A Structural Engineering Support System using Semantic Computing

A STRUCTURAL ENGINEERING SUPPORT SYSTEM
USING SEMANTIC COMPUTING

by

Thiti Vacharasintopchai

A dissertation submitted in partial fulﬁllment of the requirements for the
degree of Doctor of Engineering in Structural Engineering

Examination Commitee: Professor Vilas Wuwongse (Chairperson)
Professor Worsak Kanok-Nukulchai (Co-chairperson)
Dr. Pennung Warnitchai
Dr. B. H. W. Hadikusumo

External Examiner: Professor Kincho H. Law
Department of Civil and Environmental Engineering
Stanford University
Stanford, CA, U.S.A.

Nationality: Thai
Previous Degrees: B.Eng. (Civil Engineering)
Chulalongkorn University
Bangkok, Thailand
M.Eng. (Structural Engineering)
Asian Institute of Technology
Bangkok, Thailand

Scholarship Donor: RTG Fellowship

Asian Institute of Technology
School of Engineering and Technology
Thailand
December 2007

i

ACKNOWLEDGMENTS

I would like to thank the Royal Thai Government for granting me a
doctoral research fellowship at the Asian Institute of Technology. My pro-
found gratitude is due to Professor Vilas Wuwongse and Professor Worsak
Kanok-Nukulchai, my mentors and advisors who invariably gave highly
valuable guidance, inspirational suggestions, encouragement and support
throughout this research, to Professor Kincho H. Law from Stanford Uni-
versity who gave highly valuable suggestions as the external examiner, and
to Dr. William Barry, former doctoral advisor, master’s thesis advisor and
good friend, who introduced me to the world of scientiﬁc computing. I
would also like to express sincere appreciation for Dr. Pennung Warnitchai
and Dr. B. H. W. Hadikusumo, doctoral committee members, who gave
invaluable guidance and suggestions. I am grateful to Dr. Pannapa Her-
abat for her unfailingly support and to Dr. Wisit U. Pongsa, my former
employer, who gave several opportunities for conferences and professional
training which form the basis of the work in this dissertation. Dr. Vo-
ratas Kachitvichyanukul provided useful comments during the early stages
of this research. Wouter Rosingh was of great help particularly towards
the end. I wish to thank my parents, Chavalert Vacharasintopchai and
Benjawan Chatnarat, my younger brothers, Nithi and Nipith, family mem-
bers, Apsara Narat and Gallissara Agavatpanitch for their love, support
and encouragement. Thanks are due to friends, especially Dr. Dussadee
Satirasetthavee, Dr. Kannapa Pongponrat, Dr. Pongtawat Chippimolchai,
Neelawat Intaraksa and Anisara Pensuk for their selﬂess care and generos-
ity during my study at the Institute. Assistance from colleagues at the
Knowledge Representation Laboratory and the School of Engineering and
Technology, especially Yupa Soodjitporn and Worranuch Chumchat, are
appreciated. I am also grateful to the Asian Development Bank and the
Greater Mekong Subregion Academic Network for funding the knowledge
management part of this research.

ii

ABSTRACT

Practicing structural engineers are often faced with limitations in avail-
able computing tools as well as insufficient access to relevant knowledge for
design projects and engineering jobs. At the launch of a project the im-
mediate tasks invariably involve research, discussions with colleagues and
a series of computations. These tasks are repeated over and over again
until a final design is arrived at. The degree to which these tasks are well
facilitated in an engineering firm helps determine the productivity of its
engineers and thus the firm’s competitiveness in the industry. Semantic
computing technologies such as Semantic Web Services, Weblog, and Dig-
ital Library have substantial, hitherto untapped potential to improve the
productivity of structural engineers. This dissertation proposes a support
system for structural engineering using these semantic computing technolo-
gies. A system architecture and the key software components for it are
specified and prototype systems presented. The support system involves
two frameworks of software infrastructure: first, the Weblog and Digital
Library Framework for Engineering Knowledge Management (Blog+DL),
which uses Weblogs and Digital Libraries as core components of a collab-
orative structural engineering support system; second, the Semantic Web
Services Framework for Computational Mechanics (SWSCM), a method-
ology that unifies and utilizes scattered computing resources. Blog+DL
and SWSCM are complementary methodologies in the engineering process.
Blog+DL relies on SWSCM to integrate computing resources shared by sev-
eral parties. SWSCM can use Blog+DL to deploy and discover shared com-
puting resources, as well as to educate and exchange knowledge, including
peer opinions about underlying theories and user experiences. Two proof of
concept prototype systems were developed. The first system illustrates the
joint application of Blog+DL and SWSCM to build support systems that
facilitate the computationally oriented knowledge management process in
structural engineering. The second system illustrates the full potential of
SWSCM to facilitate a computationally intensive workflow that involves
heterogeneous engineering software components. Blog+DL and SWSCM
combine semantic computing technologies to build computer-aided engi-
neering tools that improve the productivity of individual engineers and
thereby enhance the competitiveness of engineering firms.

