Csun pse-006-presentation-2013 v2.1

IEEE LOM is not an option:
lessons to learn

Miquel Centelles, Mireia Ribera, Marina Salse
Ensenyament – Assignatura
Grup Adaptabit: Working group on digital accessibility
for teaching, research – 20xx
Curs 20xx and teaching innovation
Docent: Nom Cognoms
Departament of Librarianship and Information Science

University of Barcelona

Summary

 Rationale
 Objectives
 Methodology
 Data analysis
 Discussion
 Further steps

2/27/2013 28th Annual International Technology and
Persons with Disabilities Conference 2

Rationale:
context of the research
 A project on creating accessible
teaching resources within the University
of Barcelona.
 We want to recommend teachers a
metadata model covering accessibility
aspects of resources and processes.


Rationale:
why our (first) interest in IEEE LOM
 Adoption of SCORM: Many Learning
Management Systems support
SCORM, and SCORM uses IEEE
LOM metadata.
 Adoption in LMS: LOM as a major
development of eLearning systems
(such as LMS) and is widely used in
such systems, notably for example in
Europe.
 Adoption of profiles: LOM has been
widely profiled for particular domains.

28th Annual International Technology and
2/27/2013 Persons with Disabilities Conference 4

Rationale:
why our (first) interest in IEEE LOM


Rationale:
known IEEE LOM drawbacks
 Its abstract model is not aligned with basic
standards for semantic interoperability, such as
Resource Description Framework (RDF).
 The adaptation of the standard to the web of data
is suffering from delays in two key processes:
 The IEEE LOM mapping to Dublin Core (DCMI)
abstract model.
 The elaboration and publication of an official RDF
vocabulary.


Objectives
of the research
1. Identification of application profiles
based on IEEE LOM.
2. Descriptive review of IEEE LOM
application profiles (AP).
3. Descriptive review of AP implementation
on Learning Resource Repositories
(LRR).


Methodology:
on application profiles
Application profiles gathering:
 Literature search through key
actors, European projects, and bibliographic
databases.
 Complement with questionnaires and
interviews to AP holders.


Methodology:
 Application profiles selection:
 It must be based (mostly) on IEEE LOM, of
course
 It must be currently active
 No restrictions on:
• the practice community
• the scope of application profiles (topics…)
• the country of origin


Methodology:
 Key data of findings:
 32 different application profiles
 3 have a world wide scope
 11 are focused on Europe
 4 are focused on USA
 the remaining 17 are focused on
different, specific countries


Methodology:
on LRRs
 One LRR is selected for each IEEE LOM
application profile:
 It must offer openly accessible resources
 It could belong to one unique institution, or to
several
 If several LRRs, selection based on:
• University over lower studies
• Broad content over specialized


Methodology:
on LRRs
 10 samples of metadata records are obtained from
each LRR:
 Search period: 29th August-8th October 2012.
 Search strategy (descending order):
• 1st Criteria: first learning resources published during 2012
• 2nd Criteria: learning resources of the type “Lecture”
• 3rd Criteria: keyword “education”

Finally, we got search results concerning 24 APs


Methodology:
10 samples of records of each LRR


Data analysis:
2 different purposes

APs versus base standard IEEE LOM

Metadata records versus APs


Data analysis:
different evidence levels
 Not all AP provides the same quantity and quality
of evidences for the analysis.
 All of them: documentation about schema and data
values
 Other evidences, depending on each AP:
• Full evidence level: records in XML binding.
• Medium evidence level: records in some human readable
format (not XML).
• Low evidence level: no metadata records (8 APs), mostly due
to LRR out of order during the test period.


Data analysis:
APs vs. base standard
 # of simple data elements in AP versus 58
total in the base Standard

 # of mandatory simple data elements in AP

 # of non allowed modifications within AP:


Data analysis:


Data analysis:
 Non allowed modifications:
1. Altering the relative location of an existing data
element (e.g. moving a parent element to a child
one)
2. Creating a new element that mimics the semantic
intent of an existing element
3. Changing the meaning of an existing element
4. Changing the name of an element
5. Extending a schema other than at a specified
extension point


Data analysis:
 Non allowed modifications (cont.):
6. Extending cardinality of an element
7. Adding new items in a controlled vocabulary list
8. Modifying the value space and data type of data
elements from the base schema.
9. Defining data types or value spaces for
aggregate data elements in the base schema


