Distributed Open Asynchronous Information Access Environment






B.S., Bilkent University, Turkey, 1992

M.S., Syracuse University,1996




Submitted in partial fulfillment of requirements for the degree

of Doctor of Philosophy in Computer Science

in the Graduate School of Syracuse University


November 2000






Professor Geoffrey C. Fox





© 2000 Mehmet Sen


All Rights Reserved






Recent technological improvements have been adopted by both professionals and non-professionals in many fields. Information access and its management in the collective work of distributed objects have become quite critical in many areas such as in distance learning.  This thesis discusses the major issues in designing a multi-tier software architecture for a distributed, open, and asynchronous information access environment and in building and analyzing our experience both in implementing and in using its prototype, which we called a Virtual Classroom Manager.  

The Virtual Classroom Manager exploits recent multi-tier architecture approaches. User-level services using open interfaces on Web browsers are connected to services at the middle-tier, which further communicates with another service tier, i.e., object providers from legacy storage devices, or from other distributed services.

We present a scientific and architectural view of asynchronous support in distance education. For this research, we studied technological manipulations of Learning objects based on a developing taxonomy of Learning objects rather than by examining fields outside the computer science such as instructional design theory. This work is pioneering in the sense that it fills the gap between the high-level concepts of Learning systems and the small-scale technology applications of distance education. Our novel interpretation of Learning technology applications is derived from Web objects and is a rather generic approach that is applicable to other fields.




Table of Contents

Chapter 1 - Introduction. 1

1.1  Problem Statement 2

1.2  Contributions of This Research. Error! Bookmark not defined.

1.3  Organization of the Thesis. 8

Chapter 2 - Literature Review.. 10

2.1  Learning Technologies. 10

2.1.1 Learning Technologies Standards and Specifications. 10 PAPI 11 SCORM... 12 IMS. 13

2.1.2 High-Level Learning Technology System Architecture Model 16

2.1.3 Current Practices. 18 WebCT. 18 Blackboard. 20 VPL. 22 Synchronous Delivery. 23

2.2  A Taxonomy of Distance Education Components. 24

2.3  Web-Mining Technologies. 25

Chapter 3 - Multi-tier Architecture of the Virtual Classroom Manager. 27

3.1  Motivations. 28

3.2  Multi-Tier Architecture. 33

3.3  Implementation Technologies and Issues. 38

Chapter 4 - Middle-Tier Service Components. 48

4.1  Services. 50

4.2  Service Objects. 51

4.2.1 Learner Objects. 51 Functionality. 55 Data operations. 57

4.2.2 Course Objects. 59

4.2.3 Assignment Objects. 63 Submission of assignments. 66

4.2.4 Quiz Objects. 67

4.2.5 AssessmentNugget Objects. 73 Data Analysis. 74 Information Analysis. 75

4.2.6 Survey Objects. 76

4.2.7 Performance Objects. 77

4.2.8 Supervisor Objects. 79

4.2.9 System Administration Objects. 80

Chapter 5 - Lessons on Interoperability Issues of Distributed Components. 82

5.1  Is It a Dream or Doable?. 87

5.2  Interactions within Other System Architectures. 89

5.3  Description Framework of Learning-Web Objects. 90

5.4  Discoverability of Learning-Web Objects. 93

5.5  Event System of Distributed Web Object Services. 95

Chapter 6 - Front-Tier Customizable Open Interfaces. 98

6.1  HTML-Like Template Files. 99

6.2  XSL Templates. 102

Chapter 7 - Security Issues of an Open Access Environment 106

7.1  Communication Channel Security. 106

7.2  User authentication. 108

7.3  User access lists. 109

7.4  User privilege levels. 109

7.5  Presentation security. 111

Chapter 8 - Lessons Learned. 112

Chapter 9 - Concluding Remarks and Directions for Future Research. 116

Bibliography. 120

Vitae. 129

Glossary. 131


List of Tables

Table 4‑1 VCM Object Service List. 50

Table 4‑2 PAPI information access control versus VCM support. 56

Table 4‑3 PAPI operations on data sets (collection of information types)[PAPI] 57

Table 4‑4 PAPI application-specific data operations[PAPI] 58

Table 6‑1 Sample HTML template tags and parser functionality. 100

Table 7‑1 Privilege categories in VCM. 110


List of Figures

Figure 2‑1 The LTSA architecture system components. 18

Figure 3‑1 Multi-tier architecture of VCM... 34

Figure 3‑2 Distributed service distributed objects architecture of VCM. 37

Figure 3‑3 Starting architecture. 40

Figure 3‑4 Implementation view of multi-tier architecture. 43

Figure 4‑1 Some applications of Learner Information categories in VCM... 53

Figure 4‑2 Course structure format of SCORM... 60

Figure 4‑3 VCM course object structure. 61

Figure 4‑4 VCM assignment object structure view. 64

Figure 4‑5 A client view of assignment object in VCM. 65

Figure 4‑6 Centralized virtual file management through open interfaces in VPL. 67

Figure 4‑7 Logical relations of QTI elements at IMS QTI. 68

Figure 4‑8 VCM logical data structure and elements in the Quiz object. 70

Figure 4‑9 IMS QTI Response-type taxonomy. 71

Figure 4‑10 VCM Quiz service approach. 73

Figure 4‑11 An automated assessment architecture to extract hidden nuggets in the learning process by mainly tracking course content page accesses. 74

Figure 4‑12 Survey Object Service display at client site. 77

Figure 4‑13 General hierarchy of the performance object in VCM. 78

Figure 4‑14 Screen captures of performance object views in VCM. 79

Figure 4‑15 Client view of system administration object in VCM. 81

Figure 5‑1 The need for container framework to hold frameworks. 85

Figure 5‑2 Varying consumers of the same information object. 91

Figure 5‑3 Information transformation chain before the client get the resource. 93

Figure 5‑4 Global Object Service Locators. 94

Figure 5‑5 High level event mechanism of distributed object web services. 96

Figure 6‑1 An HTML-like template and parsed result 101

Figure 6‑2 Different outputs for varying clients through XSL mechanism.. 104

Figure 7‑1 Autentication screens for access into VCM... 108



I would like to express my appreciation to my advisor, Professor Geoffrey C. Fox, for his guidance throughout my research and for his wise and acute observations on how to improve my work. I also want to express my sincere thanks to Dr. Nancy J. McCracken for her co-advising and help throughout my graduate study.

I gratefully acknowledge Dr. Geoffrey C. Fox, Dr. Nancy J. McCracken, Dr. Jack Graver, Dr. Shiu-Kai Chin, Dr. Roman Markowski, and Dr. Ehat Ercanli for serving on my defense committee. I am especially grateful to Dr. Graver for serving as committee chair on my defense.

I am enormously grateful to Mrs. Elaine Weinman for her patience in proofreading this manuscript.

I owe many thanks to my parents, Aydin Sen and Gulser Sen, my father and mother, respectively, for their continuous moral support and prayers. I thank my wife and children for their enormous patience during the course of my work toward a Ph.D. degree.  I could not have finished it without their help.

I would like to express my appreciation to my many friends who contributed to my work with their encouragement. Of these, Dr. Kivanc Dincer, Dr. Haluk Topcuoglu, Dr. Erdogan F. Sevilgen, Dr. Erol Akarsu, and Mr. Ozgur Balsoy deserve special thanks. Their moral support and suggestions were of great value.


Chapter 1        


The last decade was a time of widespread usage of the rapid technological improvements related to networking distributed resources. In the first half of the decade, research grants mainly focused on how to achieve a background of world wide web technology. High-speed networking across geographically distributed computers was followed by providing a uniform computing and communication environment on top of the various hardware and software platforms, e.g., PC-Windows and Unix-X-windows. Having the potential for a well-achieved and broad networking backbone, the second part of the decade mainly focused on building software architectures of distributed resources that would be exploited in any field, including science, business, education, health, government organizations, etc. 

As a result of the remarkable research done and its technological applications, distributed computing and remote information access became popular and common. Currently, almost everyone, from the professionals of practically any field to non-professionals is connected to worldwide-distributed resources. This has led to the resources on legacy systems being turned into distributed objects. Information access and management in the collective work of distributed objects that have become quite critical in areas such as distance learning.  The contributions of a worldwide and diverse collection of programmers have made a remarkable new software environment of astonishing functionality. The future seems to demand software of collective interactions of being developed components in both fine and coarse grain. The interactions will be of a rather high level through common open interfaces among specific interfaces of heterogeneous machine types, e.g., personal computers with Windows-based user interfaces and Unix workstations with X-Windows-based user interfaces. Furthermore, the collection not only of heterogonous machines but also  of heterogeneous models has appeared as a challenge in the software development arena. Terms such as “metacomputing” and “meta-applications” are already well known nowadays.

1.1      Problem Statement

New technological improvements have enhanced our experiences. We believe that the above-mentioned recently structured software environment can benefit our range of educational experience with such contributions as distance learning. As stated by [Wayne99], the global hardware and software force is shaping excellence in education, training and learning. Improvements in education are no longer decided only by instructional designers, but also through the contributions of computer science researchers among others. The concepts of education, training and learning are gaining meaning outside of their traditional scope. The tremendous increase of easy-to-access information and production tools make their conception by humans almost impossible using traditional learning skills. We must therefore develop new environments for learners by using our new technological abilities. Investors in this new Information Age seem to be willing to invest much more in learning technologies. Technology-enhanced learning environments are becoming vital in science, industry, government, and other areas.

In our experience with web-based distance learning, we realized that the off-campus geographically-distributed students had more difficulties than on-campus students in following a course presented via recent web technologies. A new, complete and integrated environment needs to be presented that will make off-campus registered students feel that they are attending a virtual university and that will also benefit on-campus students with a complete course environment on-line.

Not having student records in a reliable, easily-securely accessible database environment is a great disadvantage for both on- and off-campus students. Web-based interfaces should allow users to access specific information, perform related operations, and store the results in a stateful environment.

Our team offering the web-based courses also felt overloaded by the amount of technical work required for the management of class. The construction of class lists, e-mail lists, Unix accounts, lists of student home pages and the submission of passwords to students, keeping track of add-drop students, and assessing class levels by surveys required many hours of human effort because of the lack of an automated environment. Grading in such courses became a problem for distributed teaching teams. For example, issues such as online progress tracking, a categorized questionnaire databank, automatically generated evaluation reports, and similar services became a necessities rather than luxuries. As a result, the geographic distribution of instructors and students required the use of the latest technology environments to develop, deliver, evaluate, and administer training. The collaboration of users, e.g., instructors with students, students with students, instructors with instructors, both synchronously and asynchronously turned out to be a definite necessity.

The Web is not just for semi-statically loaded document accesses. Rather, it is a distributed-object technology charged with producing general, multi-tiered intranet and internet applications. Immediate results of several Web-related pop-up technologies brought a fast growing number of easy-to-develop applications. Although technologies continuously make it easy to manipulate distributed objects, we observed that current distance learning applications lack a well-designed architecture. Some existing works are founded on client-server architectures, while these approaches are becoming historic in today’s computing platforms. Unfortunately, not every use of wonderful technological tools produces a final product that is subject to further integration and usability inside the continuous-development environments. For example, the systems currently developed for distance education have short-termed taxonomies and for-the-day architectures, and some well-known systems have already became unpopular. Today’s education systems, however, need to be supported by new open architectures for managing courses, instructors, students and performance records, for grading home works and exams, and for supporting assessment, legacy systems, and security.

As current efforts to set standards focus on an area that is still in its infancy, we believe that future knowledge and learning management systems (LMS) should have open distributed access with the properties of interoperability and reusability as well as discoverability and accessibility on a set of scattered web objects. We need to find a way to describe our web objects that can be easily discovered by others through open interfaces. To illustrate the shortage of today’s learning management systems, none of the existing systems can move or copy a course from one system to another. We believe that none of the efforts thus far has given enough importance to an architectural view of a learning system capable of a broad deployment of web learning objects, which, in fact, would shape the abilities and usability of future systems. In the future, traditional classes may gradually be replaced by learner-centric, personalized performances by the increase of learning-web object environments.

1.2      Contributions of This Research

The biggest contribution presented in this dissertation is designing a multi-tier software architecture for a distributed open, asynchronous information access environment. We developed a novel architecture including a collection of Web object services. Moreover, building and analyzing our experience from both implementation and use are complementary contributions. This thesis discusses the major issues in designing the environment architecture and describes the development and implementation issues of Virtual Classroom Manager as a prototype of the multi-tiered architecture on top of the Internet software infrastructure. 

Virtual Classroom Manager interprets resources as web objects through the incorporation of emerging Web technologies inside an environment of multi-tier architectures. 

We present a scientific and architectural view of asynchronous support in distance education. For this research, we studied technological manipulations of Learning objects based on developing taxonomy of Learning objects rather than examining fields outside the computer science such as instructional design theory. This work is pioneering in the sense that it fills the gap between the high-level concepts of Learning systems and the small-scale technology applications of distance education. Our novel interpretation of Learning technology applications is derived from Web objects and is rather generic approach, which is applicable to other fields.

We showed how to use individual standards, currently specifications, in a large application domain while providing interaction and correlations between individual standard-specific components, i.e. Web-Object services. We presented how to use technology to solve interoperability issues of distributed services. We also contributed to those standards in the sense of how one can use modern technology and architectures to adopt them widely.

The Virtual Classroom Manager exploits recent multi-tier architecture approaches. Open interfaces on Web browsers are connected to services at the middle-tier, which further communicate with another service tier, i.e., object providers from legacy storage devices or from other distributed services. Our contribution is in the integration of modular objects at the back-end with middle-tier services and flexible display interfaces at the front end.

Middle-tier object providers are designed to produce objects in well-understood formats such as XML or DOM trees through robust technologies that include secure http protocol for easy collaboration with any other system or to allow linkage as a component of a requesting system by allowing further exploration of discoverability, accessibility, interoperability, and reusability.

We experimented with different user interface strategies and implemented an easy-to-customize, well-designed graphical user interface scheme. Interface customizations are as easy as HTML editing. [1]

As an important issue in its context, we researched security issues from the very-back tier to the end-user rendering.

To perform assessment, we surveyed web-mining technologies and designed a supportive architecture component as a part of our general assessment progress. This component placed in general architecture as a service for assessment related web objects.

We experimented with different technologies for implementing our architecture. The final system is a reliable, easy-to-customize component within a distributed education arena. The proposed multi-tier architecture model provides asynchronous distance education. Our prototype system is also used to support another synchronous collaboration system and provided content management in distance education.

The multi-tier architecture model is applicable to many other applications areas to develop information management systems in industry, government, etc.. Any application that target the connecting of back-end repositories to web browsers in terms of highly functional services may find our approach useful. Specifically the designs, which manage human information in a distributed and open environment, will find our study helpful.

1.3      Organization of the Thesis

The organization of the rest of this thesis is as follows:

Chapter 2 briefly surveys related work in the literature and illustrates the related terminology used in the rest of the thesis.

Chapter 3 introduces our multi-tier architecture model and its evolution as well as the motivations behind it. It also describes implementation issues of the Virtual Classroom Manager (VCM), which is a prototype of our architecture model.

Chapter 4 provides an in-depth discussion of the middle-tier service components in our multi-tier architecture. The services that cover a whole set of distributed open asynchronous information management environments are explored. The common characteristic of the services is providing a learning management system through the open interfaces of commodity technologies, which we call Web-Object services.

Chapter 5 discusses a distributed environment from the perspective of the lessons we learned from designing our architecture and form building and using the VCM. A new distributed-component computing framework paradigm is proposed based on experiments with our Web-Object services and ongoing specification and standardization efforts.

Chapter 6 discusses major implementation issues and models for preparing convenient open interfaces for a distributed information management environment with heterogeneous clients.

Chapter 7 touches on practical security issues starting from the back-end services and ending at front-tier clients while concentrating on  presentation issues such as displaying public and private information in a distributed environment and explaining the security concerns for the VCM.

Chapter 8 summarizes the overall lessons learned from our research.

Chapter 9 summarizes the work done in this thesis, highlights major contributions, discusses possible extensions of the VCM, and presents future directions for our research.

The glossary at the very end of the thesis provides key definitions of the terminology used throughout, which readers from outside the learning technology field may find useful and even necessary. The glossary also defines some recent commodity technology terms and some terms from other fields.





Chapter 2        

Literature Review

Several research projects have influenced the work presented in this thesis. This section presents some of those related works and current literature in our research area, which includes distributed commodity computing, learning technologies, data mining, etc.


2.1      Learning Technologies

Recent technological improvements have advanced the education area and brought new learning technologies.


2.1.1 Learning Technologies Standards and Specifications


As a result of the appearance of new learning technologies, a definite necessity that has come into account is the widespread adoption of common standards.  Currently, producing standards for learning technologies is very important for producing a revolutionary set of technology and tools in large-scale learning architectures. In our research, we have observed an ongoing effort in   the last half of the decade that covers the common standard for metadata, learning objects, and learning architecture. Among much learning-object metadata, computer-managed instruction, course sequencing, and learner profiles are the areas subject to producing the accredited standards proclaimed by various committees that include the IEEE Learning Technology Standards Committee (LTSC), Advanced Distributed Learning (ADL), IMS (Instructional Management System), Aviation Industry CBT (Computer-Based Training) Committee (AICC), PROmoting Multimedia Access to Education and Training in EUropean Society (PROMETEUS), and the Dublin Core. In following sections are brief overviews of selected standards efforts. PAPI

The Public and Private Information (PAPI) specification is a standard effort to describe so-called portable learner records. It specifies the syntax and semantics of a Learner Model. PAPI aimed to be a data-interchange specification for conforming collaborative systems. Data can be exchanged via external specification, control transfer mechanism to facilitate data interchange, or via data and control transfer mechanisms. In the first case, only PAPI codings are common for any method of communication that is subject to being chosen jointly by data exchange participants. In the second case, PAPI API is used to exchange the data. In the final case, PAPI protocols are used [PAPI.]

