Learning object metadata: a wolf in sheeps clothing

1
Learning object metadataa wolf in sheeps
clothing?

• Jane Barton John RobertsonCentre for Digital
  Library ResearchUniversity of Strathclyde

2
Overview

• Its just a sheepLO metadata and its origins
• The librarians cry wolfidentifying the LO
  metadata problem
• Okay, its a wolfscoping the LO metadata
  problem
• Taming the wolfthe Strathclyde Metadata
  Workflow Study
• metadata quality workflow
• factors influencing quality workflow
• issues to resolve

3
Its just a sheep

• Standardised approach to metadata integral to the
  development of the learning object economy
• international standard
• application profiles, best practice guidelines,
  etc
• But
• The authoring of metadata itself will be
  straightforward... Because metadata filesare
  machine-writable, authors will simplyaccess a
  form into which they enter theappropriate
  metadata information.
• (Downes, 2001)

4
The librarians cry wolf

• Metadata quality
• errors, omissions and ambiguities result in
  problems with recall and precision
• essential for functionality, interoperability and
  meaningful participation in aggregated services
• Metadata creation is not a trivial exercise
• Creation of good quality metadata requiresa
  community-wide approach

5
The librarians cry wolf

• Even when theres a positive benefit to creating
  good metadata, people steadfastly refuse to
  exercise care and diligence in their metadata
  creation. Take eBay every seller there has a
  damned good reason for double-checking their
  listings for typos and misspellings. Try
  searching for plam on eBay. Right now, that
  turns up nine typoed listings for Plam Pilots.
  Misspelled listings dont show up in correctly
  spelled searches and hence garner fewer bids and
  lower sale-prices. You can almost always get a
  bargain on a Plam Pilot at eBay.
• (Doctorow, 2002)

6
The librarians cry wolf

• Metadata quality
• errors, omissions and ambiguities result in
  problems with recall and precision
• essential for functionality, interoperability and
  meaningful participation in aggregated services
• Metadata creation is not a trivial exercise
• Creation of good quality metadata requiresa
  community-wide approach

7
Okay, its a wolf

• Useful case studies from within the community
• mature repositories experiencing precision and
  recall problems
• non-specialists recognising need for education
  and training
• new repositories looking for advice and guidance
• some recruitment of librarians
• Better awareness and appreciation of thenature
  of the beast and the size of its teeth

8
but its a special wolf

• Collaborative metadata creation model with
  limited specialist input
• Resources not formally published and often highly
  transient
• Concept of reuse and repurposing, with associated
  digital rights issues
• Complications arising from content packaging

9
Taming the wolf

• The Strathclyde Metadata Workflow Study
• supporting study for X4L and FAIR programmes
• focus on metadata creation within UK LORs and IRs
• arises from issues identified by early
  implementers
• Primary aim was to develop a good understanding
  of
• the factors which influence both metadata
  qualityand the workflows used to create it
• the context within which recommendationsare to
  be made

10
Defining metadata quality

• Quality is fitness for purpose
• Aspects of metadata quality
• structure, semantics and syntax
• metadata metrics
• Metadata requirements
• local needs
• community-level requirements
• cost-benefit and compromise

11
Managing metadata quality

• Metadata quality in practice
• contextual factors determine actual quality
• immediate and long-term priorities
• The quality cycle
• Quality assurance quality by design
• elements, records, repositories, aggregations
• tools and techniques
• embedding quality assurance in metadatacreation
  processes - workflow

12
Metadata workflow why bother?

• Metadata for libraries
• well-defined purpose, stable context
• workflow embedded and tightly controlled
• community-wide approach to optimise quality
• Metadata for repositories
• poorly defined and diverse purpose, evolving
  context
• workflow distributed, collaborative or both
• potential of community-wide approach notfully
  exploited

13
Designing metadata workflow

• Determine purpose of metadata
• Determine level of quality required
• Determine level of quality that can be achieved
• Design and implement workflow
• Refine workflow
• Review
• Opportunities for refinement
• arise from both local and wider context
• produce both direct and indirectimprovements in
  quality

14
Factors influencing metadata quality

• What is the repository for, locally and within
  the wider context? Does this give rise to any
  conflicts?
• What type of objects will the repository contain?
  How will they be used? And by whom?
• What functionality is required locally? How will
  it be interfaced? What entry points will be used?
• What is required for interoperability? Are
  requirements formal or informal, direct or
  indirect?
• Will access restrictions be imposed locally?And
  in the wider context?
• Will metadata be meaningful withinaggregations
  of various kinds?

15
Factors influencing workflow design

• What resources are available locally? Who will be
  involved? What skills do they have?
• How can these resources be used to best effect?
• Are resources available within the wider
  community? Does their use require compromises to
  be made, and if so, is it worth it?
• Are resources sufficient to produce the required
  metadata quality, and if not, what are the
  priorities?
• What level of commitment exists locallyand in
  the wider community?

16
A typology of factors

• Repository-level factors
• Does the repository have a subject specialism? Is
  it required to interoperate with the local VLE?
• Object-level factors
• Can objects be repurposed? Is their use
  restricted?
• Metadata-level factors
• Does participation in the wider community impose
  specific requirements? Is training available?
• Local factors
• Is there a strategic commitment to the repository?

17
Issues to resolve

• Repository-level issues
• need a better understanding of what an individual
  repository is for, both locally and as part of a
  wider system of repositories and services
• Object-level issues
• need a better understanding of how objects are
  created, used, repurposed, managed
• Metadata-level issues
• need to integrate and optimise metadatacreation,
  enhancement and QA processes throughout the wider
  system

18
Developing the metadata lifecycle model

• derived from scenarios based on observation of
  existing metadata creation/aggregation processes
• how metadata moves,
• multiple instances, versions or formats
• various transformations at different locations
• interactions between different metadata records
  for the same object or related objects
• implications of movements transformations
• include community-level metadata creationand QA
  activities
• development of vocabularies
• design and testing of tools