Collection Specific Vocabularies

1
Collection Specific Vocabularies

• March 2004

2
Terminology

• CB - abbreviation for collection builder
• CV - abbreviation for controlled vocabulary

3
Controlled vocabularies in DLESE

• Enables browsing and searching by
• Resource type and grade range
• Educational standards
• Collection names
• Subject (browse only)
• Ensures consistent metadata content in some
  required and optional fields
• Applies to all collections
• Encourages consistent terminology usage

4
Collection specific vocabularies

• Terms/phrases CBs think describe collections
  better than the framework capabilities of
• Free text descriptions for each framework section
• Annotation records
• Controlled vocabs
• Use only in optional fields of ADN framework
• Applies to a single collection
• Should not use terms in DLESE-wide vocabs
• Should have easy to understand definitions

5
Putting collection specific vocabs into the ADN
framework

• Keywords method
• Collection builder controlled terms method
• XML schema method
• Notes -
• When collections first ask about this they
  generally envision method three
• Three collections implemented method 1

6
Keywords method

• Develop a list of terms or phrases
• Use the terms or phrases consistently in the
  keyword field of every record
• No browse capability
• Use keyword search to find records
• No definitions needed
• CB does all the work

7
Keyword methodPros Cons

• Simple for DPC and CB
• Searching yields good results
• CB has complete control and responsibility
• No metadata framework changes
• No impact on other collection builders

• No specific browse
• CB has complete control and responsibility
• No integration into DLESE systems and services

8
Collctn bldr contr. terms method

• Develop a list of terms/phrases
• Write definitions for terms/phrases
• Provide a URL to the definitions
• Use terms consistently
• Be very consistent in spelling, spacing and
  capitalization when creating metadata content
• DPC can do browse if CB is consistent with
  metadata content

9
CB controlled terms methodPros Cons

• Simple for CB
• Searching yields good results
• Browsing capability can be developed
• CB has complete control and responsibility
• No framework changes
• No impact on other collection builders

• Browse capability requires DPC work and close
  synchronization with CB
• CB must use terms consistently and be good about
  spelling/spacing
• Vocab not easily integrated into other DLESE
  systems or services

10
XML schema method (1) (CB does)

• Develop a list of terms
• Define the terms
• Provide attribution for the terms
• Encode the list of terms into XML that is
  compatible with the XML of the appropriate DLESE
  metadata framework
• Register the vocab and XML files with DLESE
• Keep the terminology updated

11
XML schema method (2) (DPC does)

• Develop a vocab registration system
• Encode the vocab terms and definitions into the
  DLESE systems for use in discovery
• Issue a new version of the particular metadata
  framework (because we do so when vocabs are
  added)
• Upgrade all existing records and collections to
  the new framework version (impacts every metadata
  record and CB in DLESE)
• Provide browse capability

12
XML schema methodPros Cons

• CB must know XML
• Searching yields very good results
• Good browse capability
• CB has some control and responsibility
• Vocab can be integrated into DLESE systems or
  services

• Not simple extensive time and work is need from
  DPC and CB
• Requires metadata framework changes
• Impacts other CBs (records and vocabs)
• CB does not have complete control or
  responsibilty

13
Probing questions

• Will users actually benefit from such vocabs?
• Will programming, web devel, metadata devel
  overhead outweigh benefits of vocab?
• Technical challenges exist in implementing
• Who will update terminologies when CB is done
  with the project?
• Searching by CB vocabs creates user interface
  issues, i.e. multiple custom interfaces
• How will DLESE system keep up?

14
Recommendation for DLESE

• Use the Keyword method because
• Simple for both CB and DLESE
• CB has complete control
• Searching (no browse) is mostly likely just as
  good as without the vocabs because
• Most library users search using the keyword text
  box not the controlled vocab menus
• Stemming proximity help improve search results

15
Complications of combining multiple controlled
vocabs (CV)
Then how do we know what to use in the browse page
CV 3
CV 1
CV 2
CV 1
CV 3
CV 2