The PAPI specification contains elements for recording knowledge acquisition (from coarse- to fine-grained), skills, abilities, learning styles, records, and personal information. This specification allows these elements to be represented in multiple levels of granularity, from a coarse overview down to the smallest conceivable sub-element. The standard allows different views of the Learner Model (learner, teacher, parent, school, employer, etc.) and substantially addresses issues of privacy and security.

PAPI provides logical categories of learner information, which provides separate security and administration of several types of learner information. Specifically, learner information is divided into six types of information: (1) personal information, e.g., name, id number; (2) preference information, e.g., useful and unusable I/O devices, learning styles, physical limitations; (3) performance information, e.g., grades; (4) portfolio information, e.g., accomplishments and works; (5) relations information; (6) security information.  These are also known as "learner information" and "learner profiles." SCORM

The Department of Defense (DoD) established the Advanced Distributed Learning (ADL) Initiative to develop learning and information technologies focused on  DoD needs. ADL aimed to build up a strategy for using learning and information technologies to modernize education and training in DoD-wide, which has existed for a long time. The ADL initiative produced high-level requirements for learning content such as content reusability, accessibility, durability, and interoperability. In this way, technology-enhanced education investments on a sound economic basis will be possible on a DoD-wide scale [SCORM].

The major work of ADL, to promote reusability and interoperability, is the definition of a reference model for sharable courseware objects. The existing reference model definition reflects ADL requirements. On the other hand, mapping with other existing learning models and practices is possible through common interfaces. Interaction with other systems needs further common data definitions and standardizations. ADL’s effort, known as the Sharable Object Reference Model (SCORM), is in its early stages of test and evaluation at this time and is subject to further revisions with respect to feedback from uses in courseware management systems and development tools. Though our architecture does not directly include content development, we keep compatibility issues with the SCORM model in our design for highly possible future interactions.

On the other hand, SCORM is focused mostly on computer-based tutoring instead of on a more flexible collaborative learning environment. However, they also included metadata for describing the course content originated from the IMS Learning Object Metadata mentioned below, which makes the course content useful to other systems. IMS

The Instructional Management Systems Global Learning Consortium, Inc. (IMS) is building open specifications for assisting online distributed learning activities such as locating, transferring and using educational content, tracking learner progress, reporting learner performance, and exchanging learner records between administrative systems. Since launched in 1997, the IMS Project has become one of the leading learning technology centers. IMS promoting finds a rather wide range of audience, both in developers of Internet-specific environments (such as Web-based course management systems) and off-line electronic resource environments.


The IMS Project has two main objectives: (1) preparing the technical specifications for interoperability of applications and services in distributed learning environments; (2) supporting the incorporation of IMS specifications into available products and services worldwide [IMS].


IMS Learning Object Metadata (LOM) is based on the IEEE LTSC Learning Object Meta-database. The LOM specification describes learning content cataloging information such as the names, definitions, organization, and constraints of IMS meta-data elements. This specification describes technical interoperability in terms of functionality, conceptual model, semantics, bindings, encodings, and extensions.  Interoperability can be measured via conformance testing [LOM99].

LOM defines a learning object as any entity, digital or non-digital, that can be used, re-used or referenced during technology-supported learning. Examples of technology-supported learning applications include computer-based training systems, interactive learning environments, intelligent computer-aided instruction systems, distance learning systems, web-based learning systems and collaborative learning environments. Examples of learning objects include multimedia content, instructional content, and instructional software and software tools that are referenced during technology supported learning. In a wider sense, learning objects could include learning objectives, persons, organizations, or events.

LOM standards focus on the minimum set of properties needed to allow learning objects to be managed, located, and evaluated. The standard accommodates the ability for locally extending the minimum set of properties. The LOM group enumerated their purposes as follows: To enable (1) learners or instructors to search, evaluate, acquire, and use learning objects; (2) to share and exchange learning objects across any technology-supported learning system; (3) developing learning objects in units that can be combined and decomposed in meaningful ways; (4) computer agents to automatically and dynamically compose personalized lessons for an individual learner; (5) to document and recognize the completion of existing or new learning and performance objectives associated with Learning Objects; (6) a strong and growing economy for Learning Objects that supports and sustains all forms of distribution: non-profit, not-for-profit and for profit; (7) education, training and learning organizations (including government, public and private) to express educational content and performance standards in a standardized format that is independent of the content itself; (8) to complement the direct work on standards that are focused on enabling multiple Learning Objects to work together within an open, distributed, learning environment; (9) to provide researchers with standards that support collecting and sharing comparable data concerning the applicability and effectiveness of Learning Objects; (10) to define a standard that is simple yet extensible to multiple domains and jurisdictions so as to be most easily and broadly adopted and applied; and (11) to support necessary security and authentication for the distribution and use of Learning Objects.



QTI Information Model

The IMS Question & Test Interoperability Information Model  is focused on the interoperable data structures between the learning systems, particularly between the question  and test systems. Moreover, those systems are planned to be Internet-based. The key data structures of the IMS Model consist of hierarchical structuring of question and test units. Those data structures are modeled using the object-oriented Unified Modeling Language (UML) and XML binding is followed by.

The main goal of the specification is to make possible the importing and exporting of question and test content between the learning systems. Basically, questions may be grouped in one data hierarchy and tests in a higher data hierarchy containing questions and assessment information in certain definitions of IMS data specification

The IMS Model quickly turned out to be a set of complex detailed specifications. It may not be easily possible for any application to fully conform to the IMS Model, which seems to be a drawback of IMS widespread usage or a reason for future shortcomings of IMS-centered systems.

Currently, the IMS QTI model is in its revision stage, and a second version is expected in the future.

2.1.2 High-Level Learning Technology System Architecture Model

The Learning Technology Standards Committee (LTSC) works for specifications within subject areas, which cover reference model and learner model standards, learning objects, task models, course sequencing, tools/agent communication, ontologies, data interchange, course management, metadata, and student identifiers. The Learning Technology Systems Architecture  (LTSA) is “a high-level architecture and layering” for information-technology-supported learning, education, and training systems, and a proposed architecture specification for the IEEE 1484 LTSC started in 1997. It describes the high-level system design and its components. The LTSA specification covers a wide range of systems commonly known as learning technology, computer-based training, electronic performance support systems, computer-assisted instruction, intelligent tutoring, education and training technology, metadata, etc.  The LTSA specification is pedagogically neutral, content-neutral, culturally neutral, and platform-neutral.  The LTSA specification (1) provides a framework for understanding existing and future systems, (2) promotes interoperability and portability by identifying critical system interfaces, (3) incorporates a technical horizon (applicability) of at least 5-10 years, while remaining adaptable to new technologies and learning technology systems.  The LTSA is neither prescriptive nor exclusive [LTSA].

LTSA explores high-level frameworks for the above-mentioned system types, their subsystems, and their interactions. LTSA is a source of more than one architecture, and their analyses, designs, interoperability and service identifications, communication, cooperation, collaboration. Many systems can satisfy the requirements of the LTSA, although they don't provide all the components, have different organizations or different designs. Figure 2‑1 illustrates the high-level architecture of LTSA and its system components [LTSA].


Figure 21 The LTSA architecture system components.


Actual systems are generally not designed as the clear divisions specified as LTSA components. Commercial, business, or technical reasons are always effective in systems’ organization. LTSA provides guidelines to learning technology systems designers, and its conceptual system components can be mapped to virtually all actual learning technology systems already designed or naturally practiced. These include a father-and-son talk, traditional classroom teaching, self-study, Web-based learning, flight simulation, etc.

2.1.3 Current Practices WebCT

The World-Wide-Web Course Tool (WebCT) is one of the initial LMS systems appeared ob open interfaces starting in 1996.  WebCT is a tool being developed at the University of British Columbia presents an environment (courses) for authoring and delivering educational material by educators, with or without technical expertise. These courses can incorporate an array of tools (continually growing) to enhance the on-line learning experience. A subset of these tools is intended to facilitate communication and collaboration. This includes a conferencing system, group presentation areas, a synchronous chat system, and electronic mail. While there are other examples of similar tools individually, WebCT assembles these tools into a single package that can be used by course facilitators without technical knowledge to create sophisticated web-based course environments. WebCT also adds features to these tools that are of particular utility in an educational setting [WebCT97.]

WebCT is entirely WWW-based, both for the student and for the course designer. A single web server running the WebCT software is used both for course creation and delivery. Students wishing to interact with course material, and designers wishing to create new courses or modify existing courses use a web browser to connect to the WebCT server. The content of a course is provided by the course developer or designer. Interactivity, structure and educational tools are provided by WebCT. WebCT, as well as allowing the incorporation of a set of educational tools in a course, also allows for the manipulation of the layout and look of the course. WebCT provides a set of educational tools that can be incorporated into any course. Examples of these tools follow

·        Content Tools: course content indexing and searching, searchable image archive (which allows the addition of annotated images to a course,) presentation tool (that allows the course designer to determine the layout, colors, text, counters, etc for the course pages.)

·        Study Tools: Searchable and linkable glossary, course calendar, reference tool (References, for example text book or URL, can be added to a course and be customized for any page of content,) student homepages, presentation area, page annotation, self-progress track (Allows students to see their individual course content visiting tracks.)

·        Evaluation Tools: Timed on-line quizzes, self-evaluation tests, grade distribution, progress tracking (Facilitator can get complete information regarding the accesses of individual students, and the access trends for each page of course content contained in WebCT. Statistics are usually based on simple countings.)

·        Communication Tools: Examples are chat, discussions, mail and whiteboard.


WebCT architecture was initially designed on the trust of CGI mechanism at the server side by omitting database back-end. These refrained their development until recently though it has become very popular. Most of the functionalities are similar to our VCM. Recently, they added some more functionality like surveys, assignments, student presentations, student tips, bookmarks to course content, CD-ROM support etc. Blackboard

Blackboard is a learning management system that provides course, course content, user and portal management services. All the users of the system can access information using Web browsers after authenticating themselves with a login and a password pair. Instructors and students first see their list of courses either teaching or taking. After selecting a course, users are provided a navigation menu, which is the same menu shared by instructors and students except a button for the Control Panel through which instructors have the ability to create course content, and alter system, portal, or students’ usage settings.

The design is mainly concentrated on an asynchronous teaching environment. Any document can be either uploaded into a shared file area, or submitted through a form with text areas to be posted inside the course content. The latter must be plain text or HTML content. Besides online document sharing, it provides discussion boards, e-mail tools, and a home page preparing wizard as well.

Blackboard environment provides a third-party software, which enables synchronous teaching. This software has very limited features including a whiteboard, image or Web page sharing on the board, a chat tool, and a question and answer history box. Blackboard also provides users a calendar, the instructors’ being the master. This means that students can update their calendar while they can also see the additions instructors make. For instructors, the Control Panel is the major tool through which they can develop course content as well as prepare assessments, post announcements, and alter system and user settings. Portal manager is also very useful feature. Instructors set up the sources, which feed information to the portal such as news sources and search tools. Any related Web content can be indexed into a nice and handy page. However, portal is only customized for each course rather than for individuals.

The software looks very easy to use, and has many features, such as discussion boards, class- or group-wise file and message sharing, assessment tool, needed for an asynchronous learning system. On the other hand, the synchronous teaching tools are very immature. The content management may limit this product in the future since it is mainly based on links to uploaded HTML pages that make users slightly restrictive. VPL

An NPAC research that has culminated into the Virtual Programming Laboratory (VPL) [VPL97ConcJ], that is a Web-based virtual programming environment based on a client-server architecture. It can be accessed from any platform (Unix, PC, or Mac) using a standard Java-enabled browser. Software delivery over the Web imposes a novel set of constraints on design. It outlines the tradeoffs in a design space, motivates the choices necessary to deliver an application, and details the lessons learned in the process. Additionally, VPL facilitates the development and execution of parallel programs. The initial prototype supports high-level parallel programming based on Fortran 90 and High-Performance Fortran (HPF), as well as explicit low-level programming with the MPI message-passing interface. Supplementary Java-based platform-independent tools for data and performance visualization are integral parts of VPL. Pablo SDDF trace files generated by the Pablo performance instrumentation system are used for postmortem performance visualization.

VPL is exceptional work that was based on emerging Web technologies a couple of years ago, and it is still a very good Web-based system that anybody can assume as a starting point. We integrated some of VPL functionalities together with VCM as well as its independent uses on various courses. Synchronous Delivery

A number of applications, systems and tools have been developed to answer synchronous needs in general. Some of them are used in distance education, others are applicable for such purposes including delivery of course content and various other interaction tools such as chat, whiteboard and real-time video and audio. Examples of such systems include TANGO Interactive, WebEx and Centra among others.

TANGO Interactive is a Java-based web collaboratory developed at NPAC. It is implemented with standard Internet technologies and protocols, and runs inside an ordinary Netscape browser window. Although TANGO was originally designed to support collaborative workgroups, in this project it was used to synchronously deliver course materials stored in an otherwise asynchronous repository. The primary TANGO window is called the control application (CA). From the CA participants have access to many tools including: WebWisdom ( a presentation environment for over 400 foil sets), SharedBrowser (a special-purpose browser window that "pushes" Web documents onto remote client workstations), WhiteBoard (for interactive text and graphics display), 2D and 3D Chat tools, RaiseHand ( a tool used to signal one's desire to ask a question), BuenaVista (for two-way streaming audio and video)[Jackson98.]

We successively used TANGO tools and VCM to provide both synchronous and asynchronous environment to Learner, who are students at Jackson State University (JSU) in Jackson, Mississippi and Supervisors, who are faculty and teaching assistants both in JSU and Syracuse University, Syracuse, New York, during the computational science course delivery over the Internet in 1997 and 1998. Our initial ideas for interoperability of various supportive systems are started during those experiments.

Centra is real-time collaboration framework, and optimized for different business interactions, including virtual classrooms, Web conferencing, and eMeetings. Centra provides education and training programs to assist curriculum developers, content developers, instructional designers, instructors, system administrators, and other business professionals in their use of Centra's products and online services.

WebEx provides real-time global communication services with integrated voice, data and video communications using any web browser or telephone that can be easily integrated into any corporate website. Further, the WebEx architecture enables a comprehensive set of functionality with maximum scalability and the fullest extensibility. WebEx has built a unique communications infrastructure based on "information switch" technology.  This technology is analogous to telephone switching systems and enables true real-time interactive communication sessions that combine voice, data and video.  This platform offers deep communication functionality, solid reliability and massive scalability and differentiates WebEx from competitors who offer database-centric communications that rely on a store-and-retrieve paradigm [WebEx.]


2.2      A Taxonomy of Distance Education Components

There are numerous approaches to interpreting and implementing distance education architectures. However, based on our experience we collected all the distance education technologies in the following taxonomy:

Content preparation: The tools to prepare and organize course contents, e.g., lessons, are classified in this category.

Information management: All the management information related to records of Learners, Supervisors, courses, quizzes, etc., in collected into this category. Learning Management Systems (LMS) mostly behaves as information management environments. Information management is  usually performed asynchronously, though some services may be performed synchronously between the partners.

Delivery: In distance education, delivery of course content is especially essential. Delivery can be done both synchronously and asynchronously. The delivery tools become more important when it is performed synchronously. A number of remarkable tools and systems have been developed to fulfill needs, such as TANGO, WebEx, Centra, etc.

2.3      Web-Mining Technologies

The accelerated usage of the Web has led many organizations to analyze web users' surfing behavior. Huge investments have been made to operate sites on the Web. The immediate benefits provided by an organization's online site need to be explored by collecting precious customer-visiting data. While improvement of the effectiveness of such investments is one concern, discovering new ways to extract potential opportunities becomes another concern. As a consequence, the understanding and using of user surfing behavior, i.e., Web Usage Mining, is important today for every organization running a web site.


There are a number of log analysis tools available for what appeared as an immediate need. However, many of these tools do not care about individual user behavior or relationships among the accessed pages, but are mainly useful for  viewing raw Web-server statistics, including total number of visits to a specific site, locations of the visitors, the peak times of visits, etc. Some examples of this kind of tool are wwwstat, Analog, SuperStats, Open Market Web Reporter and  Webtrends.

More sophisticated data mining tools begin with clarifying user sessions. Among well-known tools, one of three techniques is favored for identifying user sessions: login-ids, cookies, or heuristics on the host addresses. Some examples are WEBMINER, IBM's SpeedTracer, SILK, SurfReport and NetTracker Usage Analyst by Microsoft Corporation. After user sessions are identified, most of the analysis tools behave similarly. Straightforward statistics based on the counts of visits and/or visitors are always popular at first glance. Data-mining algorithms can be applied to further investigate user behavior or to discover navigation patterns, including the most common traversal paths and groups of pages frequently visited together. The analysis tools need to be further improved in specific areas, as has already been done to discover customer-buying patterns.



Chapter 3        

Multi-tier Architecture of the Virtual Classroom Manager


The Virtual Classroom Manager (VCM) is a software system that provides management of courses given both on and off campus. The VCM provides an asynchronous, distributed, open information access and management environment to particular users, including supervisors and learners. The VCM work is a confluence of several research ideas. As a natural result, particular aspects of its design, functionality, and implementation appear in other types of tools. However, the architectural design of VCM has unique properties that differentiate it from the others. This chapter describes the motivations behind VCM and presents its architecture and implementation details.


