How to Build, Display and Find METS Objects

1
How to Build, Display and Find METS Objects

• Nate Trail
• Digital Project Coordinator
• Network Development and MARC Standards Office
• Library of Congress
• ALA - June 25, 2007

2
Outline of Topics

• A little about the technology
• How we build and display digital objects
• Searching
• Other Tools
• Future Directions
• Conclusions

 
3
Technologies Used

• XML documents
• METS Files
• MODS Files
• XSLT processes
• Transformation
• Display
• MySQL Database
• Cocoon framework
• And a little Java

4
XSLT Transformation
MODS XML for Beaux Arts Trio recording
Stylesheet transforms it into a simpler
structure of elements, labels and values.
5
XSLT for Display
METS XML for Beaux Arts Trio recording
XSLT transforms it into an HTML or XHTML
display for the screen
6
Web Framework Concepts

• User input via URL

• Create small programs for easy-to-modify parts
• Separate Data from Action and Design

7
Key Elements in Describing an Object

• Bibliographic data
• Files
• Object type
• (sheetMusic,printMaterial
• audio,video, compactDisc
• recordedEvent, bibRecord,
• photo, Article, Score)

8
Bibliographic data (1)

• Harvested from Voyager ILS (SRU/z3950)
• Inserted into MySQL database

9
Bibliographic Data(2)

• Direct Data entry for items not previously
  cataloged

10
METS Maker Pipeline Step 1

• Bibliographic Data is extracted from the database
  and converted to MODS

11
METS Maker Pipeline Step 2

• Combine queries of
• Object type (SheetMusic, RecordedEvent etc.)
• Files on the server
• MODS data
• Rights metadata
• Creates a virtual file called pre-mets

Object Type
Files on Server
WebPipeline
pre- mets
MODSMD
RightsMD
12
METS Maker Pipeline Step 3

• METS Maker XSLT processes the file using the
  objects profile and builds a METS File

METS file for Beaux Arts Trio concert recording
13
Display Step 1

• Within-object Navigation

14
Display Step 2

• Descriptive information about the object

15
Display Step 3

• Page-turning or Playlists

(pre-display virtual file )
16
Display Step 4

• Add page framing and breadcrumbs

LC Presents header and high-level navigation
Breadcrumb navigation
LC Presents footer (not in view)
17
Veterans History Project Displays
Veterans History Project Experiencing War Header
Standard Veterans History Project Header
18
Searching

• Lucene is open source and scalable to large
  indexes Australia has 16 million items indexed
• Lucene indexes are easy to configure using XSLT
  METS ? Indexed item

19
Searching in LC Presents

• Single box search offered on home page
• Collection-level search offered for
  subcollections
• Searches may also be limited to certain fields
  or object types

20
Searching in LC Presents

• Canned search to create virtual collections

21
Searching in LC Presents
Browse by Subject, Name or Title

• New browsing capability

22
Searching in Veterans History
Limit by conflict, branch, gender, POW status,
etc.
23
Browsing in Veterans History
Browse by last name, state, race/ ethnicity and
war/branch
24
Searching in Minerva (across collections)
Choose among collections harvested.
MODS terms to search.
25
Other Web Tools

• SQL query any Oracle or MySQL tables from the
  browser
• Holdings queries (beyond z3950/SRU)
• Site administration
• Convert from MARC to MODS or MODS to MARC
• Convert HTML to XML
• Link checking

26
Future Directions

• New Behaviors, Profiles
• Article OCR
• Multivolume Monograph
• Generating RSS data on the fly
• Integrate JHOVE file inspection tool (and MIX
  metadata) into our METS objects
• Move to JPEG2000 file format for images

27
Conclusions

• The METS standard is flexible enough to
  describe multiple kinds of complex objects.

• Profiles in METS help define an object and its
  range of behaviors
• METS and MODS play well together

28
Conclusions, cont.

• MODS is extremely powerful as a structural tool,
  not just a bibliographic tool
• Consistent, authoritative, structured metadata
  makes search and display for an object persistent
  into future software and hardware systems

29
Conclusions, cont.

• Standards matter. Without standardized
  structures, we cant program a site to retrive
  or display items with any reliability.
• Open Source software is a major step forward for
  the ability of libraries and other lower budget
  institutions to enhance the availability and use
  of their digital assets.

30
Questions?
www.loc.gov/lcp (LC Presents Music, Theater, and
Dance)
www.loc.gov/vets (Veterans History Project)

• Nate Trail
• ntra_at_loc.gov

Library of Congress