8/13/2019 Hg Thesis

1/128

Designing a Reusable and Adaptive E-Learning System

A Thesis Submitted to the College of

Graduate Studies and Research

in Partial Fulfillment of the Requirements

for the Degree of Master of Science

in the

Department of Computer Science

University of Saskatchewan

Saskatoon

By

Honggang Wu

November, 2002

Copyright Honggang Wu, 2002. All rights reserved.

8/13/2019 Hg Thesis

2/128

Abstract

The Internet has shown its advantage in e-Learning, a new approach to online education.

It enables online learning materials to be reused by any educational organization around

the world, when there is an efficient way to find the appropriate learning materials on the

Internet and combine them together. The new concept of e-Learning is that by organizing

and disseminating the learning content into a uniform format as small chunks of learning

materials, it is possible to achieve content reuse and interoperation between different

educational institutions and training vendors.

This thesis addresses the possibility of developing learning content based on learning

objects, and evaluates the role played by XML in designing an e-Learning system. The

main goal of this research is to design a reusable and adaptive e-Learning system with

XML schemas forming the basic framework. To implement this goal, the concept of a

learning object is redefined, and a schema for learning objects is developed. Some

modifications to the existing e-Learning specifications, such as the metadata specification,

the content packaging specification, and the learner information specification, are

introduced to make them more suitable for our e-Learning systems. In this thesis, four

online tutorial courses are developed to illustrate how these schemas can work together to

make it possible to reuse learning objects in different learning contexts and to provide

learners with individually tailored learning content.

8/13/2019 Hg Thesis

3/128

Table of Contents

1 Introduction 1

2 Background 4

2.1 E-Learning, a New Way of Learning 4

2.1.1 The Features of E-Learning 52.1.2 A New Approach to E-Learning Content Development 7

2.2 Learning Objects 82.2.1 Background Literature on Learning Objects 8

2.2.2 Reusability of Learning Objects 10

2.2.3 Problems with Learning Objects 112.3 E-Learning and XML 13

2.3.1 XML Schema and its Benefits for Interoperability and Reusability ofInformation 14

2.3.2 XSL, XSLT, and Xpath for Flexible Information Presentation 172.3.3 XML DOM and Java API 182.3.4 Conclusion about Using XML 18

2.4 E-Learning Standards 192.4.1 Standards Initiatives For E-Learning 20

2.4.2 Specification for Metadata 23

2.4.3 Specification for Content Packaging or Course Structure Format 252.4.4 Specifications for Question & Test Interoperability Specification 27

2.5 Summary 27

3 Schema Design 28

3.1 Schema for Learning Objects 303.1.1 New Definition of Learning Objects 30

3.1.2 The Design of a Schema for a Learning Object 343.2 Schema for Metadata 38

3.2.1 The Design of Schema 38

3.2.1 Controlled Vocabulary 41

3.3 Schema for Content Packaging 423.4 Schema for Learner Information 46

4 System Design 49

4.1 Architecture of the E-Learning System 49

4.2 Learning Content Development 524.2.1 Idea and Principles behind Learning Content Development 52

4.2.2 Developing Learning Content 544.3 Design of the Learning Management System 64

4.3.1 Learning Management System 64

4.3.2 Design of the LMS 654.3.2.1 Learning Management Module 66

4.3.2.2 Sequencing/Tracking Module 67

8/13/2019 Hg Thesis

4/128

4.3.2.3 Testing/Assessment Module 68

4.3.2.4 Search Module 684.3.2.5 Learner Profile Module 68

4.3.2.6 Delivery Module 694.4 Summary 69

5 Implementation of the E-Learning System 70

5.1 Learning Content Implementation Details 70

5.2 Implementation of the Learning Management System 755.2.1 Environment of the Learning Management System 75

5.2.2 The Basic Components of the Learning Management System 76

5.2.3 Implementation of the Modules 775.2.3.1 Implementation of the Learning Management Module 77

5.2.3.2 Implementation of the Delivery Module 785.2.3.3 Implementation of the Tracking/Sequencing Module 79

5.2.3.4 Implementation of the Assessment/Testing Module 82

5.2.3.5 Implementation of the Learner Profile Module 835.2.3.6 Implementation of the Searching Module 845.3 System Testing 855.4 Conclusion 88

6 Conclusions 89

6.1 Summary of Thesis Work 896.2 Research Contributions 906.3 Future Work 92

6.3.1 Improving Search Functionality 92

6.3.2 Generating Learning Objects from Databases or HTML files 93

6.3.3 Displaying Learning Objects with Different Formats 936.3.4 Expanding Content Packaging Schema and Learner Information Schema 94

6.3.5 Developing Tools for Authoring Learning Objects and Content PackagingFiles 94

6.4 Conclusion 94

References 95

Appendix A-Schema for Learning Objects 101

Appendix B-Schema for Content Packaging 110

Appendix C-Schema for Metadata 118

Appendix D-Schema for Learner Information 121

8/13/2019 Hg Thesis

5/128

Figures

Figure 2.1 Content Packaging 26

Figure 3.1: Structure of Element 34

Figure 3.2: Structure of Element 36Figure 3.3 Importation of QTILite Schema 37Figure 3.4 Structure of Metadata 39

Figure 3.5 Structure of Content Packaging 43Figure 3.6 Sample Course Structure 44

Figure 3.7: Structure of Element 45

Figure 3.8 Structure of Element 46Figure 3.7: Structure of Learning Information 47

Figure 4.1 Architecture of the E-Learning System 50Figure 4.2 Course Structure for Encryption Tutorial 51

Figure 4.3 Sample Element 57

Figure 4.4 Sample Element 58Figure 4.5 Prerequisite for a Quiz 59

Figure 4.6 Difficulty Range for Learning Content 59Figure 4.7 Learning Object RSA.xml 61

Figure 4.8 Learning Object RSA1.xml 62

Figure 4.9 Sample Metadata 63Figure 4.10 LMS Modules 65

Figure 5.1 Course Structure for Encryption 71Figure 5.2 Course Structure for Firewall 72

Figure 5.3 Course Structure for Network Security 73

Figure 5.4 Course Structure for E-Commerce 74

Figure 5.5 Sample Learner Profile Information 75Figure 5.6 XSL File for Learning Objects 79Figure5.7 System Interface 80

Figure 5.8 Learning Content 81

Figure 5.9 Quiz 82Figure 5.10 Records of Learning Activity 83

Figure 5.11 Changing Learner Information 84Figure 5.12 Searching Results 85

Figure 5.13 Record of Learning Paths for the Encryption Tutorial 87

8/13/2019 Hg Thesis

6/128

Chapter 1

Introduction

As more and more individuals are connected to the Internet, it will penetrate deeper into

our everyday activities, including the way we learn. E-Learning, or Web-based learning is

becoming a new research area in which the web and learning converge on all levels,

whether in elementary school, college, or business. The rapid development of e-Learning

is based on the astounding growth of the Internet and the emergence of new advanced

technologies. For the first time in history people may have access to all kinds of learning

materials 24 hours a day, seven days a week, at any possible location around the world.

The Internet also shows its advantage in course development. For example, if some

educational content, such as a description of computer hardware, is available online, then

it is available worldwide. It could be accessed by each of the thousands of educational

organizations teaching the same topic. Therefore, online learning materials may be easily

reused by many organizations, if there is an efficient way to find the appropriate learning

materials on the Internet and join them together. This is the reason why the new concept,

learning object, has been proposed. The core idea here is that by organizing and

disseminating the learning content into a uniform format as small chunks of learning

materials that are referred to as learning objects, it is possible to achieve content reuse and

interoperation between different educational institutions and training vendors.

In order for different systems to communicate and interoperate with each other, it is

important to have a common language among the systems. Nowadays, the common

language adopted by most learning organizations is eXtensible Markup Language (XML),

since XML can facilitate significant features in the e-Learning framework, such as

personalization, interoperability, reusability and flexibility [XML 2000].