3.1      Motivations

We observed that the WWW technologies provide a standard high-level open interface for the delivery and access of information served by distributed Web servers. This has opened opportunities to develop more sophisticated software systems. Technological improvements have enhanced our range of educational experience with such things as distance learning. A remarkable development of pragmatic learning technologies has been witnessed in education in the second half of the last decade. For example, distance learning paradigms are involving both synchronous (such as online distance lectures over the Web) and asynchronous (such as independent learning from web materials) techniques. In distance learning courses, collaboration tools, mailing lists, and bulletin boards are used, and a rich set of online materials and reference links to the World Wide Web (WWW) are presented to students.


The computational science education group at the Northeast Parallel Architectures Center (NPAC) has developed a huge repository of online course material that includes lectures, tutorials, and programming examples in various languages. To provide synchronous interaction with students involving teachers and other learners, in addition to asynchronous learning materials, the TANGO Interactive collaborative system is used to deliver classes over the Internet [TANGO]. One major use of this system has been to deliver Syracuse University academic courses with instructors in Syracuse, New York, to Jackson State University students in Jackson, Mississippi [Jackson98].


All the immediate improvements in technology-based education methods brought new needs. Our technical and educational experience showed that the new face of learning should be supported more elegantly by evolving commodity technologies into a more advanced architecture design. We believed that we should go out of traditional two-tier client server architecture understanding. The requirements of the user community had driven the evolution and realization of our multi-tier architecture. The later-developed architecture models of commodity technologies proved that our approach is right.


On the other hand, in addition to architectural design flaws of current systems, the new education environments made traditional educational methods subject to revision. For example, educators can no longer continue assessments with methods that need human interactions within traditional systems. In a collapse of previously completely different fields, we designed technological abilities to shape the assessment world of educators by integrating several research-field approaches in computer science, such as data mining, into a more general architecture. Many types of distance education involve the learner’s asynchronous remote access to on-line course materials. Basically, using the online resources asynchronously reflects the student’s learning abilities, which opens up additional possibilities for assessment that use access patterns to the materials. Analysis of how students are using a site is critical for assessing both the students and the course materials. Current education experience does not provide the ability to evaluate students’ informal responses to the given materials.


A new architecture for assessment over the WWW is necessary for today's innovative distance-education technologies. This architecture should involve the collection of students’ access data to web resources, transformation of the collected data to assessment information, integration of personal registration information of students and continued performance data collected from surveys, quizzes, and assignments in the final evaluations, as well as the processing of all the data and information available with the recent data mining techniques.


More important than an assessment architecture is the architecture supporting management of the courses, supervisors (instructors, etc.), learners, performance records, grading of assignments and exams, legacy systems, and security. Because of the advance of technology applications in education, an asynchronous collaboration between learners and educators became crucial. Furthermore, continuous open access to personal information and other records from distributed access points through the help of WWW is essential.


As long as we use technological advancements to build or to modernize learning environments for users, we may encounter new needs and new architectures supporting them, such as the two we mentioned above. We observed that a strong correlation is an absolute necessity between the implementations of these architectures. Even if internal architecture design and implementation tools differ, all the final environments should support a strong communication with each other. As emphasized in the recently finalized LTSA by the IEEE Standards Committee, learning systems can emphasize different aspects of a large architecture even if the entire set of components is implemented [LTSA].


Common language(s) between various systems may lead to their further utilization, both in functionality and in usability. An existing system with a narrowed scope can boost its functionalities in a world of other entire existing systems. For example, an individual assessment system has much value if its feedback is interoperable with a synchronous delivery system or an asynchronous information management system.


Interoperability would also cause positive side effects such as reusability of resources. An information resource produced in one component should be able to be used by other components in different architectures. Furthermore, a strong architecture of today will be usable tomorrow, even if it turns out to be a legacy code.


On the other hand, we believe that interoperability and reusability are not sufficient for tomorrow’s software world. Already, we have a remarkable number of information resources distributed on the Internet, but as long as we are not aware of the sources for useful information, we cannot use them no matter how easy it is to integrate our systems with these resources. This brings us to two other necessary properties of future software architectures, which are discoverability and accessibility.  Certain ways to easily find resources, such as reusable web objects through services, i.e., discoverability, and the mechanisms to be able to access these resources should be defined. Accessibility of resources covers a wide range of concepts from authentication to the cases of online availability of remote resources, reference changes, agreeing transfer protocols, line bandwidths, etc.


As a natural result, all the software components and their information resources should be manageable in an architectural environment. For example, those matters pertaining to education should be especially reusable and manageable in other architectures. Any learning software design product can be part of another architecture, or at least it should be possible to easily integrate the information output produced by this product into the new architecture.


From this point of view, we targeted  the interoperability, reusability, discoverability, accessibility, and manageability properties of a software system architecture.

As we decided to design a learning system architecture to apply to these targets, we observed that architectures of the existing work were rather straightforward for carrying out specific purposes. Most of the learning systems started with a two-tier architectural view, i.e., client-server architectural view.  Our belief that future systems should be designed as multi-tier architectures is being realized by the computing community of researchers and industry.


A multi-tier architecture is innovative because of its natural ability to produce object services in an open environment. Having next-generation objects in an open environment, we can provide the above-mentioned rich set of properties within a software environment. We believed that easily discoverable and accessible, reusable and manageable, interoperable, integrative and expandable object services would drastically change next-generation software development and be a significant milestone in the computing world.


One may argue that the programming effort for two-tier client-server architectures is much easier, and that a two-tier architecture is capable of achieving what can be achieved in a multi-tier architecture. However, today’s software packages have already found a huge audience. A strict two-tier system does not scale well when thousands of clients are awaiting service. One technical advantage of multi-tier systems is that they are capable of multithreading on middle tiers instead of using fatty server code or client code. In two-tier systems, mostly client applications are loaded with additional service codes as well as service delivery  codes such as user interfaces, i.e., fat clients. The only other option is, similarly, to have a fat server, which is more inconvenient since the server also carries the main management of the software system. In contrast, in multi-tier systems, while the middle tiers handle decisions of transaction policies and the allocation of resources and coordination between various services, the back tiers carry the low-level service actions, and clients become responsible mainly for the rendering of results.


With  this perspective, we designed a multi-tier architecture with learning web objects support, which is our main experimentation area. The next section describes the design issues of our architecture, which is implemented using recent commodity technologies. The final system is called Virtual Classroom Manager, i.e., VCM.

3.2      Multi-Tier Architecture

The VCM system has a multi-tier architecture. Conceptually, the architecture tiers can be seen as having clients at the front end, services at the middle tier, and data storage at the back end in a three-tier architectural view (Figure 3‑1.) However, the middle tier should be considered as a multi-tier, where after the web server, the first tier is handling client requests and managing the system at a high level, and the second and third tiers are responsible for producing web objects by communicating with the storage devices of the back tier.


Figure 31 Multi-tier architecture of VCM

The client site gets services through distributed open interfaces.  As its nature, the web server provides services to distributed users. The rendering of information is dependent upon the abilities of the client. In our application case, we expect the client  to be able to handle common web pages. However, the object broker in the middle tier can leave the entire rendering process to the client, which will allow other less capable devices to render the information according to their capabilities.

 The middle tiers provide access to conventional back-end services such as relational databases or other file systems. The system manager in the middle is responsible for handling all requests, coordination of services and, finally, for producing responses to the clients. The system manager works at a rather high level and does not deal directly with back-end services. The application layer, i.e., clients perception of services, is close to an abstract  module, which does much of its work through object services working with high-level object request brokers. The object services comprise the main part of the middle tier. An object service is connected with the related object broker. Each service handles different aspects of the system, based on the underlying object scheme. The object brokers are responsible for keeping and serving well-formatted objects, which may come from object providers. The object provider works at a rather low level, communicates with back-end devices, gets data and converts them into an object format.


Communication with the back end is provided by appropriate bridges present in current commodity technologies. For the most part, a bridge to a relational database is used, yet bridges to legacy system codes are still present today. We have a limited number of legacy system connections in our architecture.


Our approach is very compatible with modern industry’s view of commodity applications. Numerous applications have been developed to offer open interfaces for accessing “traditional”  back end services in the company of the availability of standard interfaces with a middle tier. At the very beginning, file systems were popular together with emerging commodity applications and, later, relational databases again became popular in commodity computing. However, some large applications got stocked with file systems and could not evolve well later.

On the other hand, the distributed nature of Internet technologies causes fast evaluations in current architectures. We believe that the “Pragmatic Object Web” mixture of distributed services and distributed objects will be a common characteristic of the web. In our architecture, the object providers help to convert the back-end relational data as information in the form of  web-objects. The “High-Level Object Request Broker” component at the middle tier helps serve the web objects through today’s most common access way, i.e., web servers, on Internet architecture.

In our case, the web objects served to the outside world are “Learning Web Objects.” Currently, the existing learning systems and architectures have failed to follow even the rather old-fashioned, three-tier commodity architectures. However,  considerable efforts are being made to standardize the interoperability learning technology applications. Without doubt, the Web is the best place to develop next-generation learning systems. Our distributed web-object-emphasized research  architecture was targeted to be a prototype of those systems. More characteristics of the architecture are shown (Figure 3‑2) which elaborate the primary architecture (Figure 3‑1).


Figure 32 Distributed service distributed objects architecture of VCM.

Although it is possible to connect directly to another back-end database in our architecture, it is not recommended. Rather, accessing information through the open interfaces of another management system is the more elegant solution in presently developing modern architectures. Well-designed management systems are responsible for providing Web-Objects in a common syntax and for understandable semantics using recent commodity technologies. In our architecture we serve the web objects through the object services in the middle tier to the open interfaces offered by web servers. We do not require the compatibility of system architectures to use other systems’ objects, but those structures of objects made available by numerous methods. Of course, the methods for discovering and reaching those objects should be integrated into the system for specific cases.

As an illustration, a learner may transfer from one university to another. The learning system in the new university may keep track of his performance records; however, to better analyze the student’s abilities and progress, the performance records of the previous university may be necessary. In this case, instead of physically transferring all of his records, a Web Object containing performance information of the student can be directly questioned through the open interface of the previous university’s open learning system by providing the necessary authentication information. Another example can be the buying of a questionnaire on a specific topic from a well-accredited education site.

3.3      Implementation Technologies and Issues

The last decade’s Internet technologies tremendously affected our computation styles. While we were developing our system, we observed a  remarkable software environment of unprecedented quality and functionality  being  developed by  current industry, as well as pretentious contributions of  the loosely organized worldwide collection of commercial, academic, and freeware programmers. We have always benefited from this software environment and have experimented with different implementation strategies and tools. The important lesson we have learned from these technology advancements is that future systems will have largely independent and distributed clients and will include final users and other application services connected to a distributed collection of servers hosting certain services.

We started our research with a much simpler architecture (Figure 3‑3) as a result of emerging Internet technologies. In our very first architecture, we used a web server to connect the back-end database to the outside world through commodity interfaces. The connection between the web server and the database, i.e., Oracle, was provided by a CGI supported by both Oracle and the web server in a corresponding WOW (Web-Oracle-Web) Gateway.  More specifically, the CGI script, called WOW, and the Oracle bridge, i.e., wowstub, established the database connection and marshaled data to the related PL/SQL package and function and, finally, the output was sent in a backward direction. Alternatively, another Gateway of Oracle, called OWA, embedded the CGI mechanism inside the Oracle’s own web server, called the Oracle Web Application Server. In this architecture, the middle tier was very thin, and as a result the  back tier was thick. Every computation and most of the interface preparation were handled by a back-end database management system. The PL/SQL code modules were responsible for producing HTML pages using relational data. The middle-tier components worked as a simple bridge to reach the PL/SQL modules. The front site web browsers were needed only to render an HTML page, which was enriched by JavaScript functions.


Figure 33 Starting architecture

 This architectural view became common within emerging commodity technologies. It answered their needs with a fairly lightweight client to distribute cross-platform. It was more elegant, considering the stateless CGI-Perl environments. PL/SQL is one of the most suitable languages for developing complex database-enabled web applications. Although this architecture was rather common in enterprise computing, and still has large use in industry, it has some shortcomings.

The architectural philosophy is that of a client-server even though it has a three-tier skeleton. One main drawback of the system is that its scalability is strictly bounded by one database server at the back end. Since the database server is also responsible for management of the system and for rendering issues, its performance is subject to doubt. Another issue is that the computation carried out at the back-end does not allow the use of distributed web objects of the later systems. The system is stand-alone and is not intended to collaborate with others in a distributed computation environment, though its services may be used as legacy services in one-way collaboration. Another drawback of the system is that when using WOW, the database user account login-id and password are kept in the script, which is a text file. This causes an easy-to-penetrate security hole in the system.


The common modern industrial view of commodity applications is an architecture of three tiers. Our latest architecture is a multi-tier architecture (Figure 3‑1) based on supportive technologies of this architectural view.  We believe that the three critical technology areas are the Web, distributed objects, and databases. Their linkage will provide bases for the object-web technologies of the next generation. Numerous commodity areas bring into being remarkable software artifacts that are subject to continuous evaluation. The fundamentals of these artifacts are provided by lower-level standards and tools such as client technologies, HTML, DHTML, JavaScript, ActiveX, VRML, applets, etc., communication protocols, e.g., HTTP, CGI, MIME, IIOP, TCP/IP, etc., programming languages and environments, e.g., Java, JavaBeans, CORBA, COM, RMI, JDBC, SQL, and servlet servers. They are followed by data representation format and frameworks such as XML, XSL, DOM, XHTML, and RDF. On top of these technologies, some of the many applications and services include collaboration, security, electronic commerce, multimedia, etc. All of these technologies are currently leading distributed pragmatic web-object services that are available through a set of open interfaces at each level of the various technologies.  Modular software development in a distributed powerful software environment is popular and has been researched in the Gateway project by NPAC [Erol00.] We experimented with and chose different technologies at each tier with which to make the final system a prototype of the emerging future generation of software systems.

The richness of commodity technologies at present provides developers with various ways in which to use a specific technology, and also how to use a mixture of technologies in a specific job. Our main purpose, while emphasizing a specific technology in our implementation, was to construct a prototype of our  architectural view instead of to develop pointed solutions.

In the middle tier, we selected pure Java implementations. Java is an object-oriented language and is currently the best supportive language of object web technologies.  Using Java provides us with various possible forms of service objects, i.e., pure Java objects, Java DOM objects, or XML objects. Having XML and Web servers at the middle tier leverages the customization and installation of distributed objects and their services on top of the Internet backbone. We observed that various profound framework arenas such as CORBA, COM, RMI and Enterprise JavaBeans appeared in the distributed-object technology arena. Each of these frameworks has specific advantages with respect to others, depending on the platform and priorities used. On the other hand, we learned the lesson that it does not matter which distributed object technology is used as long as an efficient communication and integration mechanism can be provided between these notable technology artifacts. Each distributed technology has its pros and cons. Furthermore, numerous artifacts are already established using different distributed technologies. Instead of forcing a specific technology on the applications, XML provides integration among various technology standards when their proper objects are specified in XML. Various commodity tools supporting Java and XML are in use today, thus, using them together is helpful, both in developing and in supporting Object-Web.

Besides the proper nature of Java within our context, the language is also a good choice because of its cross-platform applicability. The same software implementations can be installed on any machine without any single-line editing, in theory.

Figure 34 Implementation view of multi-tier architecture

An implementation view of the architecture is shown in Figure 3‑4.  The middle-tier services are connected to the back-end servers through appropriate bridges as the JDBC-bridge connects to the relational SQL database. The back-end databases are old technology and not appropriate for today’s distributed-object paradigm. On the other hand, they serve well as a storage mechanism for middle-tier services. To live in a web-object world, the stored data in a SQL database should be converted into more informative and complete objects by the Object Providers. In our case, we represent web-objects at the middle in the format of CML DOM objects. At the back end we used Oracle and mSQL databases in our research. Although object providers can reach any relational data storage devices through a generic JDBC bridge, having connections at this lower level does not reflect next-generation systems. At a higher level, object brokers can connect to other distributed services to get XML objects using available protocols like HTTP, DCOM, IIOP, etc. Although we did not implement all these technology protocols, they can be integrated into the system as small object broker modules. Our implementation of object providers should be seen as just another  module that is using a legacy protocol, i.e., database network.

Performance is an important issue in distributed web object technology when providing a remote object, or even when querying a database multiple times to construct a single information object. Therefore, an object cache mechanism is included in our architecture. The consistency of the cache level has been a well-known research issue for years. Even if we need to check each time to see if it is updated in originating repository, we do not need to transfer whole object data each time as we do with the web browsers’ caching mechanisms.

Other than the relational databases at the back end, an application may need file system storage for XML/XSL file objects, various other legacy system services, or system access logs on the file system. We keep XSL and HTML template files in the template library. Our previous experimentation was based on HTML template files, which are extremely easy to edit, but not powerful enough to handle all the rendering of XML objects. As an alternative, XSL templates have become more useful with their rich functionality for rendering XML despite their complexity for the programmer. Furthermore, the emerging tools have made them easy to edit for non-professionals, too. The middle-tier services return objects after parsing with available templates in the library. In addition, to serve as part of another system, they may serve the plain objects allowing the requester to parse and render the object.

Legacy systems are usually part of modern architectures until they are replaced with better implementations. To illustrate this, we used a Majordomo mail server activated through legacy CGI-Perl scripts.

The front site of the middle tier consists of a group of services and an abstract manager of all the services.  Each service is specialized to serve a certain kind of object. The functionality of service objects is coupled with the object types in the system context. All the service objects are written in Java. These Java service objects are coupled with servlets, which provide the interaction with the web server. The servlet technology is developed much later than a CGI mechanism. Among several, one important advantage to using servlets is that they are stateful objects and allow object-oriented development at the middle tier. 