Keywords: engineering support systems, knowledge management, productivity,
software interoperability, Semantic Web Services, Semantic Web, Web Services,
Internet, ontologies, artificial intelligence

iii

TABLE OF CONTENTS

Chapter Title Page

Title Page i
Acknowledgments ii
Abstract iii
Table of Contents iv
List of Figures vi
List of Tables vii

1 Introduction 1
1.1 Motivation 1
1.2 Problem Statement 1
1.3 Objectives 2
1.4 Scope and Research Approach 2
1.5 Contributions 3
1.6 Organization 3

2 Literature Review 4
2.1 Knowledge Management 4
2.1.1 Definition and Significance 4
2.1.2 Classification of Knowledge 5
2.1.3 Theory of Organizational Knowledge Creation 5
2.1.4 Knowledge Management Process and Tools 6
2.2 Knowledge Management in Engineering Design Firms 6
2.2.1 DAR Knowledge Management System 7
2.2.2 ADD and Arup Intranet Systems 7
2.3 Weblogs and Digital Libraries 8
2.3.1 Weblogs 8
2.3.2 Digital Libraries 9
2.4 Web Services 10
2.5 The Semantic Web 12
2.5.1 Background 12
2.5.2 How the Semantic Web Works 13
2.6 Semantic Web Services 15
2.6.1 Background 15
2.6.2 How Semantic Web Services Work 16

3 Development of a Support System for Structural En-
gineering 18
3.1 Structural Engineers and Construction Projects 18
3.2 Design Principles 18
3.3 General System Architecture 20
3.4 Summary 21

4 A Structural Engineering Support System Using
Semantic Computing 22
4.1 Introduction 22

iv

4.2 Blog+DL Framework for Engineering Knowledge Man-
agement 22
4.2.1 Technology Application Framework 22
4.2.2 System Architecture 24
4.3 A Semantic Computing-Based Support System for Struc-
tural Engineers 27
4.3.1 Implementation 27
4.3.2 A Proof of Concept Prototype 29
4.4 Summary 31

5 Engineering Software Interoperability 44
5.1 Introduction 44
5.2 A Semantic Web Services Framework for Computa-
tional Mechanics 44
5.2.1 SWSCM Model 45
5.2.2 Components of SWSCM Service 46
5.2.3 Communication between SWSCM Services 47
5.2.4 Inference Engine and Reasoning Process 48
5.2.5 Matchmaking Algorithm 50
5.3 Application to Structural Analysis and Design 51
5.3.1 Implementation 51
5.3.2 A Proof of Concept Prototype 53
5.4 Summary 56

6 Conclusions 59
6.1 Conclusion and Discussion 59
6.2 Recommendations for Further Research 61

References 63

v

LIST OF FIGURES

Figure Title Page

2.1 Conceptual Web Services Model 10
2.2 Example of RDF Metadata Embedded in Web Page 14
2.3 Sample RDF Statements about Person 14
2.4 OWL-S Service Profile Description of GetSineValue Web Service 17

3.1 General Architecture of Support Systems 20

4.1 Blog, Digital Library, and Knowledge Conversion Modes 23
4.2 Blog+DL Framework for Computer-Aided Engineering Design 25
4.3 User-specific Initialization of Prototype 30
4.4 Blog Publication Features of Prototype 33
4.5 Blog Consumption Features of Prototype 35
4.6 Blog Socialization Features of Prototype 37
4.7 Knowledge Archiving Features of Prototype 38
4.8 Knowledge Retrieval Features of Prototype 41