1

8/13/2019 Hg Thesis

7/128

XML was developed to facilitate the description and exchange of data on the Web by the

World Wide Web Consortium [Goldfarb and Prescod 2000]. It is a means of representing

information according to its internal structure. Such a structure makes the information in

the XML files meaningful and machine-readable, and therefore achieves interoperability

and reusability of information [Bosak 1997]. The great potential of using XML has been

predicted by many developers and technology-driven companies. In fact, several learning

organizations, including a group newly formed by IEEE, are trying to develop e-Learning

standards using XML [Gerber 2001].

However, several key problems remain unsolved for developing e-Learning content based

on learning objects and XML. Firstly, the exact definition of a learning object is still

unclear. Several different definitions exist, and most of them are so broad that they may

lose any useful meaning. Secondly, though it may be possible to find the appropriate

learning objects by their metadata, it is not clear whether it is possible for computer

agents to integrate these learning objects in an appropriate way to form a higher level of

course unit that makes instructional sense. Most e-Learning specification initiatives have

not mentioned this issue or just leave it to specification adopters to make the decision by

themselves. Thirdly, there are many e-Learning specifications in XML available now,

however each of them has its own emphasis, and none of them provides a complete

solution for developing an e-Learning system. Therefore, most online instructional

systems are still developed in proprietary ways without adopting the existing

specifications. Finally, although XML was introduced several years ago, it is far from

mature. Many technologies associated with it are still under development or change

frequently, and therefore learning specifications based on XML have to change

accordingly. Moreover, XML is difficult to work with, needing a careful design by

organizations that want to adopt it. All of these issues make the use of XML in

developing a reusable e-Learning system difficult to implement.

This thesis addresses the possibility of developing learning content based on learning

objects and the role played by XML for designing an e-Learning system. The main goal

of this research is to design a reusable and adaptive e-Learning system with XML

2

8/13/2019 Hg Thesis

8/128

schemas forming the basic framework. To implement this goal, the concept of a learning

object is redefined, and a schema for learning objects is developed. Some modifications to

the existing e-Learning specifications, such as the metadata specification, the content

packaging specification, and the learner information specification, are introduced to make

them more suitable for our e-Learning systems. In this thesis, four online tutorial courses

are developed to illustrate how these schemas can work together to make it possible to

reuse learning objects in different learning contexts and to provide learners with

individually tailored learning content.

Chapter 2 of this thesis provides background information on e-Learning. First, e-Learning

and the requirements of future e-Learning systems are reviewed. The concept of a

learning object is selected for a deep examination, since learning objects form the basic

foundation of an e-Learning system. Then the present state of XML, and issues such as

why XML should be used, and what are the benefits are discussed. Finally, the emerging

international standards for e-Learning systems are reviewed, along with the benefits of

open standardization.

In Chapter 3, the features of learning objects according to the new definition are

summarized. The schemas for learning object, metadata, content packaging, and learner

information are defined and explained.

Chapter 4 describes the architecture and detailed design of the experimental e-Learning

system, and Chapter 5 describes the implementation of this system. Chapter 6 summarizes

the thesis work, and discusses research contribution and future work related to the e-

Learning system.

3

8/13/2019 Hg Thesis

9/128

Chapter 2

Background

2.1 E-Learning, A New Way Of Learning

Education has evolved considerably because of Web technology. The Internet enables the

ordinary person to have access to never-ending quantities of information and knowledge

more efficiently and conveniently. The growth of the World Wide Web, high-capacity

corporate networks, high-speed desktop computers and all kinds of mobile devices will

make learning available to people 24 hours a day, seven days a week around the globe.

Since many traditional education organizations are using Web technology to deliver

educational content, it is possible now for a high school student to seek assistance with

mathematics questions at any time of the day or a graduate student at home to take some

courses through long distance education.

Web-based learning not only improves the achievement of students from kindergarten to

university, but also enhances the productivity of the corporate workforce. Turbulent

corporate environments, caused by market dynamics, have made knowledge and skills

indispensable for effective performance in the workplace. Knowledge in the workplace is

no longer implied but required at different times and different quantities. Traditionally,

corporate training has existed in organisations to impart knowledge to individual workers

as off-the shelf learning packages. In this model, learning takes a reactive approach to

problem solving encountered by organisations, and learning programs take place in a

specific location. However, recent advances in the fields of distributed and ubiquitous

computing, artificial intelligence, cognitive learning theory, and multimedia have

converged to provide more distributed learning systems over the Internet and World Wide

Web (WWW).

4

8/13/2019 Hg Thesis

10/128

A requirement for knowledge and skills distribution across different systems, space, and

time is pertinent to unique learning requirements of individual learners within all kinds of

organisations. The infrastructure to support such knowledge distribution is in the form of

electronic learning, normally referred to as e-Learning.

Commonly, e-Learning is defined as Internet-enabled learning, or convergence of

learning and the Internet, including any use of computers and the Internet to facilitate

education [Downs 1998]. The components of e-Learning can include content delivery in

multiple formats through the Internet, management of the learning experience, and a

networked community of learners, content developers and experts.

E-Learning covers a wide set of applications and processes such as Web-based learning

systems, computer-based learning systems, virtual classrooms, and digital collaborative

learning GroupWare packages. E-Learning content is mainly delivered via Internet,

intranet/extranet (LAN/WAN), audio- and videotape; satellite broadcast, interactive TV,

DVD and CD-ROM, and the still to emerge wireless application protocols (WAP)

[ASTD, 2001].

It is estimated that the e-Learning market will grow substantially over the next five years.

Moreover, with the improvement of bandwidth, video, and storage technology, the

demand for e-Learning products and service will increase exponentially [Wiley 2001].

2.1.1 The Features of E-Learning

E-Learning has the potential to revolutionize traditional education, because it could

provide faster learning at reduced costs, increased access to learning, and clear

accountability for all participants in the learning process. It enables businesses or schools

to distribute training and critical information to multiple locations easily. Employees and

students can then access training when it is convenient for them, at home or in the office.

5

8/13/2019 Hg Thesis

11/128

In today's fast-paced culture, organizations that implement e-Learning will provide their

work force with the ability to turn change into an advantage.

However, e-Learning is just now in its infancy [Downes 1998]. As pioneers struggle with

new technologies and new practices, the discipline evolves almost daily. Despite the rapid

change, some significant features of future e-Learning can be identified as the following:

Personalization. The education of the future will become deeply personalized.

The learning topics will be selected based on student interest, student aptitude and

educational level, and societal need. The menu of available courses presented to

any given student will be determined dynamically by the student's prior learning

assessment, by the prerequisite for the new course, and by the learning

management system. A student's daily menu will be varied and constantly

changing, building on each day's achievement.

Interoperability and reusability. E-Learning systems with different

environments and contents from multiple authors must have the ability to work

together. There must be a semantic relationship between different e-Learning

systems. Learning content may be reused in multiple applications andenvironments regardless of the tools used to create them. This requires that

content be separated from context-specific runtime constraints so that it can be

incorporated into other applications. For reuse to be possible, content must also

have common interfaces and data.

Flexibility. Courses could be generated in a variety of forms based on standard

style sheets. Different forms of layout could be available depending on the

purpose of the course and the preferences of the learner. A student can use various

kinds of devices with different processor speeds and memory capacity, from

desktop computers, laptop computers, and mobile devices such as Palm

computers, to access the learning content.

6

8/13/2019 Hg Thesis

12/128

2.1.2 A New Approach to E-Learning Content Development

Despite the wide spread use of e-Learning infrastructure in corporate and educational

environments, current approaches to the development of e-Learning content are expensive

and time consuming. It is common that content developed by a single vendor or

educational institution can be difficult to reuse by a second vendor or institution, even

though the content shares the same meaning and quality [Downes, 2001]. Failure of

systems to interoperate or exchange content and differences in content ontology between

institutions make content reusability and sharing difficult, although content sharing and

reusability will reasonably reduce production cost.