Depending on the request type, a service may provide an XML web object or  its HTML page format to the front tier. In the case of an XML object service, the client is responsible for how to render data. A client can be a web browser with advanced functionality, a hand-held device with limited capacity, or another system service located in another place in the network. The XML object representation of the information gives flexibility to the client in choosing how to manipulate information with respect to its own needs. For general purposes, HTML page service produces an XSL or HTML template-parsed HTML page format of the information. The client is a regular Web browser in this second scenario.

JavaScript enriches HTML pages at the front end. JavaScript and possibly Java applets and DHML objects, provides dynamism to HTML pages. Advanced user interfaces are possible with these developments tools. JavaScript is fast and provides much functionality to the represented text. Java can produce much professional interfaces with the sacrifice of the speed. DHTML and JavaScript together can utilize the browser’s capabilities in almost the same professional manner as Java or other interface tools.  Furthermore, using images or emerging GUI tools like Flash [Shockwave] has turned web tools into virtual showrooms. One essential GUI property of next-generation systems is that they can represent the same information in many different ways with respect to users’ preferences in such things as templates and skins. Our template library allows the inclusion of more than one template to be chosen by the user for each information page.

An important issue of the emerging distributed web object architectures is providing security. In our architecture the communication channel is secured with SSL, i.e., it has a secure web server. To provide front-end presentation security we used JavaScript. Caching of secure documents is prevented in most popular browsers unless users directly request it. In addition, our system provides a new key to users each time they log in, which prevents using the same URL for the second time. Furthermore, a user authentication allows only one session to be kept alive at a time.  The very back-end security of the system between middle-tier object providers and  storage devices is subject to question. The current solution is to keep both servers on the same machine behind a firewall. We left secure communication at the back end to the vendors. For our purposes, the current way is sufficient, and we concentrated on other aspects of security. More details will be discussed in Chapter 7.







Chapter 4        

Middle-Tier Service Components

The Web infrastructure began with a number of early engineering decisions, some of which became awkward for the large-scale applications that were developed later.  One example is the stateless CGI mechanism; another is that there was a tendency to have a document-centric collection of non-interactive servers. We believed that we could shorten these limitations of the current infrastructure by using recent web technologies and tools. At the same time, we could still use this extraordinary infrastructure encountered in the history of technology. The middle-tier service components in our architecture explained in Chapter 3 are the realization of our beliefs. Our services combine the benefits of distributed object technology and the Web infrastructure in the form of distributed Web-object services. The aim of those services in our architecture is not to isolate themselves from other distributed infrastructures such as CORBA and DCOM, but to utilize them as needed through the integration of proper adapters as a new set of services.

The trend for applications presently being developed in general areas is to become more functional and powerful by integrating inside existing applications or services. If an easy standard way of integration can be provided, we may witness a much bigger boom of application development with open interfaces. Architectures such as CORBA have already made efforts to gain conventional libraries, though we need a much broader integration model for providing simplicity and usability without sacrificing performance and up-and-down scalability.


Distributed Web-object services overcome the failures of centralized applications in many cases by providing us the ability to share information across applications through open interfaces. This allows us to distribute some application development to other vendors as well as to offer more computing power  across a network of computers.

The web, with its distributed web-object services, is an accommodation space for an unpredictable growth. Specific services on one dedicated machine can be performed more efficiently, and a single and simple application at the client site that uses a set of services may utilize many resources on PCs, UNIX workstations, etc. A simple application manager on a UNIX processor may be in charge of activating a parallel program on a set of parallel processors. Furthermore, distributed web-object services can be modified, upgraded, or replaced by new ones without redesigning the main application, which uses a set of services.




4.1      Services

The capability to reuse learning components from other applications (and vice versa, i.e., to be able to produce reusable objects for other applications) are the biggest dreams of every software architecture designer today, who would like the architecture to be able to survive into the future. Therefore, we adapted our information model for different technology advancements. It is clear that not just one research group will be effective in all the areas of a learning management system. We tried to be compatible with different leading research groups in each service area as well as to develop our information model with respect to our own needs. Since in the best of scenarios most of the standards efforts are just now becoming finalized, our implementations are not necessarily final and may need future work. On the other hand, we give here the basic concepts of the architecture that allow it to be compatible with the learning management systems.

VCM has the following services, as listed in Table 4‑1.


Table 41 VCM Object Service List.

Service Name

Object Information and/or Functionality

Learner Object Service

Learner information such as student name, id, passwords, grades, etc. Management of student information.

Course Object Service

Course information containing general course metadata and system management.

Assignment Object Service

Information for assignments given in a course, and their management.

Quiz Object Service

Quiz and question information and management ( when quiz starts, how long it lasts, which questions will appear, question types, etc.).

Performance Object Services

Learner performance information and management containing assignment and quiz grades, assessment comments, etc.

Supervisor Object Services

Learning System supervisors’ information and management.

AssessmentNugget Object Service

Automatic assessment of learning progress from secondary indirect sources of information, e.g., Web logs. Usually, available through technology uses.


4.2      Service Objects

4.2.1 Learner Objects

Learner information is a significant part of the learning technology information managed by the VCM system. Having portable learner (students, etc.) records is one of the central issues in a learning system. In fact, the model for keeping public and private information about individuals is a common sensitive concern of many organizations, including government, universities, and industry. One of the main purposes of this research is to make personal information portable across the different learning systems.

The framework of our implementation is the PAPI specification effort. Indeed, PAPI specification is still not finalized. A PAPI learner implementation will represent human information for use by learning technology systems or communicate this information for use in PAPI learner applications. The standard effort tries to satisfy a minimal set of functionalities for wide acceptance and sufficient capability for usability in contrast [PAPI.]  In practice, however, there is no PAPI implementation at present, because the specifications are not final and have been constantly changing for the last four or five years. Even the original name, "Personal and Performance Information," has been changed to "Public and Private Information.” On the other hand, the general overview of PAPI is applicable to learning systems or any other such system needed by industry, hospitals, financial institutions, etc. We left a strict conformance with PAPI as future work. Our experiments with Learner information have parallels to PAPI’s efforts in its developing stage.

VCM has six major categories of information at various levels of support, based on the PAPI division of learning information into the following [PAPI:]


·        Learner personal information.  Information mainly for administration. Example fields are name, address, telephone number, etc.

·        Learner relations information.  Information related to other users of learning technology systems such as teachers, proctors, and other learners.

·        Learner security information.  Information consisting of the learner's security credentials such as passwords, challenges/responses, private keys, public keys, biometrics.

·        Learner preference information.  Information to improve human-computer interactions. This type of information may include the learner's technical, learning, and physical preferences, e.g., preferred, useful, and unusable input or output devices.

·        Learner performance information.  Information containing the learner's history, current work, or future objectives produced or used by system components at various stages.

·        Learner portfolio information.  Any learner work objects or references to them  supplementing the learner's portfolio of work.


Figure 41 Some applications of Learner Information categories in VCM

Figure 4‑1 illustrates the different categories of Learner information in VCM. In Figure 4‑1 (a) several Learner information categories are listed for the Learner to explore. The emailing post office presented in Figure 4‑1 (b) is a sample view of the Learner relations category. The emails of classmates and instructors with other staff such as teaching assistants are listed as a sign of close relations with the Learner in the context. Figure 4‑1 (d) and (e) illustrate how security information is prevented from being sent to others than Learner. Figure 4‑1 (c) and (f) illustrate two different views of performance information that are for Learner and Supervisors. Each Supervisor level has more colons at the right to see comments about the Learner, who is not aware of these comments. In fact, the last sample shows the same category of information; the performance information is presented at varying degrees of privacy and public availability. Both public viewers including Supervisors, and private viewers (Learner himself) are restricted at some level.


Building multiple categories for single users has various rationales behind it. For example, in a remote learning system the learner’s local security information has no value and also should not be transferred. During an interactive learning session, only the performance information is necessary for leading the experience. Because of the advance of distributed access to resources, it is critical that public and private information be available at different levels.

Sometimes the learner information is not only generated by the learning system manager, but also by the learner himself. For example, student homeworks are done by the student himself. Furthermore, this information is stored and kept track of by the student at an independent repository. To locate and use these works, a URL reference should  be kept in the portfolio information object.


Moreover, a finalized Learner object does not necessarily reflect a single information type. A learner object can be a collection of information in the existing learning management system, e.g., VCM. With respect to the request type, a related service can collect a varying number of information (Figure 4‑1).

The security issues of Learning Objects are important ones, both in VCM and in PAPI. We have left these issues for Chapter 7. Here we give only conceptual concerns related to public and private. Functionality

Note that the information presented   in a Learner Object may not be directly stored in the database. The Learning object can be finalized by querying the back-end database multiple times, or by using other system services. Based on the above divisions, personal information and performance information may be kept separately in repositories. Since each category covers different usage areas of information, and some information is private and secret, different access levels are necessary for the stored data. While presenting performance records, there is no need to provide personal and private information. Giving applications too much information (such as security and private information beyond their requests) is also subject to regulations. On the other hand, on some occasions it is better to keep all or some of the records together in the database, both for performance and for usability at a low level. For example, personal and preference information can be kept together.


While PAPI does not impose any model for the storage, our object representation of the information frees the database implementer, and one may easily produce different information categories such as Learner Objects through middle-tier services. Through our object service implementation, Learner objects can be accessed remotely or locally, which allows seamless access to the information that comes remotely from other applications in some cases. The Learner objects are represented as XML objects. Identification of a Learner object in the distributed arena is provided by allowing global namespaces in the XML objects. This provides storing, searching, and retrieving information using the RDF framework in a namespace. As long as global namespaces are supported within objects, they can be stored singly or in multiple repositories.

For both storage types, in order to increase the performance of the general learning system we improved an object cache in the middle tier. In that way, we dramatically reduced certain accesses to the back-end repositories, or remote sources. However, the cache mechanism usually works better in read accesses, which is still an important gain. For the write accesses, persistency should be kept by communicating with the back end.

Whatever back-end storage is planned, PAPI suggests restrictions (Table 4‑4) on the information access, which are largely supported by VCM.

Table 42 PAPI information access control versus VCM support.

PAPI information access controls[PAPI]

VCM support

·                    Some information is kept private. Example: learner personal information.


·                    Some information has restricted access. Example: learner performance information, learner relations information, learner portfolio information.


·                    Some information has varying degrees of public accessibility. Example: learner preference information, learner relations information, learner security information.


·                    Some information must be available to certain application components. Example: learner performance information, learner relations information, learner security information, learner preference information.


·                    Some information is available to administration and management components. Example: learner personal information, learner relations information, learner security, learner performance information.


·                    Some information is primarily useful for humans. Example: learner portfolio information.

Some Data operations

Since we have not focused our research only on our implementation, i.e., VCM, we should have the same semantics for the Learner objects in the same name space as in the next-generation learning systems. To allow the same meaning for information compatibility across the systems, certain operations on data sets are defined in the PAPI (Table 4‑3), where the data set term is equivalent to our object definition. VCM supports most of these operations in its implementation. VCM does not have some operations such as the move operation and copy operation at this time. In addition, VCM needs to implement all data types compatible with PAPI. On the other hand, we just developed a prototype instead of dealing with some very specific details of coding.

Table 43 PAPI operations on data sets (collection of information types) [PAPI.]

PAPI definition of general data operations

Current VCM support

·        Create Operation.  Creating a new instance of some information type, such as personal information.


·        Destroy Operation.  Discarding an instance of an information type in the context of its storage.  Note: Compare destroying a record in application memory, in temporary storage, and in a database.


·        Copy Operation.  Creating a new instance of an information type with identical contents.


·        Move Operation.  Changing a label associated with an instance of an information type by changing the storage of the information (implicit label change) or changing the label itself (explicit label change).  Example: An implicit label change might be affected by creating new "hard links" to a new label, then deleting "hard links" to the old label.  An explicit label change might be affected by changing the label in some "directory" of information.


·        Label Operation.  Creating (or removing) a name, specified by the "caller", to be associated with an instance of information.


·        Navigate Operation.  Using a naming method (absolute, relative, complete, progressive) to locate an instance of an information type.


·        Search Operation.  Finding instances of an information type that match search criteria and returning the found information via references, labels, or copies.


·        Reference Operation.  Creating a handle to an instance of an information type.  Note: The difference between a label and reference is: the "caller" chooses the name for a label, while the "callee" chooses the name for a reference.


·        Dereference Operation.  Using a handle, created through reference, to access an instance of an information type.


·        Aggregation Operation.  Combining several instances of one or more information types into a single container.


·        Decomposition Operation.  Extracting instances of information types from a container.



PAPI standards force applications to support other application-specific data operations on data sets (Table 4‑4).

Table 44 PAPI application-specific data operations[PAPI]

PAPI application specific data operations

Current VCM support

Accumulation Operation.  PAPI learner records may be accumulated, aggregated, or analyzed.  Examples: "What is the average score among third graders?” "What is the grade point average for a particular learner?"


Time Compression and Expansion Operations.  PAPI learner records may be recorded at various levels of granularity.  Time compression reduces the set of records to larger granularity.  Time expansion creates records of finer granularity by interpolation.  Example: quarterly grades and a final exam are rolled up into a final grade; after compression, the data set is reduced, because only the final grade remains.

No, we prefer to keep records for re-evaluation.

Sort Operation.  PAPI records may be ordered, based upon sort criteria.  Example: users may be ordered alphabetically by name.



Furthermore, PAPI forces applications into having data compatibility when they are exchanging data between less capable systems and more capable systems. Converting data formats back and forth is also necessary in some situations. However, these issues are solved in XML-based systems today by directly allowing promotion and demotion of data types and some conversions of data. Now, the systems responsible for handling the XML representations of information perform these operations as a necessity for using XML formatting. XML code binding is also supported by PAPI, which supports various useful features such as allowing different namespace conventions. The following is a high-level screen shoot of XML Learner object DTD

4.2.2 Course Objects

Course Objects is one of the principal development areas in the learning systems, which are synchronous delivery, information management, and learning content preparation, as classified in Section 2.2. Although our research area is not directly in course content preparation, we represent and manage a certain amount of course information other than the course content itself.

The Department of Defense (DOD) researched a reference model that has the traits for sharable courseware that objects need to meet Advanced Distributed Learning (ADL) high-level requirements. At present, SCORM consists of three main sections: an XML-based specification for representing course structures (so courses can be moved from one server/LMS to another); a set of specifications relating to the run-time environment, including an API, content-to-LMS data model, and a content launch specification; and a specification for creating meta-data records for courses, content, and raw media elements [SCORM].

Figure 42 Course structure format of SCORM.

To have shared objects as specified by SCORM, or more generally, to interoperate between LMSs, we need to represent information objects in a common syntax and semantics. Both VCM and SCORM use XML syntax in their representations of objects. The SCORM course structure framework describes a course using three groups of information.  The first group, called globalProperties, is the data about the overall course.  The second, called block, defines the structure of the course, and the third group, objectives, defines a separate structure for learning objectives with references to course elements within the assignment structure [SCORM]. Our approach to the course objects is illustrated in Figure 4‑3. The illustrated DTD diagram shows the general structure of an XML course object. We expect that some other SCORM compatible LMS will fill the missing course content in the structure. We developed and used a number of global properties in the VCM for management purposes. Our globalProperty part corresponds to SCORM’s globalProperty part (Figure 4‑2); however, most of the properties in VCM should be considered as extensions of SCORM.

Figure 43 VCM course object structure.

Differently from SCORM, we interpret course objects as a framework in the learning process. Considering that we still depend on traditional instructional theories, the Learners (students) are registered for courses, assignments are given at each course, quizzes or exams are prepared to pass or fail a Learner in a class, various assessment methods are applied to keep track learning process, etc. Note that a course object is a container, but it does not necessarily have all the learning components in it. For example, a course object is valid and well formed (in terms of DTD structure), even if it does not contain Learner objects or supervisor objects. Furthermore, since those kinds of objects are not directly a subclass of the course object, only references to them are placed in the course object.

Some may say that a course object is superior with respect to others, thinking with an administrative perspective, while others may find it useful to have a Learner-centric representation of all the objects; one student thinks himself to be at the center and taking specific courses, but another may think that the Supervisor objects are the main containers; an Instructor may give several courses and may have many students. Whatever the perspective, each LMS may emphasize different structure hierarchies.

 In fact, we are not imposing any specific object as the main container of the others. Each of our services has already emphasized its object-type-related container understanding. We expressed every service object as a container and as a component at the same time. In this way adapting the objects into different learning management systems became much easier.

The Figure 4‑3 structure is only a high-level representation. The real design is more detailed, with more categorical divisions of the structure. For example, the elements under the “evaluation” branch may not be so clear to the reader. Some, such as the quizRef objects, represent only the questions for evaluating Learners, but not always the results,  and some others, like assessmentNuggetRef, may represent coarse-grained to fine-grained assessment details.

One may argue that the definitions under the root course element correspond to metadata definitions of the SCORM objects. Though this may be correct, our interpretation of a course object and the SCORM interpretation are different. We prefer to call a SCORM object a content object or lesson object since the target is focused on reusable course content object. Our definition, of course, reflects a classroom management hierarchy in traditional understanding. The course lessons, like SCORM objects, can be included, i.e., referred to as objects in the container. Additionally, the different structuring between various learning systems is a natural expectation of standard efforts; in fact, their main purpose is to provide interoperability between them. Therefore, we did not try to fully adapt our design to any other system, but answered our own needs and interoperability issues at a satisfactory level. Even if we do not have exact matches of the structures, we do not expect any reason our application and full SCORM object compatible application not to work collaboratively in the future. Furthermore, our work and other standard efforts started during the same period of time.