5.1 Semantic Web Services Model 45
5.2 Semantic SOAP Message 46
5.3 Components of Semantic Web Service 46
5.4 System Configuration of Prototype 52
5.5 Windows Mobile User Interface 54
5.6 Sample Request Message from PDA Client 55
5.7 OWL Definition of ASTM A36 Steel 56
5.8 OWL Definition of Pinned Node 57
5.9 Output Screen on PDA Client 57

vi

LIST OF TABLES

Table Title Page

6.1 Features Comparison with Existing Work 60

vii

CHAPTER 1
INTRODUCTION

1.1 Motivation
Two obstacles that practicing structural engineers often face in a design project
are limitations in available computing tools and limited access to relevant knowledge.
Several tasks are involved when a project is launched, in terms of research, discussion
with colleagues and computational workflow. These tasks are repeated over and over
again until a final design is arrived at. If they are not well facilitated, multiple bottle-
necks will occur and an engineering firm suffers low productivity among its engineers.
Not all engineering firms can afford expensive design software packages, and
the software itself often does not fully address their needs. Some firms may share a
commercial software package or use it on a pay-per-use basis; whereas others develop
software or subroutines in-house. Some in-house tools may be so well-developed or
have such well-developed components that it is more productive for engineers to reuse
them rather than to develop new systems from scratch. However, both commercial
and in-house tools are typically based on specific assumptions that need to be satisfied
beforehand, and their input and output data are usually based on specific conventions
and formats that must be strictly followed to obtain sensible results. As the number
of shared tools grows, it becomes difficult for users to find appropriate tools that fit
their design problems best, and it is tedious for them to manually adapt the potentially
incompatible input and output data of the tools to a computational workflow, such as
the discretization–analysis–postprocessing process in finite element methods.
In terms of knowledge management, on the launch of a new project, file cabi-
nets are often searched to retrieve reference documents of similar past projects. For
less common projects, significant time is typically invested in research to come up with
a solution strategy with the assistance of senior engineers or experts. This process
is time consuming and can become a bottleneck if not well facilitated. Many knowl-
edge management solutions are available in the literature or as commercial software
packages. However, most of them aim single-sidedly to capture tacit knowledge from
experts into a knowledge repository, rather than facilitating the sharing, discussing
and developing of new knowledge among individuals, which are key practices for the
sustainable improvement of knowledge in an organization according to the theory of
organizational knowledge creation (Nonaka and Takeuchi, 1995).

1.2 Problem Statement
This research aims to create a support system that improves the productivity of
structural engineers during the engineering process and, particularly, to alleviate the
two obstacles introduced earlier which hinder the performance of practicing engineers.
The realization of a support system that alleviates productivity obstacles involves ad-
dressing issues in the following areas:

1. Identification of the necessary features for a structural engineering support sys-
tem;
2. Knowledge management best practice according to knowledge management the-
ory; and
3. Effective methodologies for precise storage and retrieval of knowledge.

1

4. Development and sharing of in-house software tools as well as general-purpose
software packages;

5. Automated integration and execution of in-house and general-purpose software
tools;

Knowledge and tools in structural engineering as well as well-established com-
puter technologies, research, and standards are integrated in this study. The former are
required to ensure that the obstacles specific to structural engineering are addressed
and the latter are required to ensure the scalability and sustainability of the proposed
solution.

1.3 Objectives
The primary objective of this study is to develop a support system that improves
the productivity of structural engineers during the engineering process. To fulfill this
primary objective, the following secondary objectives are set:

1. To develop a framework and a prototype support system for structural engineers
that facilitates the improvement and mutual sharing of knowledge and skills in
the form of multimedia documents and computational tools. The framework
includes

(a) the identification of the features of a good knowledge management system
based on knowledge management best practice,
(b) a methodology to appropriately apply computing and information technolo-
gies in the knowledge management context, and
(c) methods and enabling modules that allow knowledge and computational
tools to be created, shared and improved by semantic computing technolo-
gies.