In order to make e-Learning content less expensive to produce and portable across

different hardware and software systems, a new way of developing e-Learning content has

been proposed. This new approach assumes that e-Learning content can be organised and

disseminated in a uniform format as small chunks of learning materials commonly

referred to as learning objects or knowledge objects [Clayton 2000] [Feemster 2000]. It

seems that developing and delivering learning content as objects will promote reusability,

interoperability and content sharing between different training vendors and educational

institutions. When combined, the learning objects, due to their reusability in different

learning scenarios may form educational resources that can be used in different

environments by different individuals. This realization leads many course developers to

believe that the learning object can become the foundation of adaptive instructional

systems that deliver individually tailored learning materials to large number of people at

the same time.

With standards and compliance in place, it is possible for learning materials to be reused

and to travel on different systems. However, another problem comes up due to

unstructured nature of most of the information available: how can someone search

through a vast online repository of objects to enable them to get what they need? The

answer is that learning objects must be associated with appropriate metadata. They must

7

8/13/2019 Hg Thesis

13/128

be labelled as to what they contain, what they teach and what technical requirements are

needed for their use. It should be also noted that issues of content relevance, systems

compliance, and the nature and structure of content are important problems to solve when

designing reusable e-Learning content. Currently, the main technology used for tagging

learning objects is the eXtensible Makeup Language (XML) [Fox 2000] [Gerber 2001].

2.2 Learning objects

2.2.1 Background Literature on Learning Objects

The complexity and contradictions about what constitutes a learning object are reflected

in the different views expressed in its definition. The IEEE Learning Technology

Standards Committee (LTSC) describes learning objects as any entity, digital or non-

digital, which can be used, re-used or referenced during technology supported learning

[LTSC 2002]. It goes on to argue that learning objects include multimedia content,

instructional content, learning objectives, instructional software and software tools, and

persons, organizations, or events referenced during technology supported learning.

However, this definition is so broad that anything related to learning can be interpreted as

a learning object. Therefore, several groups outside the LTSC have accordingly created

some new definitions of learning objects, which normally narrow the scope to something

more specific. Following are some definitions that have been adopted by various

organizations and companies:

[IDC 2001] white paper asserts that a learning object is a standalone piece or

chunk of education that contains content and assessment based on specific

learning objectives and that has descriptive metadata wrapped around it.

[Shepherd 2001] defines a learning object as a small, reusable digital component,

which can be selectively applied - alone, or in combination - by computer

software, learning facilitators or learners themselves, to meet individual needs for

learning or performance support.

8

8/13/2019 Hg Thesis

14/128

[Wiley 2001] concludes a learning object is any digital resource that can be used

to support learning. He claims that this definition is sufficiently narrow to define a

reasonable scope: reusable digital resource, and is also broad enough to include

the estimated 15 terabytes of information available on the publicly accessible

Internet.

Given all these definitions, there still seems to be little clarity, specifically on what forms

a learning object. However, important features of learning objects, shared by most of the

definitions, are centred on the grounds that the use of learning objects should be focusing

on reusability and sharing.

Those definitions explicitly rule out any non-digital and non-reusable resources, such as

actual people, events, books, or other physical objects. Examples of learning objects

includes the smaller digital resources, such as images, paragraphs of text, questions,

animation, audio or video clips, and also the larger resources, such as entire web pages

that combine text, images and other media applications to deliver complete instruction.

These objects, which are reusable, should be stored in repositories and the copies of their

metadata should be available in computers easily accessible by users [Downes 2001];[Shata 2001].

The idea shared by these definitions is that by building learning resources as reusable

learning objects, developers of learning materials, learning managers and learners

themselves will all stand to gain [Klassen 2000]. The benefits can be summarized as the

follows:

Courses can be constructed using learning objects from a wide range of resources.

Course developers do not need to prepare all course materials from scratch,

therefore they can produce courses more economically;

9

8/13/2019 Hg Thesis

15/128

Courses can be customized to suit the needs of different learners by selecting

different learning objects according to the requirement and the interest of learners;

Learning objects can be reused to meet a range of learning needs. It is not

necessary to develop similar learning objects for multiple times;

The same objects can be employed across a variety of hardware and software

platforms, when common standards are followed by the learning objects.

2.2.2 Reusability of Learning Objects

Reuse has existed in the field of software engineering for decades. Software reuse is the

process of creating software systems from predefined software components [McClure,

1995]. The greatest benefit of reuse arises from the possibility for rapidly assembling

small components into complex ones.

With object-oriented design and programming, a completely new way of thinking about

the construction of complex systems was originated. Object orientation allows software

components to be used as building blocks for future software developments and takes

components created by others rather than creating new ones from scratch. In this sense,

learning objects are an application of object-oriented design to the world of learning.

Reusable learning objects represent an alternative approach to content development.

Learning objects treated as small components are pedagogically broken down into small

chunks. Better yet, from a pedagogical perspective, each chunk plays a specific role

within an instructional design methodology. The basic requirement for each chunk is its

ability to communicate with any learning systems using a standardised method that does

not depend on a specific system. Akin to the behaviour of small software objects in the

object-oriented programming model, what happens within an individual learning object is

trivial, since this is obscured from the designer and the user.

10

8/13/2019 Hg Thesis

16/128

Each learning object must have a description that enables designers and computer agents

to search for and find the right objects for the right job. This implies that objects must be

wrapped in metadata. Whatever the properties the learning object has, the metadata itself

should be straightforward. Moreover, since the metadata is machine-readable, it must be

possible for a specific system to interpret the metadata from other sources and then reuse

the learning objects. However, it should also be noted that metadata is limited because it

is only a wrapper for the search engine to identify one object from another. The

knowledge bits inside a learning object cannot be distinguished by metadata.

2.2.3 Problems with Learning Objects

Although it is a good idea to develop course content based on learning objects, some

problems still exist. The learning objects are designed not only for direct human

processing but also for automatic machine processing. They should allow processing by

intelligent services such as information brokers and search agents, which provide greater

functionality. For example, one important benefit that these reusable learning objects can

offer is that, by mixing and matching them, an e-Learning system may customize learning

for individuals. It is also described in the proposal of the Learning Objects Metadata

Working Group formed by LTSC, which tries to enable computer agents to

automatically and dynamically compose personalized lessons for an individual learner

[LTSC 2002]. However, several questions will be raised when we design an e-Learning

system to fulfill this goal:

How should we define the level of granularity of learning objects? Does it make

sense to view any single image, paragraph of text or a question as a learning

object?

Is it possible to use search agents to select and integrate the learning objects in an

appropriate way to form a higher level of course unit, which make instructional

sense?

11

8/13/2019 Hg Thesis

17/128

Are metadata sufficient to facilitate the combination of learning objects?

The important issue not clearly addressed in the literature on learning objects is the

granularity and combination of learning objects. Granularity refers to the size of the

learning object and combination refers to the manner in which learning objects can be

combined and assembled into larger structures to enhance learning [Wiley 2000] [Jamlan

2001].

For two extreme examples, a learning object can be as small as a piece of an image or as

large as a complete course. For the former, such a learning object may be reused by a lot

of courses, however it is difficult, if not impossible, to select those individual learning

objects and combine them directly by computer agents in a way that makes instructional

sense. For the latter, although one can easily re-sequence a complete course in a new

context, the potential reusability of this learning object will be low.

[TechLearn 2001] has reported that there are no clear standards for the size (or

granularity)of a learning object. However, studies show that larger learning objects are

typically harder to reuse than smaller ones [Daniel, 2001]. From an efficiency point of

view, the decision regarding the granularity of learning objects can be viewed as a trade-off between the possible benefits of reuse and expense of combination.

Granularity and combination issues in the design of learning objects are tantamount to

issues of scope and sequencing of learning materials in instructional design [Wiley, et. al,

2000]. The way in which learning objects can be combined with other learning objects is

very much dependent on their scope and structure. However, traditional instructional

design theories that provide explicit scope and sequencing support are not applicable to

learning objects. [Wiley 2000] further argues that the structure of learning and the

combination of objects are like molecular bonding process; molecules of the same nature

can be combined to form complex structures.

