some examples
• ltdiv1gtltdiv7gt
• Describe major structural divisions
  hierarchically
• ltpgt, ltpbgt
• Marks boundaries between pages of text
• lttitlePagegt, ltepigraphgt, ltsalutegt,
• For the title page, epigraph, salutation, and so
  on
• ltrefgt, ltxrefgt
• For pointers to other places, possibly with
  explanatory text
• lt/bodygt

22
Item Level Description

• Barren Lands tags included
• ltteiheadergt and its sub-elements to capture
  descriptive metadata
• ltxrefgt for linking documents
• ltpgt, ltpbgt to demarcate pages and hold structural
  metadata

23
Text Encoding

• Descriptive metadata encoded in TEIheader
• Used optical character recognition (OCR) and
  manual rekeying to obtain text files
• Placed each page of text within paragraph tags
  ltpgt in the body of the TEI record

24
Structural Metadata

• Structural metadata needed to keep track of image
  sequence
• Encoded in the empty page break ltpbgt element
• Example
• ltpb idtyrrell/text/T10001/0004 n0
  typeill titleA Valley in the Barrens
  rotateyesgt

25
Barren Lands Two tiered approach
TEI Item Description
EAD finding aid Item
Item Item
Programmed link
Extrefloc link
TEI Item Description
Programmed link
Extrefloc link
TEI Item Description
Programmed link
Extrefloc link
26
(No Transcript)
27
Did this approach work?

• Modifications preserved intent of DTD, but added
  information via attributes
• Minimal structural encoding allowed us to meet
  project needs cost-effectively
• Use of standards allowed us to use standard
  crosswalks to extract information for data
  warehousing

28
Further resources standards

• XML in 10 points (W3C)http//www.w3.org/XML/1999/
  XML-in-10-points
• TEI lite tag listhttp//www.tei-c.org/Lite/U5-tag
  list.html
• Encoded Archival Description An Introduction and
  Overview by Daniel Pittihttp//www.dlib.org/dlib/
  november99/11pitti.html
• Library of Congress standards http//www.loc.gov/
  standards/
• Digital Library Federation standardshttp//www.di
  glib.org/standards.htm

29
Further resources digital archives

• Images Canadahttp//www.imagescanada.ca/
• Making of Americahttp//moa.umdl.umich.edu/index
  .html
• American Memoryhttp//memory.loc.gov/
• Barren Landshttp//digital.library.utoronto.ca/T
  yrrell/
• Virginia Digital Libraryhttp//www.lva.lib.va.us
  /dlp/