SUPPORTING COURSE SEQUENCING IN A DIGITAL

LIBRARY

Usage of Dynamic Metadata for Learning Objects

Raul Morales Salcedo and Yano Yoneo

Information Science and Intelligent Systems Department, Tokushima University, 2-1 Minamijosanjima-Cho Tokushima

770-8502, Japan

Keywords: Course sequencing, space-administration, digital libraries, virtual spaces, reusability and customizability.

Abstract: The production of interactive multimedia content is in most cases an expensive task in terms of time and

cost. Hence, optimizing production by exploiting the reusability of interactive multimedia elements is

mandatory. Reusability can be triggered by a combination of reusable multimedia components and the

appropriate use of metadata to control the components as well as their combination. In the same way, digital

libraries comprise vast digital repositories, a wide range of services, and user’s environments and interfaces,

all intended to support learning and collaborative research activities. In this article, we discuss the

reusability and adaptability aspects of interactive multimedia content in a digital library’s learning

environment. We extend a component-based architecture to build interactive multimedia visualization

within digital library’s learning environment with the use of metadata for reusability and customizability

1 INTRODUCTION

The rapid advancement in computer

communications and presentation technologies

produce new forms of media communications than

can be used to increase the quality of educational

documents for visualizing complex technical

problems. To help student learn difficult concepts,

interactive learning software needs specific

capabilities for communication, administration,

visualization and real-time data collection, as well

tools for analyzing, visualizing and sharing digital

data. Such interactive and dynamic digital data

becomes part of a vast digital repository and also

part of the wide range services that the digital

libraries offer, all intended to support learning and

collaborative research activities. These services and

digital data have to be flexibly combined in many

kinds of contexts: diverse classrooms presentations,

tutorials, virtual spaces and standardized

assessments. To archive this goal, the

standardization of so-called learning objects

becomes an important issue.

As stated in the specification of the IEEE’s Learning

Objects Metadata (LOM) (Roschelle et al. 1999), “a

learning object is defined as any entity, digital or

non-digital, which can be used, re-used or referenced

during technology-supported learning”. Examples of

learning objects include multimedia content,

instructional content, instructional software and,

software tools referenced during technology-

supported learning. In a wider sense, learning

objects could even include learning objectives,

persons, libraries, universities, organizations or

events. A learning object is not necessarily a digital

object; however, the remainder of this article will

focus on learning objects that are stored in a Digital

Library.

IEEE’s learning object (LO) model is characterized

by the belief that independent chunks of educational

content can be created to provide an educational

experience. This approach assumes that these chunks

are self-contained, though they may contain

references to other objects and maybe combined or

sequenced to form longer (larger, complex, other)

educational units. These chunks of educational

content may be of any type, interactive (e.g.

videoconference) or passive (e.g. simple video), and

they may be in any format or media type.

Another requirement for learning objects is related

to tagging and metadata. To be able to use such

objects in an intelligent fashion, they must be

labeled as to what they contain, what they

communicate and, what requirements with regard

319

Morales Salcedo R. and Yoneo Y. (2004).

SUPPORTING COURSE SEQUENCING IN A DIGITAL LIBRARY - Usage of Dynamic Metadata for Learning Objects.

In Proceedings of the Sixth International Conference on Enterprise Information Systems, pages 319-324

DOI: 10.5220/0002626903190324

 SciTePress

the use exist. Hence, a reliable and valid scheme for

tagging learning objects is necessary.

The LO model provides a framework for exchange

of learning objects between systems. If LOs are

represented in an independent way, conforming

instructional systems can deliver and manage them.

These efforts gained leverage from the rise of

interactive web technology and its associated

emphasis on standards-based interoperability.

Although the component-based solutions developed

to date are useful, they are inadequate for those

building component-based interactive learning

environments in which the components must

respond to the meaning of the content as well as its

form and presentation. We see the development of

techniques for sharing semantics across components

and applications as a critical research direction for

the field.

The approach described in this paper addresses in

the issue of developing and customizing dynamic

multimedia objects within personal and group spaces

using dynamic metadata.

2 CONTEXT

To explain our starting point and to communicate

the motivation for our work, we first present an

overview of LeComm project.