12

8/13/2019 Hg Thesis

18/128

Currently a clear ontology for defining learning objects is lacking. Furthermore, if

learning objects are solid entities, which can be referenced by others, as proposed by the

principles of object orientation, then a learning object should support some kind of data

structure.

To address those problems, [Daniel 2001] has initiated a new attempt to characterize

learning objects as pieces of instruction but not just pieces of digital information. He

claims that a learning object should provide instruction on a relatively small, discrete skill

or unit of knowledge, and the content of a learning object should be self-sufficient,

independent of the context. Even so, it should be possible for a learning object to be used

in more than one sequence of instruction.

In this proposal, we will further discuss the features of learning objects in chapter 3,

where we redefine the concept and develop a schema for a learning object.

2.3 E-Learning and XML

For all the new features and technical demands of future e-Learning, XML seems to be a

reasonable answer [Adolphe 2000]. XML is an important step in the direction ofpromoting the interoperability and flexibility of Internet applications. As a simplified

subset of SGML, XML was created as a way to structure, store and send information on

the Internet. Unlike HTML, XML allows someone to create his/her own tags and define

the DTD (Document Type Definition) or XML Schema. The DTD or schema supports a

tree structure, which is much richer than a simple flat list and also respectful of cognitive

and data processing requirements for economy and simplicity.

When XML is used to store unstructured or semi-structured data, for which the traditional

relational database is not suitable, it gives developers the ability to manipulate the

information easily and quickly. With XML, course developers may put semi-structured

information, such as the course content or course structure, into a discrete relational field,

13

8/13/2019 Hg Thesis

19/128

and then work with this information as with structured blocks of data, not as with a string

of bytes.

Therefore, for e-Learning XML provides a flexible approach to represent the content and

the structure of a course, and to keep such information separate from the software used for

delivery and presentation. Moreover, content stored using XML can be independent of

any course, and is in a form ideally suited to re-use in any number of different courseware

and other learning-related products. For example, learning content in XML may be

transformed into PDF to form a part of a book, or into HTML to provide online education.

With the current intense interest in XML and rapid progress in the area of XML data

management, more and more attention has been paid to using XML as an intermediate

format for Web-based information representation, exchange, retrieval and reasoning

[Nauer et al.2000]. It is widely acknowledged that XML will form the standard for data

interchange in electronic commerce, and that the management of large and complex web

site will be much improved by the use of XML.

As the e-learning market grows dramatically, an increasing number of organizations and

vendors are using XML as a standard way to tag or mark up learning information, such as

learning materials, management resources, and student data, so that they can be easily

referenced, read, and exchanged across applications and systems. [Gerber 2001] points

out that XML allows e-Learning designers and administrators to develop applications

faster, reuse course content easily, and facilitate data exchange between Web-Based

courseware content and learning management systems.

However, XML is not a single specification. A rich environment surrounding XML has

been developed by the World Wide Web Consortium (W3C) and others since 1998. This

environment includes many important specifications, such as DTD, XML Schema,

Namespaces in XML (XML Names), Xpath Language (Xpath), Extensible Stylesheet

Language (XSL), XSL Transformation (XSLT), XML Query, XML Document Object

14

8/13/2019 Hg Thesis

20/128

Model (DOM), XML Linking Language (XLINK) and XML Pointer Language

(Xpointer), many of which can be used in an e-Learning system.

2.3.1 XML Schema and its Benefits for Interoperability and Reusability of

Information

In online systems, the most popular use for XML is to create a separation of content and

presentation [Box et al. 2000]. In this situation, we are defining application content as the

data that needs to be displayed to a client or processed by a computer agent, and

application presentation as the formatting of that data. XML is a markup language for

documents containing semi-structured information. Information is stored in XML

documents with a logical structure, therefore it is meaningful and machine-processible.

This kind of format improves the interoperability and reusability of information, and

makes flexible Web-based information representation, exchange and retrieval possible.

Both a DTD and an XML Schema can be used to define the tree structure and establish a

set of constraints for an XML document. However, DTDs have no formal mechanism to

support the declaration of semantic integrity constraints. They show some critical

limitations: DTDs are not themselves XML documents; they have no knowledge of

hierarchy, they have difficulty in handling namespace conflicts, and they have no means

of specifying the types of relationships allowed in XML documents.

For these reasons, the XML Schema specification was developed to replace and amplify

DTDs. Schemas express shared vocabularies and allow machines to carry out rules made

by people. They provide a means for defining the structure, content and semantics of

XML documents. XML Schemas have the same purpose as DTDs, but provide several

significant improvements [Fallside 2000]:

XML Schema definitions are themselves XML documents;

XML Schemas provide a rich set of datatypes that can be used to define the values

of elementary tags;

15

8/13/2019 Hg Thesis

21/128

XML Schemas provide a richer means for defining nested tags (i.e., tags with

subtags);

XML Schemas provide the namespace mechanism to combine XML documents

with heterogeneous vocabulary.

These new features are important progress in the development of XML. They will

influence the ways in which XML can be used. XML Schemas offer an XML-centric

means to constrain XML documents, or to bring DTDs back into line with XML itself.

Every tool that handles an XML document can also be used to deal with schemas.

XML Schemas provide a rich set of datatypes and a richer means for defining nested tags.

This makes XML languages more like an object-oriented programming language such as

Java. Like the definition of data structures of objects in Java, schemas define and model

complex object semantics for XML documents, where semantics fundamentally means

an intricate web of constrained relationship and properties [Cover 2000].

XML Schemas supports a flexible way to use namespaces. Namespaces are the mappings

used for handling definition collisions, when several data type definitions are adopted by

the same XML documents. Namespaces were invented after DTDs and are not fullysupported by them. However, XML Schemas provide a more advanced namespace

mechanism to combine XML documents with a heterogeneous vocabulary. A customized

XML Schema would allow an XML document to make a reference to standard schemas

whenever necessary. This method could improve the interoperability of XML documents.

On the Internet, interoperable exchange and retrieval imply the existence of a sharable

ontology, or common set of object semantics [Klein et al. 2000]. XML Schemas fulfill a

major goal in common with ontology, namely providing semantic vocabulary and

structure for describing information sources that are aimed at data exchange and

information retrieval. Using XML in building learning objects and other learning

information enables designers to share content with their colleagues or with course

designers in other organizations. It should be noted that if a common schema is developed

16

8/13/2019 Hg Thesis

22/128

and adopted, information tagged in XML for one system could easily be integrated with

any other system.

For instance, using XML as a language of metadata allows the user to create new kinds of

descriptors that feature the learning objects. Since the schema ensures that the metadata is

machine-readable and meaningful for the search engines, any learning management

system (LMS), which chooses to support this schema, can use this metadata to select

learning objects. Furthermore, through the use of XML Schema, learning objects can be

structured and presented as nodes of trees in a content packaging file to reflect different

levels of learning granularity and seqencing, which makes the combination of learning

objects possible.

2.3.2 XSL, XSLT and XPath for Flexible Information Presentation

XSL is the Extensible Stylesheet Language. XSL transforms and translates XML data

from one XML format into another. XML provides a way to store data with logical

structures and to define a semantic constraint on the data, but it does not tell how to

present the data. In order to display XML documents to the clients, it is necessary to have

a mechanism to describe how the document should be displayed. One of these

mechanisms is Cascading Style Sheet (CSS), but XSL is the preferred style sheet

language for XML, and XSL is far more sophisticated than the CSS used by HTML.

XSL actually consists of three languages [Berlund et al. 2000]:

XSLT is a language for transforming XML documents into other types of

documents, or into other XML documents.

XPath is a language for addressing parts of an XML document. XPath was

designed to be used by XSLT.