4.2.3 Assignment Objects

VCM defines a new type of objects called assignment objects that we did not encounter in other learning management systems, though the objective data structure of  the SCORM model correlated with our assignment objects. Our assignment objects start point is based on the traditional classroom, where instructors gave assignments to the students to be done in a time frame as a part of the course requirements. Assignment objects are mainly to follow course content consumption of the Learner as well as to accumulate performance information.

The assignment object structure behaves as a container to include related objects such as course. The statements of assignments are primarily listed in the hierarchy instead of having nested structures. The reason behind this is that assignments do not carry complex information as do other objects. On the other hand, they are very important for tracking the learning process, both for deciding the appropriate sequence of contents and for assessing the Learner’s progress.


Figure 44 VCM assignment object structure view.

The conceptual DTD diagram (Figure 4‑4) for the assignment objects is shown above. Most of the single elements under the root are useful for defining the identification of the object. The identification of the assignment object comes by providing associations with other existence objects such as the course object and lesson content objects. The grading element, which is generally the main reason for the existence of the object, regulates the expectations to be performed by the Learner, step-by-step. Although there is no defined current practice for assignment objects, we defined an externalMetada reference for future works. Additionally, a different LMS can define extensions with respect to its needs. The existing structure provides enough functionality in VCM. Furthermore, although some of the capabilities in the structure are not used at all, their usefulness is proven in traditional systems. We therefore added this hierarchy to cover possible requirements. Assignment objects may be revisable if any standard or specification effort emerges later, and they may provide motivation for such efforts. A current view of assignment objects is shown in Figure 4‑5.

Figure 45 A client view of assignment object in VCM. Submission of assignments

In using VCM, we practiced the ways of assignment submissions. First, we provided Unix systems accounts to the Learners, who are allowed independent access to their accounts. They were supposed to put all of their assignments files into a place that is open to the Web through another independent Web server. Their submissions showed up in the VCM by using the Learner’s registration of related URL address links to their portfolio information records in the system.  Second, we experimented using  the  Virtual Programming Laboratory (VPL) [VPL97ConcJ], which is a Web-based virtual programming environment based on initial commodity technologies that appeared during the first days of the Web. VPL is an excellent work that explored software delivery over the Web and design issues with a novel set of constraints. Despite the fact that its technology was based on emerging Web technologies a couple of years ago, it is still a very good Web-based system that anybody can assume as a starting point. We used VPL especially for geographically distributed Learners and K12 students who did not know how to use the Unix system. Systems such as VPL are becoming especially important when certain constraints such as enforcing strict deadlines, controlling cheating, etc., are a concern for Supervisors. A centralized management of assignment-related objects and providing a virtual file system to the geographically distributed Learners may be the right alternative for many systems.


Figure 46 Centralized virtual file management through open interfaces in VPL


4.2.4 Quiz Objects

Our study of quiz objects is one of the direct research areas of  IMS specification efforts that seems to find a large audience while only specification work  has been published so far. The IMS Question & Test Interoperability Information Model, called IMS QTI, is focused on the interoperable data structures between the learning systems, particularly between the question and test systems on the Internet.  IMS described their own data structures using UML modeling. The key data structures of IMS QTI is as follows:

·        Assessment – the basic test unit in QTI.

·        Section – a container holding groups of sections and items with common objectives.

·        Item – the basic self-contained entity containing individual questions/ responses.

The Logical data structure of  IMS QTI regarding relationships between Assessment, Session and Item, i.e., ASI, is illustrated below:

Figure 47 Logical relations of QTI elements at IMS QTI.

An Assessment consists of at least one Section (c); a Section may contain other Sections (b) and (f); a Section may contain one or more Items (d) and (h). With this structure, IMS allowed flexible import/export capabilities, as listed below[IMSQTI]:

·        One or more assessments only (c) and (g);

·        One or more sections only (b) and (f);

·        One or more items only (a) and (e);

·        Any number and combination of assessments, sections and items (d) and (h);

·        An assessment may or may not contain more than one section (c) and (g);

·        A section may or may not contain items (b), (c), (d), (f), (g) and (h).

As a result, IMS QTI provides the exchanging of any number and any combination of the above structures, i.e., assessments, sections, and items packed in a single data structure. The main goal of IMS QTI specification is to provide the importing and exporting of this single data structure between the systems.

In our approach, we designed different logical structures as shown in Figure 4‑8, but we provided similar flexible interoperability operations. A Quiz element may have any number of Categories, a Category may have any number of Sets, and a Set may have any number of questions. The existence of at least one element is not strictly required at any level of containment, though we did not make enough experiments on  the possible side effects of  this.

Figure 48 VCM logical data structure and elements in the Quiz object.

In fact, VCM’s Quiz, Category, and Question directly maps IMS QTI’s Questioninterop, Section and Item, respectively. However, we have an extra structure called Set to contain equivalent questions. The rationale behind the Set container is to have question items to test the same knowledge by asking differently. The Set category can be ignored and a Set element with one Question can be mapped to a single Item in IMS QTI.

Different from the IMS QTI, the VCM supports quiz objects with or without assessment at any level, (a) and (b) in Figure 4‑10. Conceptually, the aggregation of its components will be effective in the assessment of a container.


Though question and test types may vary greatly in educational communities, IMS  prepared a taxonomy related to  Questions/Items and corresponding Response times present in the specifications according to their research in the field. More specifically, IMS set its reference identity types based on Response-type instead of having Question-type or item-type concepts. IMS provided a very detailed specification to cover a wide variety of question and test types through the Response-type concept. The Response-type taxonomy of IMS is illustrated below in Figure 4‑9, which VCM architecture supports also. We regard IMS terminology in this thesis.

Figure 49 IMS QTI Response-type taxonomy.

User responses can be basic or composite. Basic response type means that only a single type of response is allowed, e.g., Single Response Type. Composite response types refers to the ones that are containers of basic types in different interpretations such as Multiple, Ordered, etc. Time is another factor in some situations and some response times are required recording of user action timestamps.  The time factor sometimes affects the sequence of items in an exam, as well as the completion requirements [IMSQTI.] Currently, VCM only supports time frame at the exam level, i.e., time-dependent on a sequence of  response-types. Note that IMS QTI allows for proprietary extensions. The specification placed proprietary extensions so that they do not cover the types defined in it.


In our first experiment with architecture design, we implemented an assessment engine in the middle tier. The assessment engine was written in Java and was able to process certain types of questions, i.e., response-types that included Single and Multiple. Even if the architecture allows a new type of Response type, the assessment engine itself cannot grade new types and needs to be revised.. We therefore suggested a new way of assessing quiz objects that is more independent than implementation languages such as Java, C, and C++. As shown in Figure 4‑10, the item/question providers should provide an assessment procedure in XSLT that is convenient to build and distribute across the Internet. The real assessment engine is responsible for coordinating the set of items for assessing and performing the calculations that are not always easily  performed by XSL, though the directions can always be provided.


Figure 410 VCM Quiz service approach

Rendering of response types completely left independent. The same question can be presented by using graphics, multimedia sounds, or just plain text. Rendering process can be done at middle-tier using XSL transform or left to the client depending on the power of the clients.


We should say that VCM assessment engine does not fully implements all the response types, however, addition of these new types is just implementation issue without changing the architecture, or playing with the code significantly.

4.2.5 AssessmentNugget Objects


The advancing technologies opened new ways for evaluating learner progress. While the learning process involves the opportunities offered by the technology, the use of these technology tools is a source of Learner progress. Though we did not completely implemented it in practice, we developed an architecture to be used together with the rest of the system evaluation components such as the above-mentioned quiz and assignment objects. The overall architecture of the system is as showed in Figure 4‑11 below;

Figure 411 An automated assessment architecture to extract hidden nuggets in the learning process by mainly tracking course content page accesses. Data Analysis


Web Usage Analysis can be divided into several steps. The data for mining is collected through the first three components. Our web resources include home pages for specified courses that contain links to a rich set of foil world presented during class time, and other useful information resources and references. The current web server stores the following data from the users' browsing home pages: IP address, time stamp, method, first line of client request, page content-type, URL address of the request, HTTP version, user identifier, return code, bytes transferred, referrer page URL, and agent.

The collected data is subject to filtering for proper analysis.  Students’ personal and performance data from other components should be merged into one center, or made available for accessing globally. Before combining the data in a warehouse, several pre-processings of data may be necessary (for example, finding user sessions on the web access logs correlated with other data sources.) Information Analysis

The combined data is ready for processing to get useful informational statistics. The most popular results produced by various data mining algorithms are association rules, sequential patterns and clustering. Available Data Mining techniques provides us with the following kinds of information: most frequently visited foils, set of foils, or sequence of foils acquired in the combination of success of the students, exam results, topics of interests by students, etc. The correlation between students’ grades and the web resources study can be checked from the perspectives of both assessment and resource improvement. For an example that can be easily reached through current data: 70% of the students who have visited Java-JDBC foils numbered 5 to 27 in the month of May got grades 90 or above from the JDBC assignment. If we have enough support, the number of the cases, for such a result in our database, then we have useful information. We are not presenting further fine-tuned assessments in our architecture but it is situations as in the following sample scenario are always possible. A student who copied the JDBC assignment was caught by being informed about his missing JDBC resource usage in the system and by further personal investigation of him.

 Clustering allows us to group together similar characteristics of both students and Web resources, e.g., students with similar work efforts, or foils useful through the semester. Such results open the door to a rich variety of technology-enhanced strategies, both in student-course assessment and in course preparation from a general education view. Combined with online quizzes and surveys and with performance records, the use of online course material shows us the effectiveness of the course as well as improving it. Continuous data provides us a way to measure the gradual learning process of the students.


The final part of our project consists of the presentation of our acquired knowledge. Presenting results is related to a number of fields, including statistics, graphics and visualization, usability analysis, database querying, and OLAP. Since the whole project was aimed at a distributed education environment, various Supervisors are thought to be present geographically in different locations. A Web interface is essential to the presentation.

4.2.6 Survey Objects

Survey object’s logical data structure is similar to the one showed in Figure 4‑8 for the quiz objects. Despite both object type’s close similarity, we preferred to have them in different object types. The rationale behind is that both object type has different functionality and usage. Mainly, survey objects are not for performance evaluations but general assessment.

The existence LMSs do not have such service functionality directly. We believe that Survey Object Service increases the richness of the VCM. It provides direct feedback of Learners, which helps to determine the time and set of contents given in a course. The service is used numerous times by the NPAC courses given at Syracuse University. Below in Figure 4‑12 is a view of Survey Object Service on client site.

Figure 412 Survey Object Service display at client site.

4.2.7 Performance Objects

Our system has several components for achieving overall assessment of learners: Data Miner, Assignments Grading and Quizzes, where inputs are coming from Web Usage, Learner Records, Manual Grading, and online Quizzes. Different assessment processes feeds various performance objects, which are hierarchically contained in Performance Object container.

Figure 413 General hierarchy of the performance object in VCM.

Performance object keeps aggregated assessment information to provide persistent performance information for a specific learner in a course enrollment. The contributions of all the sub-assessment objects are to be determined by the Assessors of the system. Below in Figure 4‑13   is a high level hierarchy of the performance objects. The gathered performance information can be presented at different granularities. The Figure 4‑14 illustrates some of the usage of performance objects in VCM.


Figure 414 Screen captures of performance object views in VCM.

4.2.8 Supervisor Objects

Supervisor object are another kind of human information stored in the system. This objects contains information similar to the Learner Personal information.. Additionally, information related to coaching of learner process is associated within the object. There are a number of categories for the supervisors, and each has different role in VCM management. The categories and their roles are further explained in section 7.4. Here we present the information object’s structure at a high level (Figure 4‑15.) Supervisor object structure has similar sub elements to the ones of Learner object. Main difference is that having capability element that regulates the access rights of a Supervisor at different levels mentioned in differences in section 7.4. Furthermore, though the performance information may be used to serve for similar demands as in  Learner performance information, it can mean different from the one for Learner such as students evaluation of an instructor at the end of a semester, i.e., comes from different source, or a professor’s research grant, i.e., comes from different source and related with different issues.

Figure 415 Supervisor information structure, a high level view.

4.2.9 System Administration Objects

The System Administration Object contains the information related with the management of VCM related with technical administration. This object type mainly designed for internal use rather than being a service object for the outside clients. The information contained in this object is completely VCM dependent. However, the design issues considered in VCM might also be useful to be a prototype for another future LMS. The information in administration object contains performance settings, system users access log setting, back-end connection settings, etc. A view of client window is presented in Figure 4‑16 below.


Figure 416 Client view of system administration object in VCM.

Chapter 5        

Lessons on Interoperability Issues of Distributed Components

This chapter proposes a new framework based on the lessons derived from our experience for the collaboration of geographically distributed component-based services. Our experience in the subject  includes the design and implementation  of a software architecture and its prototype consisting of various services, and our further analyses of our experiments both in individual uses and in supportive uses with the other systems. Specifically, we provided a distributed, open, and asynchronous information access environment; and we synthesized our environment and the collective work of other synchronous and asynchronous resources. One practical example of such an environment is that we offered distance Web-based courses to students at Jackson State University, Mississippi, from NPAC, our research center at Syracuse University, New York.


Our experience in distance education come to fruition immediately as we found tthat neither a single service nor a collection of services within a single context  are sufficient to perform next-generation, large-scale distance courses. Rather we need to find a way to combine the contributions of several service components within one container, i.e., framework, to perform a large task. Although components are well established in all engineering disciplines and have widespread usage, software components are still in their infancy today. Recent technological developments, e.g., component frameworks such as CORBA, COM, and EJB, are promising and are shaping the future’s software development methodologies. Numerous practices are performed in such frameworks, and varieties of those frameworks are still being researched.

One critical problem emerged with the component frameworks. Components, or objects as a result their usage, are not compatible with others when entered into a different framework. Objects do not have plug-and-play properties that will allow compositions in their own environments. Having a critical object in development when needed means one of two things is necessary: either design and implement it, or obtain this object including its framework from a second party if they are available. While the common first choice has already reached its limitations, the second choice is not practical since none of the frameworks are designed to allow compositions with others in its real meaning.

Moreover, most objects are not readable to non-professional clients, i.e., in semantics and/or in syntax.  Whenever the composition of objects is required, professionals need to be ready each time to use highly sophisticated developer tools, and even those professionals are dependent on various libraries, and of course the objects’  own framework, since objects are meaningful only in their own environment. On the other hand, as we stated in Chapter 1, both professionals and non-professionals are currently involved in the remarkable software environment as a result of recent technological advancements. Even if today’s component development can be left to professionals, its integration should probably, in fact desirable, be left individuals, who have the technical knowledge at the varying levels from profession to non-profession. As much as the knowledge level of composers of components approaches to the ones of non-professions, the component software development project will reach its success limits.

A good reason why we do not already live in a software component world as has happened in other disciplines may be the nature of the software technology itself. The components are actually factored in computing machines through the deployment of metaproducts, e.g. classes versus object instances. This is a rather different situation than exists in many fields and does not allow the making of clear mappings of other component model solutions. Analogies with other fields, e.g., hardware components, lego blocks, etc., are fail to answer the conceptual needs of the software world. Further, conceptual and mathematically sound solutions are failing to answer all the engineering practices and market demands [Clemens98]. Both analogies with other fields to reflect the practical needs and mathematical formalism as a natural property of  the software are required to realize software world of components.

Today we have a good background in mathematics, i.e. conceptual understandings of software component frameworks, and practical experience to forward next step towards the next-generation component-software world. We believe that future’s software integration can be realized through easy-to-use, useful, reusable, common-object structures. More specifically, the common Web-objects approach will be effective in achieving this goal. Since the Web connects numerous worldwide-distributed resources inside, it has the potential of being explored since it is well-accepted. Instead of having components, or their blue-prints, physically carried and integrated in an environment, well-defined object services scattered on the Web can realize desired functionality as the components of the world of software.



Figure 51 The need for a container framework to hold frameworks.

From a different perspective, having component frameworks such as CORBA, COM ready at one hand, we should figure out a solution to use these frameworks’ abilities without sacrificing one framework to another. More precisely, we need another software framework that contains many frameworks in it as shown in Figure 5‑1. However, it is clear that none of the existing frameworks are designed to realize a friendship with any other framework type but its own types. One may argue that CORBA allows to use COM objects, but, it did not evolved practically because of possible wrong design considerations, or market choices, or impossibility that can not be solved just by plugging the components. Though we are not arguing CORBA or any other astonishing engineering assets, we already observed that they did not fulfill all the expectations.

At this point, rather than interpreting the proposed model as a component software architecture model, or framework-framework model, we call the new software model as “Distributed Web-Object Service Model” (DWOSM), which can be mapped to the framework model. The services in the model may be in different granularities. Services’ responsibilities can change from very complex computations to the simple ones. In fact, by interpreting the model as a object service model we aimed to solve an important problem of component frameworks, which is interoperability. More clearly, many frameworks can handle connection and interaction of components with others in other frameworks as well as the ones present in the same framework. However, interaction between components is just an initial step and does not brought conceptual collaboration between them but wiring. Though conceptual collaboration is always provided in the same framework, it is questionable when it happens between  different frameworks.

Note that our use of object world is slightly different from well-known objects used in the object-oriented paradigm. A Web object can define the ones present in object-oriented terminology. However, an object may be a basic document such as an HTML page. If we were bounded with object-oriented concepts, we could not be able to realize larger-scale component-paradigms, though we should benefit from those concepts as much as possible.