2.1 The LeComm System

The LeComm project (Morales, 2002) currently

under development, is a learner-centered Digital

Library CSCL environment, where the learners have

the advantages of using the integration of databases

and search functions within a personal or group

space the Digital Library provides. These virtual

spaces provide learners the capacity to arrange and

structure the digital material to suit their own needs

and preferences. It is possible, of course, to use all

material freely available on the net or in a Digital

Libraries in different contexts. LeComm effectively

utilizes and manages such digital material; and also

provides activities like copying, transforming,

indexing, storing, and keeping references in an

appropriated virtual learning environment.

In the LeComm project, we have defined knowledge

base as Web-based software tools that enable access

of valuable information that is organized in a

systematic and pre-designed manner in the

distributed Digital Library’s databases. Therefore,

individually structured knowledge bases are

provided in the form of personal and group spaces.

The learners are then able to establish their own

learning environment, in which they could, for

example join and link documents from different

courses and different digital libraries or places on

the Internet to match their own learning path and

knowledge level.

The LeComm’s virtual space’s success depends

critically on a successful knowledge management.

Knowledge assets are the knowledge that the

LeComm’s virtual space owns or needs to own to

archive its goals. Knowledge equals information,

extracted filtered or formatted in some way.

In LeComm project, knowledge can be divided into

two types: “tacit virtual space knowledge” and

“explicit media space knowledge”. Tacit virtual

space knowledge consists of the hands-on skills, best

practices, special know-how, heuristic, ontology,

intuitions, and so on. The transfer of tacit virtual

space knowledge is by tradition and shared

experience, through for example, apprenticeship, job

training or expertise. Explicit media space

knowledge is used in the design of routines, standard

operations procedures, and the structure of data

records. Explicit media space knowledge enables

the Digital Library to enjoy a certain level of

operational efficiency and control. Those forms of

knowledge can be found in any Digital Library.

The LeComm learning environment however, is

continuously expanding, renewing, and refreshing its

knowledge in all categories. The role of the

LeComm’s knowledge is to promote the learning of

tacit virtual space knowledge to increase the skill

and creative capacity of its learners and takes

advantage of explicit media space knowledge to

maximize the learning efficiency. In effect The

LeComm’s learning environment has acquired a

third class of knowledge - meta-knowledge - that it

uses to create and integrate specific lessons tailored

to a targeted Digital Library’s group with all its

intellectual resources in order to achieve high levels

of learning. These lessons are created using a

knowledge base of multimedia elements within the

Digital Library. These lessons are created

automatically by using learner’s preferences and

style (course sequencing).

LeComm architecture includes part of the

knowledge base of the Digital Library which is

necessary to implement the course sequencing,

consist of two separated knowledge spaces. The

concept “tacit virtual space” contains a networked

model of learning topics (Fischer, 2000) and uses

well-known approaches from knowledge

management. The media bricks stored in the Digital

Library’s “explicit media space” are atomic

information units in various formats. These units are

interconnected via rhetorical relations and, each

media brick is described using IEEE’s Learning

Objects Metadata (LOM) scheme. In the following

sections we refer to media bricks as “learning

ICEIS 2004 - HUMAN-COMPUTER INTERACTION

320

objects”. Although both information spaces are

separate, each learning object can have a relation to

one or more related topics. The separation of both

spaces is the way in which LeComm generates

adaptive lessons, because a set of media bricks

(documents, magazines, books, articles, papers,

graphics, notes, audio, video, etc.) for each topic is

available. Thus, the selection of media bricks is

determined by each learner’s preferences.

The general functionality of LeComm, in others

words the generation of lessons, depends on the

knowledge base stored in the “tacit virtual space”. -

This approach is similar to the standardization by

IEEE, in that the IEEE proposes the use of

knowledge library responsible for sequencing a

lesson, while the actual compilation of the lesson is

done by a delivery component (see Figure 1).

Figure 1: IEEE-LTSC adapted architecture

To understand the automatic creation of exercises, it

is essential to understand the setup of our meta-

knowledge base.

The following order can be specified; taking into

account the way that a teacher actually sets up a

multimedia lesson:

The teacher acquires background knowledge from

the Digital Library.