XSL Formatting Objects is an XML vocabulary for specifying formatting

semantics.

17

8/13/2019 Hg Thesis

23/128

XSL is a powerful style sheet for XML documents. Instead of having to rewrite the course

content in different formats or in a different length, the developers can use XML to

separate content from the way it is presented. That allows them to re-organize the content

and make tailored courses for a specific learner in an appropriate format. For example, the

same XML document may need to be displayed in HTML, PDF, and Postscript form.

Without XSL, the XML document would have to be manually duplicated, and then

converted into each of these three formats. Instead, XSL provides a mechanism for

defining stylesheets to accomplish these types of tasks. Through some standard XSL files,

an XML file can be transformed into different formats without having to change the data

due to the requirement for different representations. In this way, XSL helps XML fulfill

reusability of content.

For mobile devices that want to access the learning content, Wireless Markup Language

(not HTML) is used. WML is defined as an XML 1.0 application, so XSLT can also be

used to transform the XML documents to WML files, and then the Palm computer or

mobile phone displays the WML files. With XSL, XML will implement a flexible

information presentation with little effort.

2.3.3 XML DOM and SAX

The personalization of learning content means that every learner may get information

matching particular requirements. Therefore, it is necessary to search and retrieve

different information from the XML documents for different learners. As a data-driven

markup language, XML can be easily searched and manipulated due to the strict structure

and semantic constraint that schemas can impose [Boyle 2000].

XML is simple and easy to use, since there are some software packages for XML parsing.

Right now, two APIs for XML, SAX (Simple API for XML) and DOM (Document

Object Model) are normally adopted. With the SAX, developers have access to the

information in XML documents as they are read, without imposing major memory

constraints or a large code footprint. Meanwhile, DOM is designed to generate a

18

8/13/2019 Hg Thesis

24/128

representation of an XML document as a tree, therefore users may traverse and

manipulate the tree structure to retrieve the data.

As W3C specifications, both DOM and SAX provide a standard programming interface to

a wide variety of applications for XML. They are designed to be used with any

programming language and any operating system. With DOM and SAX, a user can create

an XML document, navigate its structure, and extract, add, modify, or delete its elements.

Several tools currently used these APIs and provide the means to parse a XML document.

2.3.4 Conclusions about using XML

From the above review of XML and the technologies associated with it, it is clear that

XML can benefit an e-Learning system with several advances. Firstly, XML Schemas

provide a way to define a set of elements, which can establish a shared ontology among

different organizations. This helps learning materials go through platforms and be reused

without the problem of compatibility. Secondly, the separation of content and

presentation will enhance the flexibility of displaying learning materials. By adopting the

standard XSLT files, learning materials may be transformed into a variety of possible

standard forms. Lastly, information stored in XML files is easy to search and retrieve due

to the structure and constraint that XML files followed.

It seems that XML is almost ready to be used in such online systems as e-Learning.

However, it should be noted that change and development is occurring rapidly, and that

XML is far from stable. Some guides about using XML in e-Learning system have been

published, however, no practice and experimental data are available yet. Therefore, using

XML in a realistic environment is just in a tentative phase now.

2.4 E-Learning Standards

Why should we develop standards for e-Learning? Maybe we can get some inspiration

from the Lego system, the childrens construction toy. Although individual Lego pieces

19

8/13/2019 Hg Thesis

25/128

have different shapes and sizes, they always follow the standard units of measurement and

standard interfaces. Therefore, no matter how one wants to reassemble the pieces, they

always fit together. The idea behind standards for e-learning is exactly the same. If the

learning objects and metadata follow common standards, the course developers or

computer agents can eventually form an integrated course by assembling all kinds of

learning objects.

In the e-Learning community, XML forms the basic foundation for inter-application

communication, however it does not ensure that communication will happen. For

example, people speaking different languages cannot understand each other. Even in the

same language, such as English, words may have different meanings; the word football

represents different sports in England and North America. To make communication

possible, a set of vocabularies with clear definitions should be shared by speakers. That is

what some organizations are doing; developing standards in XML for e-Learning. In this

section, we will survey some important organization and potential standards in e-Learning

community.

2.4.1 Standards Initiatives for E-Learning

Strictly speaking, there are just a few e-Learning standards now. Several organizations are

concerned with e-Learning specifications that the learning community may support.

Amongst them, Learning Technology Standards Committee (LTSC) from Institute of

Electrical and Electronic Engineers (IEEE), the Aviation Industry Computer-Based

Committee (AICC), the Instructional Management System (IMS), the Advanced

Distributed Learning (ADL) and the Educational Modelling Language (EML) are the

leading ones.

IEEE LTSC

IEEE, ISO (International Standards Organization) and ANSI (American National

Standards Institute) are the major organizations that set most official computer standards.

20

8/13/2019 Hg Thesis

26/128

IEEE has developed many technology standards for electrical and information

technologies and sciences. Several years ago, it set up the LTSC to develop technical

standards, recommended practices, and guides for software components, tools,

technologies and design methods that facilitate the development, deployment,

maintenance and interoperation of computer implementations of education and training

components and systems [LTSC 2002]. Currently, the committee is carefully reviewing

the specifications created by IMS, ADL and AICC, comparing and combining them, and

making sure that they are general enough to fit the requirements of any learning

organizations. Several working groups have been formed concerning the standards in a

number of areas, such as glossary, student identifiers, learner models, course sequencing,

content packaging, learning object metadata and so on. The standard for learning object

metadata was approved on June 13, 2002, by the Standards Board of the IEEE Standards

Association.

AICC

In 1988, AICC was first formed to propose hardware requirements for running computer-

based training (CBT) software developed for the aviation industry. However, it has since

branched into several other areas [AICC 2002]. In recent years, the AICC has developed

guidelines for the aviation industry in the development, delivery, and evaluation of CBT

and related training technologies. AICC recommendations are fairly general to most types

of computer-based training and, for this reason, are widely used outside of the aviation

training industry. It also actively coordinates its efforts with broader learning technology

standards organizations like IMS, ADL, and IEEE/LTSC.

IMS

IMS may be the most influential organization in the e-Learning community. The

contributing members of IMS include many well-known academic, corporate, non-profit

and government organizations. IMS is developing and promoting open specifications for

facilitating online distributed learning activities such as locating and using educational

21

8/13/2019 Hg Thesis

27/128

content, tracking learner progress, reporting learner performance, and exchanging student

records between administrative systems [IMS 2002].

Because XML has shown its advantage in the interoperability and reusability of data, IMS

adopts XML in all of its specifications. Now five specifications are available. When

designing our e-Learning system, we were aware of these specifications and tried to adopt

them in our system [IMS 2002]:

The IMS Learning Resources Meta-data Specifications creates a uniform way for

describing learning resources so that they can be more easily found [IMS 2000c].

The IMS Enterprise Specification deals with administrative applications and

services that need to share data about learners, courses, performance, etc., across

platforms, operating systems, user interfaces.

The IMS Content & Packaging Specification is concerned with creating reusable

content objects [IMS 2000a].

The IMS Question & Test Specification addresses the need to be able to share test

items and other assessment tools across different systems [IMS 2000d].

The IMS Learner Profiles Specification looks at ways to organize learner

information so that learning systems can be more responsive to the specific needsof each user [IMS 2000b].

ADL

The initiative of ADL is to accelerate large-scale development of dynamic and cost-

effective learning software and to stimulate an efficient market for these products in order

to meet the education and training needs of the military and the nation's workforce of the

future [ADL 2002]. It achieves this through the development of a common technical

framework for computer and net-based learning that will foster the creation of reusable

learning content as "instructional objects."

22

8/13/2019 Hg Thesis

28/128

ADL has developed the Sharable Courseware Object Reference Model (SCORM)

initiative. SCORM includes four major areas: metadata, course structure format, data

model, and an application program interface or API. To avoid reinventing the

specifications, SCORM integrates specifications from IMS, AICC and IEEE. It also

created a software program to test the learning objects, learning management system, and

course delivery tools for compatibility.