Below we explained in more detail how one can achieve the above software described software environment providing properties such as discoverability, accessibility, interoperability, reusability, durability, and manageability.


5.1      Is It a Dream or Doable?


Targeting such an astonishing software world, which is worldwide, distributed, connected with legacy codes, providing the integration of all the well-known different architectures should not be easy! Reaching this goal might seem impossible at first. However, we observed that the capabilities of current commodity architectures reached this potential already. Two cases prove us that this dream-like project is simple in reality.

One is the case that consists of experiments performed at NPAC, which are called , WWVM [Kivanc97] , WebFlow [WebFlow], and Gateway [Gateway99]. For example, the Gateway experiment showed that it is possible to connect back-end high performance computation and communication (HPCC) environments such as Globus with other CORBA components, or COM components in one modular programming environment. Gateway project mainly targeted to bring HPCC resources into commodity arena by middle-tier service collections in a multi-tier architecture. Final environment is called High Performance Commodity Computing (HPcc).


Second case is the Web itself. The simple HTML specification caused tremendous usage and technological advancements in the world. The reason behind is that HTML pages, i.e., HTML objects, have been well understood by every client wired to Internet network. Both syntax and semantic of them were very clear to any client including the browsers launched on heterogeneous machines and even the humans using them. HTML objects are good and simple samples of portability, interoperability and reusability. Furthermore, fast growing number of Web search engines showed us how to locate the resources usable for the requester, i.e. a merit of discoverability. Additionally, we did not live the problem to access those pages after they are published in certain directories of computing platforms through Web servers, i.e. a merit of accessibility. Even, protection mechanisms on some directories showed us how to manage what with side effects in an open world. Furthermore, when one HTML object is not available by some reason it may be compensated by another mirror site, or hundreds of similar and better ones, i.e., a merit of durability.  Finally, the hard part of it, the control of this open environment is distributed to both professionals and non-professionals, i.e., a merit of manageability and scalability.

The below sections describes our Web object model. Section 5.2 explore how to provide interoperability and reusability, Section 5.3 and 5.4 mention about a framework to provide discoverability and durability, and Section 5.5 briefly refer to manageability issues through an event mechanism.


5.2      Interactions within Other System Architectures

As discussed above, the success of the Web objects services model depends on two critical factors, which are their interaction and collaboration. First, interaction means agreeing on common information format and sharing both the format and the data. Note that the concern here is at a higher level than the issues such as networking resources together or agreeing on a data transfer protocol between them, etc. Currently web technologies are suited for machine-to-machine interaction. Specifically, the invention of XML provided us the possibility of wiring two services in a conceptual level. Second, collaboration means reading and understanding the content of shared data as well as its original author does. In other words, having common syntaxes and semantics of Web objects is the key issue of the model.

An important lesson we learned from our research is that it is no matter what distributed object framework or communication protocol we use as long as we can specify object information in well-defined formats. The second-generation technologies in the Internet Age such as XML advanced the methods of information codings, and opened an opportunity to express every piece of information in XML syntax. From simple Web pages to advanced Java objects, every granularity of information can be represented inside XML tags.

Currently, all information on the Web is machine-readable. What is missing is that machines do not understand whatever they read. When two applications is interacting, XML becomes only useful if they agree on common XML tags, their internal structure, and their meanings which is the most important agreement. Many services connected to the web are able to talk, but a few of the rest can understand what they said since XML looks like  only a speech ability but not the common language of speeches.

The problem and a solution have been already discovered in the computing world 2-3 years ago. One right way of solving machines’ speech problem is forcing them to use the same language even if they may think of using different languages. Making analogy with the humans, the first generation of immigrants in America mostly think in their own language but talk in English, however, their next generation both think and talk in English. We are expecting and/or observing the similar outcome for the recently being developed software tools. In any event, the standard efforts such as the ones in learning technologies are best way of forcing a common object syntax and semantics while exchanging the information.

After solving interoperability and reusability issues as stated in previous section, the next related issue is how to describe the resources on the Web.

5.3      Description Framework of Learning-Web Objects

The open face of Web objects does not limit their usage with simple similar-services interactions. The gained information is subject to be used by different applications of various fields as well as by the common focused applications. In case of learning management systems, a Learner portfolio information is important while a student is taking the course, then it is important for recruiters to decide on her, and it might be primary entertainment of a child if it includes a graphical programming game exercise. The interpretation of the same information may greatly vary among its consumers. The namespaces of XML Web objects provides how to interpret the same information and at what level of granularity. Different indexing mechanisms can catalog the same information contents differently, and using application specific namespaces allow to see the same information differently. Again, the standard efforts play a great role in constructing common languages.


Figure 52 Varying consumers of the same information object.

Above in Figure 2‑1 is a picture on how the sub-contents of a general information object can fit the needs of several consumers. The assessor client consumes the whole information for the grading; the recruiter checks only the report section and ignores the rest. Similarly, the child ignores all the information content but the applet. This can be achieved only after the information object properties are defined by different authorities. When the information object is available on the Web, either its author, or independent automated cataloging agents may process its contents and produce metadata about its contents. Metadata may or may not be stored together with the data. Logically its direct producer might store it together with the object, and the other cataloging systems store it in independent places. On the other hand, this is rather an application dependent issue.

The expected question here is that how coding of information for one purpose can answer the needs of the many. Though it is not easy, one right way of doing it is that using different namespaces to describe the same object. Other applications can ignore the property sets that they do not need. In our example, the child client can ignore the assignment report section of the information by applying a different namespace to the information. On the other hand, there may be multiple parties, which are interested within the same information object and its same sub-sections. In this case, while applying the namespaces for each party, we also convert each data property name into its proper namespace corresponded as seen in Figure 5‑3. We expect different inter-disciplinary adapters to appear after well-defined namespaces are published for each field. Note that it is not necessary to write adapters between each field to use only a single source. The adapters can be chained one after another to transform the data into its final user format.

Figure 53 Information transformation chain before the client get the resource.

Finally, similar to its description transformation, the actual information-coding schema may need to be transformed to another coding schema. Specifically, assuming we use an XML binding of standards on the Web, the XML tags and their structure for an information object may need to be converted to another set of XML tags and structure.


5.4      Discoverability of Learning-Web Objects

Assuming that soon the Internet will be full of object services, we need to make it easy for clients to find them, i.e., providing discoverability property.  We have proposed a novel framework for object-web services.  After implementing each object service, we can register it in a global Web-object service table space for Object consumers, which is similar to the Persistent Uniform Resource Locator (PURL). Certain interfaces for each object service can be defined including the metadata about the object types in this space. Independent clients can find the services from the global object service locator (GOSEL) when they needed as shown in Figure 5‑4.



Figure 54 Global Object Service Locators.

One useful property of GOSEL is that after registered the service, the service developer may change the actual location or address of the service. More than that, developers may continue to upgrade the service without worrying about the object consumers. Since the GOSEL is open to the public in the format of an HTML page, the specialized search engine agents can catalog the object services and make them available just like ordinary HTML pages. In view of other application service developers, they can search the Web to find best service matching their needs such as functionality, price and availability of the object service.



5.5      Event System of Distributed Web Object Services

Having a word of distributed web objects available everywhere; the application developers may want to see the list of available objects capable of doing more than only carrying information. We believe that Web objects should be able to carry some control information to be triggered on the consumer site additional to their information packing. The stored information itself may also cause triggering of certain function calls in the consumer site.  In implementation point of view, the SAX model is appropriate to supply necessary intelligence to the objects [SAX].  After learned the published interface of a service collection, application developers may integrate different service interactions and transfer data via web object model using visual component editors managing worldwide components, i.e. web object services.




Figure 55 High level event mechanism of distributed object web services.

Here we encounter again the well definition of control mechanism formats. Unfortunately, current standard efforts are concentrated to supply reusable information object models at best. While the standard efforts are not mature yet for the information model, the architecture views of next generation computing models are not considered widely. Most of the reusable object models are planned for providing export/import capabilities to the applications. However, we targeted the applications whose functionalities can be added and removed at ease of non-professional users.

In above sections, we proposed a novel and feasible architecture model mainly for providing properties such as discoverability, accessibility, interoperability, reusability, durability, and manageability. We believe that similar properties will be key properties in the next generation of software world. We already observed different research and practical efforts to afford those properties at various scales of software development. We believe that a world of pragmatic Web-objects and their services will be leading computing power among other computing models. Our proposed architecture model, of course, is subject to evolve with respect to practical and logical needs when adopted by a large audience. We believe that continuous research on this area is necessary. New challenges will definitely appear in this novel-computing paradigm. For example, when information object is transformed a number of times to reach its final destination, can and should we be able to get the exactly same information object back? Both formal and practical studies will be shaped when we step in this software world.




Chapter 6        

Front-Tier Customizable Open Interfaces

The Web started with a static HTML page presentation, and an immediate appearance of Web-centric applications followed. The rapid evolution of these applications brought an on-the-fly page-generation idea for providing better user interfaces that are customized at the requests of users. Today, most of the applications having open interfaces follow this idea. 

On-the-fly page generation provides flexibility to applications. However, some application groups using this approach have shortcomings, too, because The method used for automatic code generation is embedded inside the application itself.  This approach has two disadvantages: The designing user interface is harder combined with an actual programming code, and programming itself is difficult when different language statements are mixed in with the actual code.

The best alternative for inside-program-embedded user interface preparation is keeping the user interface statements outside the program, i.e., as templates. Templates are kept in a separate repository, either a file system or a database. A final user interface can be prepared by parsing the specific template from the template repository by embedding output data inside it.

Another advantage to using templates is that more than one template can be prepared to display the same information type. Clients can choose different templates with respect to their needs and capabilities.

We experimented with two different template styles in our architecture. The first was HTML-like template files. The second was XSL templates processed together with XML objects.

6.1      HTML-Like Template Files

 Our first implementation was a parser for processing HTML-like template files. The HTML template file structures are very similar to their actual appearances at the client side. That way, the final system users can modify the templates by using a favorite HTML editor of his own choice. Since the template repository is kept in a file directory, separate from the binary files or any other data, editing and browsing templates are straightforward.

Inside the HTML template, a number of special tags and symbols are allowed to harvest the template with the output data. Table 6‑1 presents some examples of tag formats that are to be embedded inside the regular HTML file, with the triggered action in the parser. Although, there are only a few presented formats, they handle most of the parsing operations common in the application area.

Table 61 Sample HTML template tags and parser functionality

Sample Template Tags

Proper Actions of the Parser


Replace with identifier variable value in the program.




Check whether this variable has a value or not, and replace with proper constant value.


Add prefix to variable value.


Multiply variable value with N.


Check whether this variable has a value; if yes, replace with ‘newvalue,’ else leave blank.


Repeat the ‘record’ tag enclosed part of template for a set of entities.


Since the user interface is left to the HTML editor, the job of the parser is just to insert the service output into proper locations in the HTML file. The result is a very fast output generation with enhanced capability. We observed that the performance of the parser is pretty much comparable to the other methods. Because the HTML allows other components inside to enrich the displays such as Java Applets, Dynamic HTML components, JavaScript scripting, etc., the end users have a wide variety of choices.

Figure 61 An HTML-like template and parsed result

In our experimentation, we used JavaScript functionality primarily. Most of the pages gained dynamism from integrating JavaScript scripting. One advantage of JavaScript is that it provides interactivity to the static HTML file rendering. One can manage the interactions with the user without submitting HTML forms to the server by using JavaScript functions. Furthermore, JavaScript enhances HTML form submission by moving data at the background between different forms and collecting user responses. JavaScript provides background communication with the server without submitting the active widget state in front of the user, thus preventing undesired display changes. One may prefer using Java applets instead of JavaScript interpreting. Although our templates allow the inserting of Java applets or any other dynamic programming components, we preferred JavaScript. The rational behind that is that today we have a large enough library of JavaScript with rich functionality, from pop-up menus to mouse action tracking, from controlling static components to communicating with dynamic applets, etc. In addition, JavaScript codes work faster than applets in most cases, even if they are executed by an interpreter. The main reason for this unexpectedly good performance of the interpreter comparing to the compiled code executing is that JavaScript functions are calling binaries, which are parts of compiled browser binaries, further ready for the execution in the memory. On the other hand, other components like Java applets are executed by a second toolbox such as JRE, which is subject to being loaded first in some cases.

6.2      XSL Templates

Though we got good results with our HTML-like file templates and their parser, the increasing demand for XML file objects led us to use different mechanisms such as XSL templates. One drawback to HTML-like template files is that they do not provide enough functionality for the complex data representations, which can be achieved in the case of coupling XML objects and XSL templates. Rendering XML object files with XSL style sheets is very powerful and has become popular in the industry. On the other hand, using XSL templates has a drawback, compared to the first method, as they are complex with respect to both XML and HTML files. Therefore, it is not straightforward to edit XSL files. Nevertheless, many tools are appearing in the industry to provide easy XSL file editing.

Normally, XSL transform each XML element into a format that is well-understood by web browsers. Furthermore, XSL is also capable to add new elements, or to remove them, to rearrange and sort the elements, to test and decide which elements to display, etc. An XSL file may look like a file written in a scripting language. This capabilities of XSL mechanism provided a nice property in our research. The same XSL file can produce different outputs, in different formats and with varying contents.  This way our same service provider can manage different requests of distributed various client types.  More specifically, assuming each user has varying power of browsers, we can produce client-specific output for the same information content. Furthermore, we assumed the other systems as regular users, and send them the information in the manner they understand instead of fancy HTML display by the help of the XSL processing. The below Figure 3‑1 illustrates our approach. While the number of different learning systems and display devices are increasing, making the output readable to everybody is quite important. In case of service site is responsible from the user interface a mechanism resembling XSL is vital for the being developed next generation applications.


Figure 62 Different outputs for varying clients through XSL mechanism

In fact, there is two choice for the developers while displaying XML output; keeping XSL on server site or keeping XSL on client site.

Currently, not all of the major browser market is supporting XML support, and the clients are not always web browsers. Therefore, transformation of the XML document on the server should be done and sent it as plain HTML, or in any other desired format, to the clients, to make our XML data readable for all varying clients.

In the second case, an XSL style sheet information should be added to the XML file, and to let the browser do the transformation. Though this works fine, it is not always desirable to include a style sheet reference in the XML file, and the solution will not work in a non-XML aware browser. A much more adaptable solution would be to use a JavaScript to do the XML to HTML transformation. Using JavaScript allows to do browser specific testing and to use different style sheets according to browser, i.e., user needs. XSL transformation on the client side is one of the  key concerns in the development of new browsers or their new versions.


Chapter 7        

Security Issues of an Open Access Environment

Human records are among the most private kinds of information in any application area. Similarly, providing enough security and privacy for Learner records are vital subject education technologies. Because of its exposure to the public over the Internet, the whole environment is subject to attacks to capture, to alter, and to destroy the private information. For the sake of completeness, complete security should be inserted into the architecture. Though complete security became never possible for any application until today, it may be provided up to some acceptable risk level. We solved the security issues under the following categories:

7.1      Communication Channel Security

Currently, we are using the technology to secure the communication channels over a public network by using public key cryptography in the SSL protocol for securing the transfer of information over the Internet. The SSL channel secures the most important part of the communication. However, the back-end communications between middle-tier and back-end repositories are vulnerable for the attacks. We did not implemented a encrypted data transfer protocol as our own, but leave this part as a future work or vendor implementers. In fact, this technology should be advanced by database vendors to improve security for current technology applications, which largely use databases in a distributed environment. Despite the databases became popular again with the applications on top of the Internet infrastructure, database vendors did not improved their products competitively to current standards. This is mainly because of difficulties adapting security issues in a rapidly developing market. On the other hand, the solution is not absent at all. Security at back-end is usually provided by isolating database or file system servers from the outside world. In our system, we suggest the same solution until more elegant secure communication methods become common for the back-end services. Though there are numerous secure applications in the market, which are using back-end tools also, the security implementations are always a cumbersome for these applications while they are targeting different solutions but needing security as a side effect. Furthermore, people are always implementing their own security mechanisms each time with great efforts, and these efforts are not going beyond being specific to their applications. We believed that communication security at any level should be out of question for the application developers for advancing next generation software tools. Of course, the other security issues like mentioned below sections are the responsibility of the application developers.

Figure 71 Autentication screens for access into VCM

7.2      User authentication

In VCM, similar to many computer operating systems, a user authenticates himself by entering a user login id and a secret password known solely to herself and the system. The security credentials of the user are kept in database and they are never displayed to public or system coaches in any way. The users are responsible to remember the passwords. On the other hand, VCM administrators may assign new passwords, or login ids, to users. After that, user himself can change the password but not the login id. Login id is the unique set of the characters identifying user in the system. This approach is very compatible with UNIX user authentication.


7.3      User access lists

 The second part of the authentication mechanism is keeping access lists for the users. All the users are restricted only to access to the allowed part of the information through open interfaces. Access lists can be easily manipulated in the user records interface, simply by carrying present course names to the access list box. Each Learner and Supervisor can only access to the courses listed in their access list.


7.4      User privilege levels

 Since privacy issues are most important in such a system, user password authentication and access lists are not enough to keep privacy by themselves. Therefore, a further mechanism, user access privileges, is also included in the system. The user-access-privileges puts one more layer of restrictions to access and/or manipulate any private data. After getting a successful password authentication, and having desired course specified in the access list, a user must have powerful enough privileges to read and/or to write, update some records, or to see HTML information pages. In our design, we presented seven different privilege levels to the system users to expose public and private information at different levels. We have the following privilege levels as shown in Table 4‑1. Each privilege level provides varying amount of the information access and modification. The privilege levels are good to respect public and private data distinction categories and to provide security credentials only to the owner.