The teacher creates an outline for the multimedia

lesson in his/her personal space.

The teacher fills the lesson’s outline with the Digital

Library content and, shares the virtual space.

These steps are modeled by different spaces in

LeComm learning environment. The mechanisms

for automatic creation of exercises are given as

follows: LeComm system sends; receives and

“negotiates” learning preferences and styles with the

learner client.

LeComm system receives current assessment

information, preferences, and performance

information (history and objectives) for future

learning experiences; It makes queries to the

knowledge library and search for the appropriate

material; the knowledge library returns the catalog

info in the form of content index metadata as found

learning content, extracts the locators index from the

returned catalog info in the form of content index

metadata; and makes a choice (a lesson plan) by

invoking learning content. And finally, locators

index are sent to delivery process to identify but not

transfer learning content.

Here is very important to note that the delivery

process is not responsible for retrieving learning

content, and the knowledge library is responsible for

transferring learning content.

The tacit virtual space contains ontology in terms of

keywords, which is necessary for creating the

outline of a lesson. After sequencing the outline

(equally applicable to the creation of a table of

contents), the real content like documents,

magazines, books, articles, papers, graphics, notes,

videos, their complements, and their links) are filled

into the outline using elements of the second space,

the Digital Library’s explicit media space.

The general idea of LeComm in this point is that is

necessary to employ different relations within the

“tacit virtual space” and the “explicit media space”

to model the different goals in both spaces.

When working with media bricks and the necessary

educational metadata, an important disadvantage

becomes obvious. Due to the way in which the

history of metadata developed, static resources such

images or text documents can be described properly;

but an appropriate description of dynamic resources,

for example multimedia objects like video and

audio, are feasible only to a limited extend. The

reason is that dynamic multimedia objects can

process input parameters, generate output

parameters, and also work internally with data that

cannot be described with traditional metadata

schemes.

2.2 Multimedia Content

Learning systems enriched with multimedia

elements can be divided into two categories:

Learning objects are relatively simple, but are

described by metadata in detail. A learning system

operates on the metadata with intelligence.

Learning objects are very smart in that they change

their behavior. A learning system has to pass on

specific information, and each learning object has to

adhere to a specific stipulated set of inputs and

outputs parameters.

An example of the first category is the use of IEEE’s

LOM in the LeComm project to describe multimedia

content (see Figure 2). However, multimedia

content as part of a learning system can be text,

graphics, audio, video and, Digital library’s services.

In LeComm learning environment, we can identify

learning objects as follows:

Multicode: Use of various symbols, for example,

images, pictographs, texts.

Multimode: LOs make use of text, images and

continuous media like video and audio.

SUPPORTING COURSE SEQUENCING IN A DIGITAL LIBRARY: USAGE OF DYNAMIC METADATA FOR

LEARNING OBJECTS

321

Dynamic: LOs realize to some extend the interaction

between learner and LeComm.

Interactive: LOs can address various senses: visual,

aural, or both ate the same time.

Figure 2: LeComm’s lesson architecture

Figure 3 illustrates the above characteristics and

their relationships. For example, an active map

object belongs in the intersection between multicode

and multimode. The common denominator of all

these characteristics is what we refer to as “smart

learning object” (SLO). Digital material such as

audio and video that visualize complex procedures

dynamically and interactively belong to the group of

SLOs. Animated graphics, videos and audio are

much closer to real life than still flat documents or

graphics. Complex procedures can be experienced,

understood and, learned by experimenting in the

virtual environment offered by LeComm learning

environment.

Figure 3: Learning object’s characteristics

The behavior of smart learning objects can be

changed, as well as adapted, according to parameters

passed by the system. In the reminder of this article,

we detonate interactive visualizations as “SLOs”.

One of the key problems in developing e-learning

infrastructures in general and interactive

instructional visualization units in particular is the

integration of learner requirements that change over

time. Learning systems must be flexible to adapt

easily to new and changing user requirements.