EML

EML is a research program for educational modelling carried out by the Open University

of the Netherlands (OUNL). It tries to develop a comprehensive notational system that

allows course developers to codify units of study (e.g. courses, course components and

study programmes), in an integral fashion. EML describes not just the content of a unit of

study (texts, tasks, tests, assignments) but also the roles, relations, interactions and

activities of students and teachers. The major EML implementation is in XML [EML

2002]. However, EML pays less attention to the possibility of reusing study units in

different course contexts. It mainly focuses on how the course structure should be

described and how an instruction course should be navigated during a learning process.

This limitation impairs its potential to become an e-Learning standard.

Amongst all of these specifications, the metadata and content packaging or course

structure specifications may be most important ones. In the following section, we will

discuss these two specifications.

2.4.2 Specifications for Metadata

Metadatais known as "structured data about data." The term metadata has been used only

in the past 15 years, and has become particularly common with the popularity of the

World Wide Web. The purpose of metadata is to provide a common way to describe

resources so that they can be self-defined and searched [Gerber 2001].

23

8/13/2019 Hg Thesis

29/128

The first organization dedicated to promoting the widespread adoption of interoperable

metadata standards is Dublin Core Metadata Initiative [DCMI 2002]. It developed a set of

simple specialized metadata vocabularies for describing resources that enable more

intelligent information discovery systems on the Internet. Nowadays, many web sites

adopt DCMI to describe their web pages.

However, in an e-Learning system, DCMI is too simple to support effective resource

discovery. It lacks the elements that may be used to describe the educational features of

learning objects. Therefore, a number of organizations, such as IMS, ADL, and LTSC

Learning Object Metadata (LOM) working group, are trying to develop a new metadata

standard for learning objects. Over the past few years, the IMS have defined a standard

dictionary of metadata elements, which is referenced and used by ADL. Some other

initiatives, such as Canadian Core Learning Resource Metadata Protocol (CanCore), also

follow this specification [Gerber 2001]. The LOM, which has been approved as an IEEE-

SA standard, is also built on IMS Metadata Specification [LTSC 2002].

IMS Metadata Specification

One of the key contributions of IMS is the IMS Learning Resource Meta-data

Specification, an XML-compliant schema for indexing learning objects [IMS 2000c].

The growing popularity of this schema among e-learning projects such as SCORM

(Sharable Content Object reference Model), and MERLOT (Multimedia Educational

Resource for Learning and Online Teaching) and its adoption by a number of educational

repository projects suggests that IMS will become the standard means of describing

electronic educational materials [Lin 2001].

There are 86 elements in total in IMS metadata, which make the metadata a little

complex. All the elements are divided into 9 categories, and each describes one kind of

feature about the learning object:

general Groups information describing learning object as a whole;

24

8/13/2019 Hg Thesis

30/128

lifeCycle History and current state of resource;

metaMetadata Features of the description rather than resource;

technical Technical features of the learning object;

educational Educational or pedagogical features of the learning object;

right Condition of use of the resource;

relation Features of the resource in relationship to other learning objects;

annotation Comment on the educational use of the learning object;

classification Description of the characteristic of the resource by entries of

classification.

CanCore

CanCore Protocol is based on and fully compatible with the IMS Learning Resource

Meta-data Information Model. CanCore has defined a sub-set of data elements from this

IMS model for the purposes of the efficient and uniform description of digital educational

resources in Canada and elsewhere [CanCore 2002]. It is intended to facilitate the

interchange of records describing educational resources and the discovery of these

resources both in Canada and beyond its borders. Currently, the CanCore schema

provides 54 elements in total, which is a little less than IMS specification.

2.4.3 Specification for Content Packaging or Course Structure Format

The learning objects and metadata often need to be collected and packaged to enable

efficient aggregation, management, and deployment. Therefore content packaging or

course structure format can be used as a top-level manifest file describing the course

elements, the course structure, and all external references necessary to represent a course

and its intended behaviour. There are two kind of content packaging used by e-Learning

systems, the Content Packaging conceptual model from IMS and the Course Structure

Format (CSF) from ADL. The IMS content packaging is simpler and clearer than CSF,

and allows references to CSF to be made by using a namespace [Shata 2001]. However,

25

8/13/2019 Hg Thesis

31/128

none of them can provide an efficient way to describe the hierarchical structure of an

instructional course, which contains course design information.

IMS Content Packaging Specification

The IMS content packaging conceptual model creates a Package Interchange File which is

a single file, e.g. zip, jar or cab which includes a top level manifest file describing the

package as a whole[IMS 2000a]. A package represents a unit of usable content, which

includes all learning materials and the associated metadata. A content packaging file is

used to describe in XML all the resources comprising a package and one or more ways of

organizing the resources for presentation.

Figure 2.1 Content Packaging

As described in [IMS 2000a], all data in this e-Learning system could be organized as

Figure 2.1 depicts. The root element in the content packaging file is , which

may also contain optional (sub)manifests. Each instance of a manifest consists of two

major parts: the element describes the content organization and the

26

8/13/2019 Hg Thesis

32/128

element contains all references to all the actual resources. The organization

of the manifest and its nesting functionality permit a great deal of flexibility, which

allows developers to aggregate or disaggregate their material as they wish, thus a package

might be a unit in a course, a whole course or a whole curriculum of courses. The actual

resources form part of the package, with suitable metadata describing each resource.

2.4.4 Specifications for Question & Test Interoperability Specification

The IMS Question & Test Interoperability Specification provides proposed standard XML

language for describing questions and tests. The specification has been produced to allow

the interoperability of content within assessment systems [IMS 2000d]. It describes a

basic structure for the representation of question (item) and test (assessment) data and

their corresponding results reports. Therefore, the specification enables the exchange of

this test, assessment and results data between Learning Management Systems, as well as

content authors and, content libraries and collections.

2.5 Summary

In this chapter, we have reviewed a new approach to education, e-Leaning, and some

important topics around it, such as leaning objects, XML and open standards. It should be

noted that the key problems with learning objects, namely granularity and combination,

are still unsolved. Moreover, the use of XML is still not universal in the e-Learning

community and XML is not used in a consistent manner. There is no specification for

learning objects and the content packaging specification provides no information about

instructional design. Those problems make it difficult to develop an e-Learning system

based on XML that accomplishes reusability, interoperability and flexibility.

27

8/13/2019 Hg Thesis

33/128

Chapter 3

Schema Design

Before designing an e-Learning system based on XML, some fundamental concepts and

ideas must be clarified. For course designers, the idea of learning objects requires an

immense change in thinking. Instead of looking at learning as a fixed linear progression,

we must now look at learning as clusters of independent, stand-alone objects ofknowledge. These learning objects should be dynamically selected and sequenced by

computer agents to form an individually tailored course.

The key problem here is how we should handle these learning objects. We need to find a

way to ensure that these objects can be more easily accessed, can be located by computer,

can be easily updated, and can be seamlessly tied in with other objects to make a more

complete learning package. We also need to make sure that the learning course built on

learning objects can present clear objectives, integrated content, and carefully sequenced

instructional activities.

One answer for this is to use XML as metadata for describing the learning objects. In an

e-Learning system, we use metadata to support resource discovery. Every learning object

will have associated metadata that is written in XML and conforms to a particular XML

Schema. The schema gives specific constraints on the structure of metadata, and also

provides information about the interpretation of the metadata. A series of tags defined in

the schema are combined together to describe the learning objects. A designer or a search

engine will look at the metadata to decide how to find the necessary learning objects and

reconstruct them again.

28

8/13/2019 Hg Thesis

34/128

However, metadata alone is not enough. To make it possible for computers to make

sequencing or any other instructional design decisions, the computers must have access to

instructional design information to support the decision-making process. Therefore, it is

necessary to have some way to describe the organizational structure of the learning

objects and to include enough instructional design and sequencing information. Moreover,

to make the instructional design suitable for an individual user, background information

about learners is also required.