Table 71 Privilege categories in VCM.




Super User

Can access any information but security credentials of others, and modify any information

System administrator


Can access and modify the information allowed in access lists with the restriction of security credentials

Supervise the overall process at background, does not appear in any relation list with Learners or others.


Can access and modify the information allowed in access lists with the restriction of security credentials

Main responsible of the course management


Can access and modify the information allowed in access lists with the restriction of security credentials

Secondary helpers, i.e. staff


Can access and modify the information allowed in access lists up to some level with the restriction of security credentials.

Regular Teaching Assistant jobs


Can access public information and modify some security credentials.

Technical staff, e.g. Unix administrators


Can access public information, like class list, Learner public information, etc.

Thirty party system observers, e.g. demo users


Can access and modify personal and private information, security credentials, etc., except his identifier to the system

Learner (students); ones own record manipulations


7.5      Presentation security

 One of the challenging issues in the security is not the communication security but keeping the data private after the data is transferred. It is highly possible that students will view their grades and other private records in a public cluster. Although it is not possible to provide complete security at this level, we improved different strategies to make the risk as low as possible.

Preventing of caching of private files is left to the browsers’ responsibility for the end clients. The SSL server prevents browser caching in general as long as user directly set the browser to store them. The possibility that a user may turn an option to cache the data is a security leak in the system, but it is highly unlikely in practice.

Specific timeouts for each screen is placed for each critical display window. After the time is out, the window destroy itself for the cases where user may ignore logout his session by some reason.

Furthermore, VCM manager assign unique and random session key for each login to the users. Each user session can be set not to exceed a specific time period and can not be used for a second login. Using session keys, system allows only one login at a time for each authenticated user. This way user identification through URL addresses and copying the URL by second party to reach the information is prevented. Also, system alert the user if a second party uses his session key if the session is not ended. The unique session key prepared by generating a random number, and combining it with the current time.


Chapter 8        


Lessons Learned

In this chapter we highlight the lessons that we learned from our research presented throughout this thesis:


·        While the last decade was for the advance of commodity computing technologies, learning has become important, and our learning styles have changed significantly. Technology-based learning has found a large audience. Learning management systems became important and are a direct research area of computer science.


·        Though almost every field of research or commerce in the world has benefited from the web, not many of them are utilizing web technologies as well as they could. The principal problem of some application fields such as Learning systems is in having old-fashioned architecture styles.


·        We need better architectures that support learning management systems as well as others. Furthermore, we need additional architectures and models because of technology advancements such as a specialized Web Mining Architecture for assessment.


·        Although the Web initially started as a distributed document-retrieval system, it now has gained unreachable computing power. However, Web technologies offer more power than their current usage. What is missing is that their well-integration. Web can be a framework for distributed Web-Object World Services: it has the noteworthy property of being able to include everything inside it, which can make it a framework for all frameworks such as CORBA, COM, etc.


·        We have observed that individual technologies, their computing frameworks and platforms are no longer important. We do not need to choose a winner among computing technologies, but we can utilize all of them by agreeing to provide common interoperability issues.


·        On the contrary, we believe that previous computing technologies will make each other stronger if we can achieve their integration, which is technically possible. For example, we can use CORBA and COM services together by defining uniform XML interfaces.


·        Furthermore, from our experiments, we have concluded that one system and one technology is not enough to perform a world-wide application such as providing distance courses. It would be preferable to find a collaborative way between different components of learning technology systems.


·        The various efforts for standardization current learning object types are one of the best works to reach the collaborative systems goal. We specifically followed learning technology standards and analyzed their work. Some application based standard studies may not be long-lived, since they currently force limitations in so fine-grained scale and impose certain models on other systems.


·        Making analogies with the software model of the Web and the component models of other fields might not be as long lasting as hardware components model, or Lego game pieces, etc., because software is different than other entities. It needs to carry properties that pertain to both practical needs and mathematical consistency. Solutions that are too complex may not achieve market values and solutions that are too analogous may fail quickly.


·        Additionally, we learned that technologies, i.e. services, should be made simple enough to be used by a large audience. We have observed that some technologies are well designed, but not explored because of their complexity. In fact, we believe that  non-technical users should be able to build their systems using the latest technologies without the help of experts. Other than spreading to a large audience, further, easy-to-use services will solve conflicting problems between the fields. For example, it is not always effective for computer scientists to work on an area of instructional theory at each level of detail, because these are established as a result of lengthy research and literary background of this field.


·        We concluded a software model for Web-Object services with several vital properties, but not limited to them, such as interoperability, reusability, discoverability, accessibility, manageability. We have proposed this distributed computing service model based on the success of the HTML model with the same properties.


·        Our experiments in online quizzes showed that LMS systems should be tailored to answer practical social issues based on the feedback from educators. Otherwise, a very powerful system can be wasted because of various missing manipulations that are technically easy. For example, the cheating of Learners during  exam should be considered in a technology-based exam system.



Chapter 9        

Concluding Remarks and Directions for Future Research

Despite the fact that the World Wide Web was initially intended to be used as an online multimedia document retrieval system, its very powerful infrastructure caused the rapid development of commodity technologies.  The main characteristics of this remarkable software development environment are the following:

·        WWW provides a distributed software environment naturally by being a network-based infrastructure.

·        Any heterogeneous combination of machines with different software and hardware architecture can benefit from servers available on the web through standard open interfaces.

·        Numerous tools are available to access back-end tools located on server machines.

·        The Web browsers presently still being developed have a huge capability for handling a rich set of graphical user interface components and for utilizing operating system functionalities running in the background while being independent of the operating system.

·        Having Web-related standards are well accepted, the second part of the standards is being developed to support server-to-server communication that may enhance distributed computing.


Learning technologies were among many others that were dramatically enhanced in the recent past. However, instead of fully utilizing all the above important features, many of the initial approaches focused on specific features of the Web. Most applications used standard open interfaces of the Web and locked with old-days’ client-server architectural paradigms. Learning technologies are a good representatives of this type of approach.

We recognized the missing interpretation of the Web by the learning technologies and designed a multi-tier architecture that is implemented as an asynchronous, open, distributed information-access environment, i.e., provided by the Virtual Classroom Manager (VCM). VCM evolved together with learning standardization efforts and various learning system applications, as well as the second part of the Web standards such as XML and XSL.

Our architecture consists of front-end user interfaces provided through standard open interfaces, middle-tier service components, and back-end repositories. Front-end user interfaces may be improved by system administrators without a strong technical background. Middle-tier services are capable of communicating with other systems as well as with front-end users through object models that are well agreed upon. The back-end repositories may be independently designed either as file system repositories or as traditional relational databases, independent of middle-tier object structure, which are connected later through bridges and adapters. The data of these back end repositories are converted into  objects by middle-tier service components.

In the middle-tier, we designed each information type representation corresponded to the related standard effort being developed in the learning technology arena. In this way, we showed a wider picture of future learning-system architectures. In our design, we did not disregard a specific standard effort but, rather, searched for a way to adopt one into the architecture, since most of the standards are focused in different areas of the general problem. We understand that systems of future generations will be utilizing distributed paradigms and component technologies instead of realizing everything in one big application. Different distributed services can be rented for a specific time, or one specific service can be replaced with another and will need to be integrated into the system forever. Therefore, the emphasis of our research was to experiment with commodity Web objects. We especially interpreted the output of each service component as a Web object in this context.

Though VCM illustrated a number of important issues as presented throughout this thesis, it continues to be the source of interesting research issues and challenging implementation problems: for example, preparing course content correlated with personal performance records. As future work, we see that a more general and comprehensive implementation of VCM may be necessary, based on suggested design issues and experiments already being done. As standards finalize in the near future, the final adapters should be written to provide interoperability between systems.

Web portals have been gaining more attraction and usage as the number of online users increases while the time people are willing to spend for online searches decrease. In addition to current commercial Web portals, computing and educational portals, which are more specific to their respected topics, will be more important for researchers and academia. An educational portal would perfectly fit as an interface for an LMS, i.e., a VCM, developed in a multi-tier Web application environment. For a Learner, the portal might contain a list of his-or-her currently registered courses, up-to-date announcements (e.g., assignments, surveys, exams and other roster information, dates and deadlines for them, and grades and statistics postings), curriculum information (e.g., the next topics to be covered), and links to discussion areas, mailboxes, calendars, and reminders. As such, links to other online applications for course content development, archives of homework and exam questions and the like can be on lecturers’ portal pages. alike








[ABCH95pooma] S. Atlas, S. Banerjee, J. Cummings, P. Hinker, M. Srikant, J. Reynders, and M. Tholburn, "POOMA: A High Performance Distributed Simulation Environment for Scientific Applications," Proceedings of Supercomputing '95, San Diego, CA, December 1995.

[AKENTIWeb] S. S. Mudumbai, W. Johnston, M. R. Thompson, A. Essiari, G. Hoo, K. Jackson, Akenti - A Distributed Access Control System, home page:  http://www-itg.lbl.gov/Akenti

[AM94poet] R. Armstrong and J. Macfarlane, "The Use of Frameworks for Scientific Computation in a Parallel Distributed Environment," Proceedings of the 3rd IEEE Symposium on High Performance Distributed Computing, San Francisco, CA, August 1994, pp.    15-25.

[APACHEWeb] Home page http://www.apache.org

[ASWB95zoom]  C. Anglano, J. Schopf, R. Wolski, and F. Berman, Zoom, "A Hierarchical Representation for Heterogeneous Applications," University of California, San Diego, Department of Computer Science and Engineering, Technical Report CS95-451, January 1995.

[AVSWeb] Advanced Visualization System, http://www.avs.com/

[BBB96atlas] J. Baldeschweiler, R. Blumofe, and E. Brewer, "ATLAS: An Infrastructure for Global Computing," Proceedings of the Seventh ACM SIGOPS European Workshop: Systems Support for Worldwide Applications, Connemara, Ireland, September 1996.

[BDGM96hence]   A. Beguelin, J. Dongarra, A. Geist, R. Manchek, K. Moore, and V. Sunderam, "Tools for Heterogeneous Network Computing," Proceedings of the SIAM Conference on Parallel Computing, 1993.

[BM95hetero] F. Berman and R. Moore, eds, "Heterogeneous Computing Environments," Working Group 9 Report from the Proceedings of the 2nd Pasadena Workshop on System Software and Tools for High Performance Computing Environments,  January 1995.

[BRRM95gems]  B. Bruegge, E. Riedel, A. Russell, and G. McRae, "Developing GEMS: An Environmental Modeling System," IEEE  Computational Science and Engineering, Vol. 2, No. 3, Fall 1995, pp. 55-68.

[Blackboard] E-Learning software platform, http://www.blackboard.com/.

[CD96netsolve] H. Casanova and J. Dongarra, "Netsolve: A Network Server for Solving Computational Science Problems," Proceedings of Supercomputing '96, Pittsburgh, PA, November 1996.

[CDK94ds]  G. Coulouris, J. Dollimore, and T. Kindberg, Distributed Systems: Concepts and Design, 2nd Edition, Addison-Wesley, Inc., 1994.

[COMWeb] COM Home Page http://www.microsoft.com/com

[Calos96] Murray W. Goldberg, "CALOS: An Experiment with Computer-Aided Learning for Operating Systems", in Proceedings of the ACM's 27th SIGCSE Technical Symposium on Computer Science Education, 1996.

[Clemens98] Clemens Szyperski, Component Software, Addison-Wesley, New York, NY, 1998.

[Cooley97] Robert Cooley, Bamshad Mobasher, Jaideep Srivastava, Web Mining: Information and Pattern Discovery on the World Wide Web, in Proceedings of the 9th IEEE International Conference on Tools with Artificial Intelligence (ICTAI'97), November 1997

[CumulvsWeb] James Arthur Kohl and Philip M. Papadopoulos, "The Design of CUMULVS: Philosophy and Implementation," in PVM User's Group Meeting, Feb 1996. http://www.epm.ornl.gov/cs/cumulvs.html

[DARP98ACMJava] Erol Akarsu, Tomasz Haupt and G. Fox, DARP: Java-based Data Analysis and Rapid Prototyping Environment for Distributed High Performance  Computations, ACM 1998 Workshop on Java for High-Performance Network  Computing.

[DARP98Conc] Erol Akarsu, Tomasz Haupt and G. Fox, DARP: Java-based Data Analysis and Rapid Prototyping Environment for Distributed High Performance Computations, Concurrency:Practice and Experience , Vol. 10(1),   1-9(1998).

[DARPSC97] G. Fox, W. Furmanski and T. Haupt, SC97 handout: High Performance Commodity Computing (HPcc), http://www.npac.syr.edu/users/haupt/SC97/HPccdemos.html

[DARPWeb] E. Akarsu, G. Fox, T. Haupt, "The DARP System," http://www.npac.syr.edu/users/haupt/HPFI.

[DiscWHPCN97] K.A.Hawick, H.A.James, C.J.Patten and F.A.Vaughan, "DISCWorld: A Distributed High Performance Computing Environment," Proc. of High Performance Computing and Networks (HPCN) Europe '98, Amsterdam, April 1998.

[EJBWeb] Enterprise JavaBeans http://java.sun.com/products/ejb/

[Erol00] Ph.D. Thesis 2000, Syracuse University

[ETM95taxonomy]    I. Ekmecic, I. Tartalja, and V. Milutinovic, EM3: "A Taxonomy of Heterogeneous Computing Systems," IEEE Computer, Vol.  28, No. 12, December 1995, pp. 68-70.

[FK96globus]    I. Foster and C. Kesselman, "Globus: A Metacomputing Infrastructure Toolkit," Proceedings of the Workshop on Environments and Tools for Parallel Scientific Computing, Lyon, France, August 1996.

[FT96nexusjava]   I. Foster and S. Tuecke, "Enabling Technologies for Web-Based Ubiquitous Supercomputing," Proceedings of the 5th IEEE Symposium on High Performance Distributed Computing, Syracuse, New York, August 1996.

[FalconWeb] Karsten Schwan, John Stasko, Greg Eisenhauer, Weiming Gu, Eileen Kraemer, Vernard Martin, and Jeff Vetter, "The Falcon Monitoring and Steering System," 1996, http://www.cc.gatech.edu/systems/projects/FALCON/.

[Fox96] G.C. Fox, "An application Perspective on High Performance Computing and Communications," Technical Report, Northeast Parallel Architectures Center, Syracuse University, 1995.

[GEFWeb] GEF: The Graph Editing Framework home page: http://www.ics.uci.edu/pub/arch/gef/

[GHR94pse] E. Gallopoulos, E. Houstis, and J. Rice, "Computer as Thinker/Doer: Problem-Solving Environments for Computational Science," IEEE Computational Science and Engineering, Vol. 1, No. 2, Summer 1994, pp. 11-23.

[GNW95legion] A. Grimshaw, A. Nguyen-Tuong, and W. Wulf, "Campus-Wide Computing: Results Using Legion," University of Virginia Computer Science Department, Technical Report CS-95-19, March 1995.

[GSSAPIWeb] RFC 1508, RFC 2078

[GatewayGrande99te] Tomasz Haupt, Erol Akarsu, and G. Fox, The Gateway System: Uniform Web Based Access to Remote Resources, ACM 1999 Java Grande Conference    

[GatewayHPDC99et] Erol Akarsu, Geoffrey Fox, Tomasz Haupt, Using Gateway System to Provide a Desktop Access to High Performance Computational Resources, HPDC-8, 1999.

[GlobusIntlJ97] I. Foster and C. Kasselman," Globus: A metacomputing infrastructure toolkit," Int't J. Supercomputer Applications, 11(2): 115-128, 1997.

[GlobusWeb] I. Foster, C. Kesselman, "Globus", http://www.globus.org/

[HPC++Web] D. Gannon et al., "HPC++", http://www.extreme.indiana.edu/sage

[HPCCEuroPar98Gf] G.C.Fox, W. Furmanski, T. Haupt, E. Akarsu, and H. T. Ozdemir, "HPcc as High Performance Commodity Computing on top of integrated Java, CORBA, COM and Web  standards," Proc. Of Euro-Par '98, Aug 1998

[HPDForumWeb] High Performance Debugging Forum, http://www.ptools.org/hpdf/ [DAQV96Sth] Steven T. Hackstadt and Allen D. Malony "Distributed Array Query and Visualization for High Performance Fortran," in Proc. Of Euro-Par '96, Aug 1996, http://www.cs.uoregon.edu/ hacks/research/daqv/.

[HPFfeWeb] Guansong Zhang et al., "The HPF frontEnd system," http://www.npac.syr.edu/users/zgs/frontEnd/ .

[HPccGridBook]G. Fox and W. Furmanski, "HPcc as High Performance Comodity Computing," Building National Grid by I. Foster and C. Kesselman, Chapter 10, pp. 238-255

[HRJW95mpse]    E. Houstis, J. Rice, A. Joshi, S. Weerawarana, E. Sacks, V. Rego, N. Wang, C. Takoudis, A. Sameh, and E. Gallopoulos, "MPSE: Multidisciplinary Problem Solving Environments," Purdue University, Department of Computer Sciences, Technical Report CSD-TR-95-047, 1995.

[HabaneroWeb] "NCSA Habanero," http://www.ncsa.uiuc.edu/SDG/Software/Habanero/index.html

[Hood96p2d2]  R. Hood, "The p2d2 Project: Building a Portable Distributed Debugger," Proceedings of ACM SIGMETRICS Symposium on Parallel and Distributed Tools (SPDT '96), Philadelphia, PA, May 1996.