2.3 Metadata for Dynamic Learning

Objects

The existing technologies, standards and, ongoing

initiatives for multimedia educational metadata are

the starting point for our research. The Alliance of

remote instructional authoring and distribution

networks for Europe (ARIADNE, 2001), Dublin

Core metadata element set (Dublin Core, 2002),

Educom’s instructional management system (IMS)

(Educom, 2002) and, the IEEE’s learning object

metadata working group 12 are the most important

initiatives dealing with metadata for computerized

learning. These initiatives are closely related to the

resource description framework (RDF, 2003), the

warwick framework (Warwick, 2001), and to other

activities of the world wide web consortium.

All the methods specifying metadata make use of it

in the traditional sense of describing static data:

To summarize the meaning of the data. (what the

data is about);

To allow learners to search for data;

To allow learners to determine if the data is what

they want;

To prevent some learners (children) from accessing

data, and;

To instruct us on how to interpret the data (format,

encoding, encryption).

That is, the metadata descriptors are associated with

the data sets in a fixed way. Their granularity is

defined by the original metadata author.

A great drawback is that the application of metadata

is mainly limited to content. Our first observation is

that such metadata cannot adequately describe smart

dynamic LOs. Metadata cannot influence the

multimedia content itself, because metadata usually

contain universal and widely applicable descriptions

of objects. From our point of view, the use of

dynamic multimedia LOs such audio and video

objects require a new sort of metadata, which must

be dynamic in order to facilitate the I/O behavior of

a dynamic LO.

3 LEARNING EXAMPLE

Figure 4 shows the overall architecture of

LeComm’s SLOs tagging and customizing

architecture. Digital Library’s learning resources are

tagged using LeComm’s personal space (see section

3.1). For the storage of static and dynamic metadata,

we use a relational database (see Figure 2). To

access the data stored in the Digital Library, we

developed a three-tier architecture using JDBC.

ICEIS 2004 - HUMAN-COMPUTER INTERACTION

322

LeComm’s learning environment allows modify

interactive visualizations with the use of dynamic

metadata. This “customizer” is used to modify smart

learning objects for a lesson; we are then able to use

these visualizations several times in the personal or

group spaces, depending on the context, which is

described in detail in Section 3.2. In Figure 4, smart

learning objects are reused in different scenarios

with different metadata sets to show various

scenarios of the same topic.

Figure 4: SLO tagging and customization process

3.1 Personal and Group Spaces

LeComm includes a concept that we call personal

and group spaces, these areas are generated owned

and maintained by learner persistently keep resource

objects, or references to resources which are relevant

to a task or set of tasks the learner needs to perform

in the learning processes (Morales, 2000). In the

following we describe the creation of static and

dynamic metadata within personal and group spaces.

These virtual spaces can also be used to publish

metadata records for various Digital Library

resources, e.g. documents, images; audio clips, video

clips and interactive exercises. Figure 5 shows the

relationship between the Digital Library's explicit

media space and the LeComm's tacit virtual

environment using metadata.

A metadata record consists of a set of elements that

describes a multimedia resource. Examples of these

elements are: creation date, type, author, format, title

and publication of the resource.

To enable easy access and discovery of multimedia

information resources, LeComm provides

mechanisms to store and create LOM records based

on the IEEE-LOM scheme version 4.0 in a relational

database and can also be used to search the Digital

Library’s databases and to navigate the resulting

metadata set. While working with personal and

group spaces and using the base LOM scheme,

LeComm quickly turned out that SLO can be

described to only a limited extend. We extended this

feature by adding an extra category showed in the

(Table 1) for dynamic metadata that is not included

in IEEE-LOM 4.0 scheme. This extra category

includes specific parameter configuration of

visualization; used to adapt the content of and object

and/or to change the behavior of a learning object.

Tagging the Digital Library’s source material with

LeComm’s personal and group spaces turned out to

be an interesting experience. Most elements of a

lesson apply the same basic metadata such as the

author’s name, copyright and, the targeted user

group within the Digital Library. So it would be

useful to have a set of templates to tag the Digital

Library’s material. With templates we can avoid

filling a lot of fields over and over. For example, the

owner fields, library’s classification, browser

requirements, object’s representation and many

others. In out current prototypical implementation,

templates are used to stored information, which is

then only typed once but can be applied multiples

times.