Therefore, a basic framework for an e-Learning system becomes clear. A library of

learning objects with metadata forms the basis of the system. In order to construct

personalized courses for learners, a data model of a learner is needed to provide

background information to describe the learners needs. The Learning Management

System, including a search engine, uses the learners interests to select appropriate

learning objects and then pieces together courses based on the organizational structure

and instructional design information

In our project, the following aspects are defined in XML schema:

The learning object; Metadata for retrieval and reuse;

Content packaging for the learning structure, including information about

sequences and alternations of the learning objects;

Learner model, including the learners personal properties, such as skills, prior

knowledge, and background information.

Since schema development is an expensive process, in this project, we borrow from

existing standards or specifications when possible, such as the metadata specification, the

content packaging specification and the learner information specification. However, it

should be noted that some modifications are needed to make the schemas more suitable

for a reusable and adaptive e-Learning system. Moreover, we redefine the concept of

29

8/13/2019 Hg Thesis

35/128

learning object and develop a schema for it according to the new features that we have

summarized.

All of these four schemas and a learning management system form the foundation of the

whole e-Learning system. In the following section, the schemas are described in detail,

and the instance XML documents conforming to the schemas are provided in Chapter 4.

3.1 Schema for Learning Objects

3.1.1 New Definition of Learning Objects

As mentioned above, if we define learning object in too broad a way, we lose any useful

meaning about learning object. Although reuse is the core of the learning object notion, as

flexibility, adaptivity, and interoperability are all facilitated by the property of reuse, there

is still a trade-off between the benefit of reuse and the cost of combination and re-

sequencing. In this research, we prefer a much narrower definition, which makes the

automatic construction of individually tailored courses possible.

According to a common definition, a learning object is a small, reusable digitalcomponent that can be selectively applied - alone or in combination to support learning.

This definition makes it impossible to develop a schema for learning object, since any

digital resource, such as a single picture or a whole web site, can be viewed as learning

objects. It is difficult to find a common structure among those resources and to define the

granularity of learning objects. However, to travel across different platforms, learning

objects should have a uniform format. Moreover, such a format should be able to facilitate

the reuse and re-sequencing of learning objects. Therefore, we need a more specific

definition.

In our research, a learning object is defined as a combination of smaller knowledge bits,

such as text, image, and audio or video clips, which are integrated together to explain or

describe a single core concept in a course. Each learning object can stand alone as a

30

8/13/2019 Hg Thesis

36/128

collection of content items, practice items and assessment items that are combined based

on a single learning objective. The critical features of a learning object may be

summarized as follows:

A learning object is an integrated knowledge object, focusing on a core concept, not just

a piece or a chunk of information.

Normally, a learning object is smaller than a course, a module, or a lesson. However, in

order to describe or explain a single core concept clearly, a learning object may still

include some integrated knowledge bits, which may be in different formats, such as text,

pictures, video clips, maps or simulations. All materials in a learning object will be

organized to surround and describe a core concept. It is likely that the organization of a

learning object should include a definition of the core concept, detailed description of the

concept and sometimes several examples, a conclusion, and some test items or exercises.

When a single text, image, video clip or a combination of those materials cannot provide

a comprehensive instruction independently and needs be joined with other materials to

explain a core concept, it will not be viewed as a learning object.

A learning object cannot include a complex hierarchical structure.

Learning objects are the lowest level of curriculum structure; therefore unlike the upper

level structures of a course, such as lessons, modules, units or topics, they cannot have

embedded concepts or units. Normally, a learning object should just have a flat structure.

The constituent parts in one learning object are all in identical status. It is prohibited that a

description about the core concept also contains another definition and description about a

new concept. This means that if a learning content for one concept is constructed by

several smaller sub-concepts, and each of the sub-concepts can form a stand-alone

content, the instructional content about this concept may not be developed as a learning

object. Instead, several smaller learning objects should be established to describe the sub-

concepts, which are then assembled and combined to instruct a more complex concept.

31

8/13/2019 Hg Thesis

37/128

A learning object is the basic reusable unit.

The greatest potential for reuse exists when the learning object centres on a single, core

concept and does not rely on the support of other course contents or context to clearly

provide instruction on this concept. A learning object can be used in more than one

sequence of instruction. For example, when designing the hierarchical structure of a new

course, the designers can just indicate what kind of learning objects are necessary, and

then the search engine may find the appropriate learning objects from the global

repository of learning objects. However, the knowledge bits in a learning object are not

necessarily reusable units which can be handled directly by a search engine. If the

designers only want to reuse some parts of the knowledge bits in a learning object, they

should inherit this learning object or construct a new learning object by using some

knowledge bits in this learning object. However, in such case, the work must be done by

the instruction expert, not by a search engine.

Several learning objects may describe the same concept.

Since the users of a course may have different backgrounds, studying abilities and diverse

levels of interests, for one specific concept, a learning system should provide different

learning objects that may range from easy to difficult or from simple to complex.

However, the learning objects that describe the same concept will probably have some

common parts. To avoid multiple duplications of the content, some disciplines in object-

oriented programming may be adopted, such as inheritance. Therefore, from one basic

learning object, a set of extended stand-alone learning objects can be developed, which

may be more difficult or more comprehensive. In such a case, a schema can be used to

keep a consistent structure, which enhances inheritance between learning objects

A learning object often includes some test items to evaluate the learners performance.

To make learning objects independent of course context and reusable in different course

sequences, learning objects should include the test items by themselves. Directly tied to

32

8/13/2019 Hg Thesis

38/128

the core concept, test items may be used to identify whether a learner has mastered a

given learning object. With test items a learning object will not solely depend on the high-

level context to evaluate the learners performance, and therefore fulfils a special learning

objective on its own.

A learning object must be searchable.

Each learning object should be associated with metadata. The attached metadata can

implement the express goal of interoperability, which allows search engines throughout

the world to be able to successfully find and use learning objects.

A learning object will normally be embedded in a hierarchical structure.

Each learning object provides stand-alone instruction, so it is difficult to define the

connection between learning objects in the content of a learning object itself. To help the

learner comprehend the whole structure of a course or smooth the flow in the instruction,

a hierarchical structure about the instructional content is necessary, which is used to

connect learning objects.

The preferred form of representation of learning objects is XML.

Learning objects in the form of text can be directly represented in XML. The XML tags

make the data in the learning objects meaningful, so the components in the learning

objects can be searched, extracted and reused in various ways. In addition, learning

objects in XML can be transformed to a variety of forms, such as PDF or WML, based on

standard style sheets. This makes the whole system more flexible.

In general, the granularity and combination of learning materials are implemented at two

levels. At the low level, the learning object combines several knowledge bits to explain a

concept, and at the higher level, the hierarchical structure describes the aggregation of

33

8/13/2019 Hg Thesis

39/128

learning objects to form courses. This constitutes the solid underpinnings of a schema for

learning objects.

3.1.2 The Design of a Schema for Learning Object

In this project, we define a Learning Object Mark-up Language (LOML), which is an

XML-based markup language designed for instructional contents. Because our specific

interest is in the domain of Computer Science, this language is applied specially to

describe topics as they apply to the field of Computer Science. When designing this mark-

up language, we have taken into account the critical features to choose a simple set of

mark-up tags to define the structure of a learning object.

Diagram

Attribute Name Type Use

inheritance xs:anyURI optional

Figure 3.1: Structure of Element

34

8/13/2019 Hg Thesis

40/128

Figure 3.1 shows the first level of elements of LOML. In this thesis, all diagrams for the

schemas are generated automatically by XML Spy, so they will follow the same style.

The notations in this diagram are mainly adopted from UML (Unified Modelling

Language) with some revisions. Each rectangle in the diagrams represents an element,

with the name of the element inside the rectangle. If two rectangles lap together, it means

that this element may appear multiple times. The signal + at the right side of the

rectangle indicates that this element has also some sub-elements, which are not shown in

the diagram. The switch notation between the root element and the

first level elements means that one may choose any element from the first level element in

any order to occur in an instance XML document. However, if the notation is a straight

line, this means that the elements may just appear in the sequence that has been defined.