[HpccWeb] G. Fox, W. Furmanski, "HPcc as High Performance Commodity Computing," http://www.npac.syr.edu/users/gcf/hpdcbook/HPcc.html

[JGrandeWeb] Java Grande Forum, home page: http://www.javagrande.org

[JINIWeb] The SUN Jini Technology, http://java.sun.com/products/jini

[JWORBWeb] G. C. Fox, W. Furmanski and H. T. Ozdemir, "JWORB - Java Web Object Request Broker for Commodity Software based Visual Dataflow Metacomputing Programming Environment," NPAC Technical Report, Available at http://tapetus/iwt98/pm/documents/hpdc98/paper.html

[Jackson98] David E. Bernholdt, Geoffrey C. Fox, Roman Markowski, Nancy J McCracken, Marek Podgorny, Thomas R. Scavo Syracuse University | Debasis Mitra and Qutaibah Malluhi, Jackson State University, Synchronous Learning at a Distance: Experiences with TANGO, SC 1998.

[JavaBeansWeb] JavaBeans Component Architecture by SUN, http://www.javasoft.com/beans/index.html

[JigsawWeb] Jigsaw home page: http://www.w3.org/Jigsaw/

[KPSW93chal]  A. Khokhar, V. Prasanna, M. Shaaban, and C. Wang, "Heterogeneous Computing: Challenges and Opportunities," IEEE  Computer, Vol. 26, No. 6, June 1993, pp. 18-27.

[LG96legion]  M. Lewis and A. Grimshaw, "Using Dynamic Coonfigurability to Support Object-Oriented Programming Languages and Systems in Legion," University of Virginia Computer Science Department, Technical Report CS-96-19, December 1996.

[LTSA] Learning Technology Systems Architecture (LTSA) Specification, IEEE P1484.1/D6, http://www.edutool.com/ltsa

[LegionACM97] A. S. Grimshaw, W. A. Wulf, and the Legion team, "The legion vision of a worldwide virtual computer," Communications of the ACM, 40(1):39-45, 1997

[NHSS87sigops]  D. Notkin, N. Hutchinson, J. Sanislo, and M. Schwartz, "Heterogeneous Computing Environments: Report on the ACM SIGOPS Workshop on Accommodating Heterogeneity," Communications of the ACM, Vol. 30, No. 2, February 1987, pp.   132-140.

[Netsolve97IJSP] Henri Casanova and Jack Dongarra, "NetSolve: A Network Server for Solving Computational Science Problems," The International Journal of Supercomputer Applications and High Performance Computing, Volume 11, Number 3, p.p. 212-223, Fall 1997.

[NexusJPDC97]  I. Foster, C. Kesselman, an S. Tuecke, "The nexus approach to integrating multithreading and communication," J. Parallel and Distributed Computing, 45:148-158, 1997.

[NexusWeb] I. Foster, C. Kesselman, "The Nexus Multithreaded Runtime System," http://www.mcs.anl.gov/nexus/

[OMGWeb] CORBA - OMG Home Page http://www.omg.org

[ORBacusWeb] Object Oriented Concepts, Inc., http://www.ooc.com/ob.html

[PAPI] PAPI Specification, http://edutool.com/papi/

[PAWSWeb] Parallel Application WorkSpace, http://acts.nersc.gov/paws/main.html

[PCRCWeb] PCRC, http://www.npac.syr.edu/projects/pcrc/

[PURL] Persistent Uniform Resource Locator, http://purl.oclc.org/

[Pitkow97] J. Pitkow, In Search of Reliable Usage Data on the WWW, Proceedings of Sixth International World Wide Web Conference (1997).

[SC92meta]   L. Smarr and C. Catlett, "Metacomputing," Communications of the ACM, Vol. 35, No. 6, June 1992, pp. 44-52.

[SC98Pres] T. Haupt, "WebFlow High-Level Programming Environment and Visual Authoring Toolkit for HPDC (desktop access to remote resources)," SC'98 technical presentation, http://www.npac.syr.edu/users/haupt/WebFlow/papers/SC98/foils/index.htm

[SCORM] Shrable Courseware Object Reference Model, http://www.adlnet.org/index.cfm

[SDA96support] H. Seigel, H. Dietz, and J. Antonio, "Software Support for Heterogeneous Computing," ACM Computing Surveys, Vol. 28, No. 1, March 1996, pp. 237-239.

[SSLWeb] SSL, Netscape Communications, Inc, http://home.netscape.com/eng/ssl3/index.html

[SUN] Sun Microsystems, Inc., http://java.sun.com

[SciRun97Press] S. Parker, D. Weinstein, and C. Johnson, "The SCIRun computational steering software system," In E. Arge, A. Bruaset, and H. Langtangen, eds., Modern Software Tools in Scientific Computing, pages 1-44. Boston: Birkhauser Press, 1997

[Sciviz98ACMJava] Byeongseob Ki and Scott Klasky, "Collaborative Scientific Data Visualization," ACM 1998 Workshop on Java for High-Performance Network  Computing.

[Speedtracer] K.-L. Wu, P. S. Yu, and A. Ballman, A Web usage mining and analysis tool, http://www.almaden.ibm.com/journal/sj/371/wu.txt

[Tango97SIAM] L. Beca, G. Cheng, G. C. Fox, T. Jurga, K. Olszewski, M. Podgorny, P. Sokolowski, and K. Walczak, "Web Technologies for Collaborative Visualization and Simulation" in Proceedings of the 8th SIAM Conference on Parallel Processing for Scientific Computing, March 16-19 1997, Minneapolis, MN, http://trurl.npac.syr.edu/tango/papers.html.

[TangoWeb] M. Podgorny et al; "Tango, Collaboratory for the Web," http://trurl.npac.syr.edu/tango/

[Tuchman91Vis] A. Tuchman, D. Jablonowski, G. Cybenko, "Runtime Visualization of Program Data," in Proc. Visualization '91, (IEEE, 1991) 225-261

[UMLWeb] UML Home Page http://www.rational.com/uml

[VGJWeb] VGJ, Visualizing Graphs with Java home page: http://www.eng.auburn.edu/department/cse/research/graph_drawing/graph_drawing.html

[VPL97ConcJ] K. Dincer and G. C. Fox, "Using Java and JavaScript in the Virtual Programming Lab: A Web-Based Parallel Programming Environment," Concurrency: Practice and Experience Journal, June 1997.

[WebCT96] Murray W. Goldberg, Sasan Salari and Paul Swoboda, "World Wide Web Course Tool: An Environment for Building WWW-Based Courses", Computer Networks and ISDN Systems, 28,1996.

[WebCT97] Murray W. Goldberg and Sasan Salari, "An Update on WebCT (World-Wide-Web Course Tools) - a Tool for the Creation of Sophisticated Web-Based Learning Environments", Proceedings of NAUWeb '97 - Current Practices in Web-Based Course Development, Flagstaff, Arizona,June 12 - 15, 1997.

[WebCTENBL97] Murray W. Goldberg, "Communication and Collaboration Tools in WebCT ", Proceedings of Enabling Network-Based Learning, May 28 - 30, 1997, Espoo, Finland

[WebCTWeb] http://www.webct.com

[WebFlow97Furm] D. Bhatia, V. Burzevski, M. Camuseva, G. C. Fox, W. Furmanski and G. Premchandran, "WebFlow - a visual programming paradigm for Web/Java based coarse grain distributed computing," Concurrency: Practice and Experience, Vol. 9 (6), pp. 555-577, June 1997

[WebFlowAlliance98et] Erol Akarsu, Tom Haupt and G. Fox, Quantum Simulations Using WebFlow - a High Level Visual Interface for Globus, Alliance'98 poster and demo.

[WebFlowDARP] W. Furmanski, T. Haupt, "DARP System as a WebFlow module", http://www.npac.syr.edu/users/haupt/HPFI/webflow/

[WebFlowFGCS99te] Tomasz Haupt, Erol Akarsu, and G. Fox, Web-Based Metacomputing, Special Issue on Metacomputing for the FGCS International Journal on Future Generation Computing Systems,1999

[WebFlowHPCN99te] Tomasz Haupt, Erol Akarsu and G. Fox, WebFlow: a Framework for Web-Based Metacomputing, High Performance Computing and Networking '99 (HPCN), Amsterdam,  April 1999.

[WebFlowSC98et] Erol Akarsu, Tomasz Haupt and G. Fox, WebFlow - High-Level Programming Environment and Visual Authoring Toolkit for High Performance Distributed Computing, Supercomputing '98, November 1998.

[WebSubmit] "WebSubmit:  A Web Interface to Remote High-Performance Computing Resources," http://www.itl.nist.gov/div895/sasg/websubmit/websubmit.html






NAME:                                   Mehmet Sen

DATE OF BIRTH:                6 September 1970

PLACE OF BIRTH:              Balikesir, TURKEY


DECEMBER 2000                  Ph.D. in Computer Science

Department of Electrical Engineering and Computer Science, Syracuse University,

Syracuse, NY, U.S.A.

DECEMBER 1996                  M.S. in Computer Science

Department of Electrical Engineering and Computer Science, Syracuse University,

Syracuse, NY, U.S.A.

JUNE 1992                              B.S. in Computer Science

Bilkent University,

Ankara, TURKEY


September 1996 – May 2000               Graduate Research Assistant

                                                            Northeast Parallel Architectures Center

                                                            Syracuse University

                                                            Syracuse, NY, U.S.A.

JANUARY 1997 – AUGUST 1997     Graduate Teaching Assistant

Department of Electrical Engineering and Computer Science, Syracuse University,

Syracuse, NY, U.S.A.



Applet    A partial Java application program designed to run inside a web browser with help from some predefined support classes.

Binding   An application or mapping from one framework or specification to another.

CGI   Common Gateway Interface. A non-Java technique of sending data from HTML forms in browsers to server programs written in C, Python, Tcl, or Perl. They typically do data base searches or process  data in HTML forms and send back MIME.

Coding   (1) In information interchange, a formalized or structured representation of information. (2) A process of representing information in some structure.

COM   Common Object Model.  Microsoft's windows object model, which is being extended to distributed systems and multi-tiered architectures. ActiveX controls are an important class of COM objects that implement the component models of software.

ComponentWare   An approach to software engineering with software modules developed as objects with specific design frameworks and with visual editors both to interface to properties of each module and to link modules together.

CORBA   Common Object Request Broker Architecture. An approach to cross-platform, cross-language distributed objects developed by a broad industrial group, the OMG. CORBA specifies basic services (such as naming, trading, persistence) and the protocol IIOP used by communicating ORBS.  It is developing higher level facilities that are object architectures for specialized domains.

DCOM   Distributed version of COM.

DII   Dynamic Invocation Interface. An interface defined in CORBA that allows the invocation of operations on object references without compile-time knowledge of the objects' interface types.

DSI   Dynamic Skeleton Interface.  An interface defined in CORBA that allows servers to dynamically interpret incoming invocation requests of arbitrary operations.

EJB   Enterprise Javabeans.  Enhanced Javabeans for server-side operations with capabilities such as multi-user support. A cross-platform component architecture for the development and deployment of multi-tier, distributed, scalable, object-oriented Java applications.

Encoding   The bit and byte format and representation of information

Event   A noteworthy state change of an object or signal that involves the behavior of an object. An event can signal the creation, termination, classification, declassification, or change in value of an object. For example, the creation of a new circuit design or the debit of $300 to a particular account.

Exception   An indication that some invariant has not, or cannot, be satisfied.  Mechanisms for handling exceptions are often added to OO programming languages and environments. For example, Java, C++, and CORBA all have built-in exception handling.

HTTP   Hyper Text Transport Mechanism. A stateless transport protocol allowing control information and data to be transmitted between web clients and servers.

IDL   Interface Definition Language.  A language, platform, and methodology-independent notation for describing objects and their relationships. IDL is used to describe the interfaces that client objects call and that object implementations provide.

IIOP   Internet Inter Orb Protocol.  A state protocol allowing CORBA ORBs to communicate with each other and transfer both the request for a desired service and thereturned result.

IR   Interface Repository. A container, typically a database, of OMG IDL interface definitions. The interface to the interface repository is defined in the CORBA specification. Implementations of this interface are supplied by CORBA vendors.

Javabean   Part of the Java 1.1 enhancements defining design frameworks (particularly naming conventions) and inter-Javabean communication mechanisms for Java components with standard (Bean box) or customized visual interfaces (property editors). Javabeans are Java's component technology and in this sense are more analogous to ActiveX than either COM or CORBA. However, Javabeans augmented with RMI can be used to build a "pure Java" distributed-object model.

JDBC   Java Data Base Connection. A set of interfaces (Java methods and constants) in the Java 1.1 enterprise framework that defines a uniform access to relational databases. JDBC calls from a client or server  a Java program link to a particular "driver" that converts these universal database access calls (establishes a connection, SQL query, etc.) to the particular syntax needed to access essentially any significant database.

Jini   Sun's protocol for devices to identify each other using TCP/IP protocol. It will be used in small devices such as telephones to  allow a new device to be plugged into the system while everything is running. This device automatically finds out about everything else on the net and allows other devices to find it.

Learner   An individual engaged in acquiring knowledge or skills within a learning technology system. For example, students in traditional classrooms.

Learner information   The intersection of general learning technology information and human information for learners or learner entities.

Learner information   The intersection of general learning technology information and human information for learners or learner entities.

Learner performance information   Information about a learner's past, present, and future performance that is created and used by learning technology components to provide optimum improved or optimized learning experiences.

Learner personal information   Information about a learner that is not directly related to the measurement and recording of learner performance, nor or the advancement or progression of the learner, nor or the configuration and accommodation of the learner.

Learner portfolio information   A representative collection of a learner's works or references to them that is intended for illustration and justification of his or her abilities and achievements. Note: This type of information is intended for non-automated interpretation, such as a human interpretation.

Learner preference information   Information about a learner that is intended to improve human-computer interactions but allowing information technology systems and learning technology systems to adapt or accommodate to the learner's specific needs. Preferences may be explicitly identified by the learner or inferred from the learner’s behavior. Examples: interface features, technical features.

Learner profile   Information about a learner for specific learning technology components, learning technology applications, and learner administration. A subset of learner information, in general.

Learner relations information   Information about the a learner's relationship to other users of learning technology systems. Examples: teachers; proctors; other learners.

Learner security information   Information about a learner's security credentials. Examples: passwords; challenge/responses; private keys; public keys; biometrics.

Learning management system   A system that (1) schedules learning resources; (2) assists, controls, and/or guides the learning process; and (3) analyzes and reports learner performance.

Legacy System   A production system designed for technology assumptions that are no longer valid or  that are expected to become invalid in the foreseeable future. When deploying new applications or new system  architectures, a legacy system is one that may be accessed, but will typically not be modified to support new architecture..

Object   Anything that can be referred to; anything that can be identified, named, or perceived as an object; anything to which a type applies; an instance of a type or class. An instance of a class is comprised of the values linked to the object (Object State) and can respond to the requests specified for the class.


Object Interface   The set of requests that an object can respond to, i.e., its behavioral specification. An object interface is the union of the interfaces of the object's type interfaces.

Object Web   The evolving systems' software middleware infrastructure achieved by merging CORBA with Java. Correspondingly, merging CORBA with JavaBeans gives Object Web ComponentWare, which is expected to compete with Microsoft's COM/ActiveX architecture.

OMG   Object Management Group.  An organization of over 700 companies that is developing CORBA through a process of  proposal calls and the development of consensus standards.

ORB   Object Request Broker. Used in both clients and servers in CORBA to enable remote access to objects. ORBs are available from many vendors and communicate via IIOP protocol.

ORBacus    An example implementation of an OMG CORBA specification.

Proxy   An object that is authorized to act or take action on behalf of another object.

Request   An event that is the invocation of an operation. The request includes the operation name and zero or more actual parameters. A client issues a request to cause a service to be performed. Also associated with a request are the results, which can be returned to the client. A message can be used to implement (carry) the request and an results.

SCORM   The Sharable Courseware Object Reference Model (SCORM) is a set of interrelated technical specifications built upon the work of the AICC, IMS, and IEEE to create one unified “content model.”  These specifications enable the reuse of Web-based learning content across multiple environments and products

Server Object   An entity (e.g., object, class, or application) that provides a response to  a client's request for a service.

Servlet   An application designed to run on a server in the womb of a permanently resident CGI mother-program written in Java that provides services for it, much the way an Applet runs in the womb of a Web browser.

SSL   Secure Socket Layer.  A protocol for communication securely through sockets.

Supervisor  In the VCM, the users  that performs a number of in-structional functions.

 UML   Universal Modeling Language.  A modeling technique designed by Grady Booch, Ivar Jacobson, and James Rumbauch of Rational Rose. It is used for OOAD (Object Oriented Analysis and Design) and is supported by a broad base of leading industries. It merges the best of the various notations into one single notation style.

Web Client   Originally, web clients displayed HTML and related pages but now support Java Applets that can be programmed to give web clients the necessary capabilities to support general enterprise computing. The support of signed applets in recent browsers has removed crude security restrictions that handicapped the previous use of applets.

Web Servers   Web Servers  originally supported HTTP requests for information, basically HTML pages, but included the invocation of general server side programs using the very simple but arcane CGI  (Common Gateway Interface). A new generation of Java servers have enhanced capabilities, including server-side Java program enhancements (Servlets) and  the support of  permanent communication channels.

XDR   External data representation.  A protocol for sending data among heterogeneous architectures that was developed by SUN Microsystems.

XML   Extensible Markup Language.  A W3C-proposed recommendation.     Like HTML, XML is based on SGML, an International Standard (ISO 8879) for     creating markup languages. However, while HTML is a single SGML document     type, with a fixed set of element type names (AKA "tag names"), XML is a simplified profile of SGML.


[1] Another option is to customize XSL template files.