Figure 5: Publishing metada records in LeComm’s

personal and group space

Table 1: SLO proposed dynamic metadata

SUPPORTING COURSE SEQUENCING IN A DIGITAL LIBRARY: USAGE OF DYNAMIC METADATA FOR

LEARNING OBJECTS

323

3.2 Usage Example

To be able to reuse the same content in different

scenarios that is the basic functionality that enables

dynamic metadata to operate; the LeComm learning

environment requires that both LOM and SLO

proposed dynamic metadata schema exist

concurrently, as they are both involved in the

learning process. Indeed, whereas in the explicit

media space’s LOM receives, processes and sends

the original request negotiation to the LeComm, the

LeComm’s tacit virtual space gets the actual

metadata to its extended metadata schema and

creates the meta-knowledge that we called SLO.

Here the SLO’s description plays an important role

due that each SLO contain all necessary information

for a specific lesson in different scenarios.

4 CONCLUSIONS AND FUTURE

WORK

In this article we discuss the reusability aspects of

multimedia content in a Digital Library’s web based

e-learning system. We highlight the necessity for

developing component-based interactive multimedia

visualization units within personal and group spaces.

We suggest the use of metadata for reusability

issues. The main contribution of this article is an

extension a component-based architecture (IEEE’s

learning objects metadata) to enable it to describe

dynamic multimedia learning objects in a Digital

Library, which we refer to as meta-knowledge -

“smart learning objects” -. Traditional metadata that

describe learning objects are well suited for

describing static elements (text and images), but do

not take the dynamic nature of multimedia element

into account (specially audio and video objects).

Hence, we compare various learning metadata

standards and derive an extension to solve the

problem.

The LeComm’s metadata-based framework in our

article also addresses the customization of smart

learning objects by metadata. Having explained

dynamic metadata, we described our implementation

in the LeComm’s personal and group space for

tagging, storing and, customizing smart learning

objects. As a Digital Library prototype, we

implemented visualizations artifacts dealing with the

teaching of the english language in the supply chain

management area in an specific company provider

explanation. We currently use our framework to

develop other language teaching lessons within our

Digital Library, for example, Japanese, English and

Spanish multimedia lessons to explain specific

problems in the area of computer sciences.

The research described here differs from other

related work, in that the set of dynamic metadata

items in the Digital Library’s personal and group

spaces can be defined for smart learning objects is

open-ended and not fully predefined. When the

learners add new metadata or schemas, the changes

are automatically reflected through the LeComm

learning system. Predefining all attributes would

hinder the support of multiple applications as we

already mentioned. Instead, learners are allowed to

create whatever metadata records are necessary to

support the object customization process, together

with a necessary base set of predefined parameters

in the virtual space describe dynamic resources.

REFERENCES

Roschelle, J.L., Digiano , J., Koutlis, M., Repenning, A.,

Phillips, J., Jackiw, N., and Suthers, D. 1999.

Developing educational software components. IEEE

Computer 32,9 and 50-58.

Morales, S, R. and Yoneo, Y., 2002. Effective

collaboration in a videoconference-based CSCL

environment. In ED-MEDIA, World Conference on

Education Multimedia, Hypermedia &

Telecommunications. Vol.2002, Issue:1, pp.1344-

1370.

Fischer, S. and Steinmetz, R., 2000. Automatic creation of

exercises in adaptive hypermedia learning systems. In

Proc. of the tenth ACM conference on Hypertext.

ACM Press, New York, NY.

Alliance of remote instructional authoring and distribution

networks for Europe. Available at:

http://www.ariadne-eu.org

Provisional report of the Dublin Core sub-element

working group. Available at:

http://purl.oclc.org/metadata/dublin_core

Educom: instructional management system project.

Available at: http://www.imsproject.org/

Resource description framework. Available at:

http://www.w3.org/RDF

The warwick framework: A container architecture for

diverse sets of metadata. D-Lib magazine. Available

at: http://www.dlib.org

Morales, S.R., 2000. VicFDL: extending communication

channels in DL through videoconference. Master

Thesis. Department of Computer Systems

Engineering. Universidad de las Americas-Puebla,

Mexico.

NFS: Internet2 project. http://www.internet2.org

ICEIS 2004 - HUMAN-COMPUTER INTERACTION

324