In this schema for learning objects, the root element is . Other elements

are organized into eight different categories:

title The title of the learning object;

definition The definition of the core concept on which the learning object

focuses;

description Some detailed descriptions of the core concept;

exampleSome examples about the core concept; application Simulation or demonstration used to explain the core concept;

conclusion The conclusion about the core concept.

exercise The exercise used to improve the study of learner.

test The test used to evaluate the result of study.

The inheritance attribute of the root element indicates from which learning object the

current learning object inherits instructional content. If the attribute is not empty, the

XML file indicated by the attribute will be loaded and combined with the tagged contents

in the current learning object to form a complete description of the core concept.

For each sub-elements of the root element, a unique id is assigned, which can be used as a

reference to this element. Each element may also have reference attribute and status

35

8/13/2019 Hg Thesis

41/128

attribute. If the learning object inherits content from a parent learning object, the

reference attribute indicates one specific element in the parent, and the status attribute

indicates what kind of action will be taken. The following lists the available actions:

If the status is substitution, it means that the content in the current element

should replace the content in the parent element, which is indicated by the

reference attribute;

If the status is integration, it means that the content in the current element

should integrate with the content in the parent element, and then combine together

to form a single content;

If the status is connection, it means that the current element and its content

should be inserted behind the parent element;

Diagram

Attribute Name Type Use Facetsid xs:string required

reference xs:string optional

status statusType optional substitution, integration, connection

Figure 3.2: Structure of Element

36

8/13/2019 Hg Thesis

42/128

Since a learning object may contain text content, we still need some elements to support

such textual formats as paragraph, table, list, and head. Figure 3.2 demonstrates the

structure of element and the attached attributes. In this structure, in addition

to some text elements, the element is used to indicate what kinds of multimedia

are used in this learning object, so the proper tools can be invoked to display the

multimedia files. This element includes three attributes, type, mimeType and uri,

which present the type of the multimedia file and the reference to the file. Moreover, the

element may be used to tag some special texts, such as formulae or algorithms.

However, for the element and element, it is possible to adopt an

existing specification to develop test items. In this schema, QTILite specification from

IMS is imported. This specification is based upon the IMS QTI specification and is the

realization of a subset of that model. It is presented as the entry-level specification to the

full QTI specification, just supporting the question styles of true/false and multiple

choices with single answer.

Figure 3.3 Importation of QTILite Schema

Figure 3.3 demonstrates how the QTILite specification is included in our LOML schema.

The schema file, ims_qtil_rootvip1.xsd, is loaded by the element and the

namespace, qtil, is assigned to reference the QTILite schema. Therefore, the LOML

schema may use any element type from QTILite to define new elements. By adopting the

37

8/13/2019 Hg Thesis

43/128

specification from IMS, the test items developed for learning objects may be easily reused

by other e-Learning systems. A detailed description about the use and implementation of

OTILite can be found in [Brooks 2002].

3.2 Schema for Metadata

3.2.1 The Design of Schema

There are currently several metadata XML schemas focusing on e-Learning. However,

though those schemas may be used, a specific metadata XML schema should still be

created in this research for the following reasons:

The purpose of most standard schemas having the goal of interoperability is to

allow everyone throughout the world to be able to successfully find and use

learning objects. Therefore these schemas are complex, often with about one

hundred element tags. An alternative is to choose some necessary elements from

them and to build a new schema which will provide a good enough description for

the learning objects in this specific project;

A designer wants to be able to search for the specific topics within the library of

learning objects. To do this effectively, the schema must have a controlled

vocabulary, only allowing certain words to be entered between tags. While this

makes the searches fast and accurate, it is not possible to do this without creating a

unique schema. In fact, we need to add more constraints in our own XML schema

and make the metadata document more meaningful.

However, using a unique XML schema will not significantly harm the interoperability of

the e-Learning system. We choose the tags from the standard schema, so every tag in our

schema is still meaningful to others. A third-party search engine that can handle the XML

metadata documents conforming to the standard schema could also handle ours. In

practice, we selected about 23 tags from the IMS Learning Resource Meta-data

Specification. Figure 3.4 shows the first and second level elements in the tree structure of

the metadata schema:

38

8/13/2019 Hg Thesis

44/128

Figure 3.4 Structure of Metadata

39

8/13/2019 Hg Thesis

45/128

1. general Context independent features of the resource

identifier - Globally unique label for learning objects;

title Name give to the resource;

language The human language used by the learning object;

description A textual description of the content of the learning object;

keywords Keywords describing the resource.

2. technical Technical features of the learning object

format Technical data type of the resource;

location A location or a method that resolves to a location of the resource;

3. educational Educational or pedagogic features of the learning object

interactivitytype the type of interactivity supported by the learning object;

learningresourcetype specific kind of resource, such as text or figure;

context the typical learning environment where use of learning object is

intended to take place;

semanticdensity subjective measure of the learning objects usefulness as

compared to its size;

difficulty how hard it is to work through the learning object for the typical

target audience;

4. relation Features of the resource in relationship to other learning objects

kind Nature of the relationship between the resource being described and the

one identified by element;

resource Resource the relationship holds for.

5. classification Description of a characteristic of the resource by entries in

classification

taxonpath A taxonomic path in a specific classification.

40

8/13/2019 Hg Thesis

46/128

3.2.2 Controlled Vocabulary

Many search engines have a problem with interpreting the content between tags. Each

person will use different words to describe the same learning object. When search engines

try to find useful learning objects, this reduces the accuracy of the search. The best way to

overcome this limitation is to employ a controlled vocabulary. Of course, there will be

elements that allow a text description without controlled vocabulary, such as

element. However, these elements will usually not be used for searching.

They will be used once a search has been completed to enable either the designer or the

student to make a decision about whether or not to use the learning object without

actually looking through it. Whenever possible, it is useful to specify a controlled

vocabulary.

It is desirable that some existing formal vocabularies can be adopted; however, this is not

always possible. In the schema for metadata, three sets of vocabularies are selected to be

used directly to describe the learning objects:

Computing Curricula 2001.This is a joint task force to undertake a major

review of curriculum guidelines for undergraduate programs in computing bythe Computer Society of the Institute for Electrical and Electronic Engineers

(IEEE-CS) and the Association for Computing Machinery (ACM) [CC 2001].

In this curriculum a set of vocabulary has been defined to represent different

levels of topics in computer technologies. In the element, this

vocabulary is used to classify the areas with which the learning objects are

concerned;

Code for the representation of the names of languages.It comes from ISO

639, in which each language is represented by a two-character code. In the

schema, the element adopts those codes to indicate what kind of

language is used;

41

8/13/2019 Hg Thesis

47/128

Multipurpose Internet Mail Extensions (MIME). MIME defines all the

medium types on the Internet, so that the element can use them to

describe the data types of learning objects.

However, when a formal vocabulary is not available, a set of vocabulary is defined

according to the necessity. For example, the element uses the vocabulary,

{very easy, easy, medium, difficult, very difficult}, to describe how hard it is to work

through the learning object.

3.3 Schema for Content Packaging

In an e-Learning system, instructional content must be collected and packaged in some

electronic form to enable efficient aggregation, distribution, management, and

deployment. The content packaging from IMS may be used to enable the encapsulation of

the required learning resources, simple course structure information, and other supporting

information such as metadata, which promotes the flexibility and interoperability of

learning materials.

However, one of the major weaknesses of the IMS content packaging specification is that

it only supports a simple structure, and does not include any instructional designinformation. Therefore, it is necessary to add new approaches, such as hierarchical

branching, or custom learning paths with conditional branching, if they are required. In

this research, the basic structure of the IMS content packaging specification is adopted,

and then some new features are introduced to make flexible generation of an indivi