Number of simple data elements
included in AP respect base schema
120.00%

100.00%

80.00%

60.00%

40.00%

20.00%

0.00%


Number of simple data elements
included in AP respect base schema

14 APs include less than the 58 elements in
the base standard (44%)
12 APs include all the 58 data elements in
6 APs include more than the 58 elements in


Number of mandatory simple data
elements stated by AP
35

30

25

20

15

10

5

0


Number of mandatory simple data
elements stated by AP

25 APs state mandatory (simple) data elements
(78%):
 At top: Biosci Education Network (BEN) states 30
mandatory elements
 At bottom: LOM-FR states 3 mandatory elements
7 APs don’t state any mandatory (simple) data
elements (22%)


AP is conformant with base schema?


AP is conformant with base schema?

4 APs are fully conformant with the base schema (12%)
25 APs are not fully conformant with the base schema
(78%)
 The less respected restriction: Adding new items in a
controlled vocabulary list (18/25)
 The most respected restriction: Defining data types or value
spaces for aggregate data elements in the base schema
(2/25)
In 3 cases, solid conclusions can not be made based on
available sources (9%)


Data analysis:
Metadata records versus APs
 Our questions are:
 Metadata records respect mandatory
conditions of simple data elements in the AP?
 Metadata records in the LRR apply controlled
vocabularies established by the AP?
 Metadata records in the LRR respect
requirement related to value spaces and data
types in the AP?


LRR follows mandatory conditions?


LRR follows mandatory conditions?

• Not applicable in 5 LRRs (21%)
• Mandatory conditions are followed in 5
LRRs (21%)
• Mandatory conditions are not followed in
11 LRRs (46%)
• In 3 cases, solid conclusions can not be
made based on available sources (12%)


LRR applies specified controlled
vocabulary?


LRR applies specified controlled
vocabularies?

• Specified controlled vocabularies are
applied in 19 LRRs (79%)
• Specified controlled vocabularies are not
applied in 1 LRR (4%)


LRR apply data types and values
restrictions?


LRR applies data types and values
restrictions?

• Data types and values restrictions are
applied in 11 LRRs (46%)
• Data types and values restrictions are not
applied in 1 LRR (4%)


Discussion:
disappointing results
 Most AP are not conformant with IEEE
LOM base standard.
 Implementation of AP on LRR don’t even
follow the application profile conditions.
 Availability of interchange formats
(XML, JSON... not to say RDF) for
metadata records is not a broad practice.


Discussion:
main conformance black holes
 Extension of controlled vocabularies with
new words created adhoc.
 Modifications in value spaces and data
types of data elements.
 Definition of data types or value spaces for
aggregated data elements.


Discussion:
lessons learned
 Keep them simple. Metadata is an
“overhead” task which should be minimum
and as automatic as possible.
 Force conformance through XML
schemas, semantic web vocabularies or
other applied constraints
 Set a standard for the display of records
and their reusability.


Further steps:
new standards in competition
 Based on those drawbacks, we have decided
to move on new alternatives for metadata
base schema:
 ISO/IEC 19788 Metadata for learning resources
(MLR) standard
…or…
 Learning Resources Metadata Initiative
(LRMI), which uses microdata and is led by
significant companies.


Further steps:
we’ll keep monitoring
 Nevertheless, we’ll keep on completing the map
of IEEE LOM based AP.
 In order to:
 Monitoring the evolution and adaptation of IEEE
LOM APs to the semantic web.
 Monitoring the solutions which LRRs adopt to
manage mentioned challenges.
 Monitoring the evolution of IEEE LOM standard in
relation with the raising of “new” learning resources
metadata standards.


Further steps:
collaborations?
 We ask you to give us information about
IEEE LOM APs and LRR using
them, answering the questionnarie for this
purpose available at:
Adaptabit http://bd.ub.edu/adaptabit/

 We will offer you the publication of all the
data about IEEE LOM APs as open data.


Thanks for your attention!

Questions,
Opinions,
Suggestions…
miquel.centelles@ub.edu


Csun pse-006-presentation-2013 v2.1

More Related Content

Viewers also liked (10)

Similar to Csun pse-006-presentation-2013 v2.1 (20)

More from adaptabit (11)

Recently uploaded (20)

Csun pse-006-presentation-2013 v2.1