HILT II: Towards Interoperable Subject Descriptions

1
HILT II Towards Interoperable Subject
Descriptions

• Report to the JISC Terminologies Workshop,
  February 2004.
• Dennis Nicholson
• Centre for Digital Library Research,
  Strathclyde University

2
Overview

• Aims
• Overview of the Problem
• Outcomes
• Server Design Summary
• Function Elements Coverage Use
• Other Points to Note Costs
• Two Alternatives

3
Aims

• Build pilot terminologies service for JISC
  Information Environment, aiming to
• Provide a practical experimental focus within
  which to investigate and establish subject
  terminology service requirements for the JISC I.E
• Make recommendations as regards a possible future
  service

4
Overview of the Problem

• JISC other services JISC users need
• Different schemes in use
• Different versions of the same scheme
• Different approaches to amending and extending
  schemes
• Schemes, variations vary in ability to reflect
  terminologies used by users when searching
• Users need to
• Identify services appropriate to their search
• Identify relevant items either by using correct
  term for scheme used or by some other
  method

5
Outcomes

• Development project to build a working
  terminologies server with specific features
• Consensus, Collaboration seen as vital, so
• Dialogue with key national and international
  players
• Mapping between schemes, rather than preference
  for a single scheme (HILT I)
• The need for a facility to allow others to
  include their own (self-provisioned) mappings
• The existence of other terminology servers

6
Server Function

• To
• Improve accurate, consistent description by staff
• Improve accurate, informed searching by users
• Map between schemes via DDC spine
• Map user terms to DDC, collections, other schemes
  
• Ultimately improve retrieval from legacy
  metadata
• Monitor, learn from user terminology sets

7
Server Elements

• Wordmap three elements
• Database (Oracle) of terminology mappings
• User front end that interacts with database
  according to staff specifications and user input
  and feedback
• Drag and drop, multi-user interface to support
  sophisticated staff interaction with database for
  creation and maintenance of maps, inter-service
  co-ordination and training

8
Server Schemes

• DDC spine captions relative index standard
  subdivisions
• LCSH to DDC mapping
• UNESCO to DDC mapping
• UK terms registry (mapped to DDC)
• MeSH, Regional, AAT options
• Pilot more limited UNESCO, MeSH illustrative,
  no AAT

9
The Server in Action

• User enters subject term
• Term matched to terminology set and mapped to DDC
• Options and disambiguation
• DDC number truncated and mapped to collections
  database to identify appropriate collections
• Information on scheme used advice sample
  retrieval
• Demonstration

10
Other Points Costs

• User Interface Facilities and Further Research
• Machine to Machine (M2M) Facilities and
  Interactivity Issues
• Limited granularity mapping
• Information Environment Services Registry (IESR)
  Issues
• Costs over 5 years

11
Two Alternatives

• A single scheme? (but)
• Automatic Categorisation Matrix?
• Auto-categorising service for every scheme
• Staff auto index and classify resources but
  correct intellectually both are recorded
  collection and item identifier also recorded
• User queries or seed documents also
  auto-categorised against each scheme mapped to
  items using either auto or intellectual
  categorisation user disambiguates at item level

12
Thoughts

• Automatic Categorisation Matrix
• If it worked, it would remove the need for
  expensive mapping process
• Interim project that investigates this approach
  worth considering first?
• Speculative needs researched
• One possible breakout session topic
• Mapping may be the only way
• Doing nothing not an option

13
Further Information

• Website http//hilt.cdlr.strath.ac.uk/
• e-mail
• d.m.nicholson_at_strath.ac.uk
• Ali.shiri_at_strath.ac.uk
• emma.mcculloch_at_strath.ac.uk