2. To develop a framework and a prototype system that enables the dissemina-
tion and use of intelligently unified computational tools shared in the structural
engineering knowledge system. Specifically, these include

(a) a procedure for structural engineers to develop and/or disseminate in-house
and commercial computational tools,
(b) the data model and format requirements for structural engineering compu-
tational tools to interoperate,
(c) the software services and components necessary to receive, interpret, and
process client requests, and
(d) the decision processes, knowledge, and rules by which a software agent
searches, evaluates, selects, and executes computational tools in a fully au-
tomated structural engineering workflow.

1.4 Scope and Research Approach
The scope of this study covers both knowledge management and computational
interoperability in the structural engineering context. Beyond the literature review

2

traditional for research of this type, the work is anchored in observations of real world
constraints on practicing engineers in engineering consulting firms as experienced by
the author. This led to the development of a Web-based proof of concept knowl-
edge management prototype as well as a set of software infrastructure for the sharing
and interoperability of computational tools. The former facilitates the creating and
sharing of knowledge in the form of multimedia files and computational services. It
allows users to exchange opinions about posted issues and services and to save some
of them to personal knowledge portfolios for later reference. It also allows access to
shared computational tools which can be invoked and interoperated by the software
infrastructure developed in this study. To demonstrate the full potential of this infras-
tructure, a prototype is separately developed to illustrate an intelligent collaboration
of many structural analysis software agents located on personal computers and parallel
computer clusters on the Internet to solve problems in linear elasticity.
The implementation of the system prototypes is based primarily on open-source
software tools. Several widely-accepted semantic computing standards, such as XML,
Web Services and Semantic Web standards sanctioned by the World Wide Web Con-
sortium (W3C), have been used during the development of the frameworks and the
prototypes. As the term “semantic” implies, artificial intelligence techniques are used
extensively.

1.5 Contributions
This study is aimed at improving the productivity of structural engineers in
the engineering process by dealing with computational workflow and knowledge man-
agement issues using semantic computing. The results of this study are an intelligent
structural engineering support system and two underlying frameworks: one which cre-
ates a collaborative environment where structural engineers, both novice and expert,
can share and sustainably improve their knowledge and another which turns the Inter-
net into a large platform for numerical analysis of structures where computational tools
individually developed and located on heterogeneous computer platforms intelligently
collaborate. It is expected that the application of the frameworks and support system
developed and demonstrated in this study can alleviate computational and knowledge
management problems in project design and other engineering work and thus improve
the productivity of individual engineers and the competitiveness of engineering firms.

1.6 Organization
The rest of this dissertation is organized as follows: Chapter 2 gives background
information about the theory, technologies and research that pertain to the develop-
ment of support systems for the engineering process. Chapter 3 describes common
obstacles in the structural engineering process, identifies necessary features and con-
ceptualizes the general design for structural engineering support systems. Chapter 4
presents the structural engineering support system using semantic computing and de-
scribes the architectural framework on which it is based. Chapter 5 addresses the
interoperability of computing resources which enables the sharing of computational
tools via support systems. It proposes a framework for the interoperation of structural
engineering software modules and demonstrates the framework by a system prototype
and a case study. Chapter 6 concludes this dissertation and discusses the benefits and
limitations of the presented work, with recommendations for further research.

3

CHAPTER 2
LITERATURE REVIEW

“Semantic Computing encompasses all aspects of computing where data is encoded,
processed, stored or transferred using techniques that communicate the meaning of the
data in addition to the data itself” (Boakes, 2007). An overview of the theory, technolo-
gies and research that pertain to the development of a support system for structural
engineering processes using semantic computing is presented in this chapter. First, the
theoretical background of knowledge management (KM) and two illustrative KM sys-
tems for engineering design firms are presented. Weblogs, a socialization and knowledge
sharing enabling technology for the Web, and Digital Libraries, a technology for fast
and accurate archiving and retrieval of electronic documents and multimedia content
in organizations, are discussed next. Web Services technology, a modern paradigm for
distributed computing on the Internet is then presented. Finally, the Semantic Web, an
emerging technology that brings artificial intelligence to the Web, and Semantic Web
Services, the joint application of Web Services and the Semantic Web, are discussed.

2.1 Knowledge Management
2.1.1 Definition and Significance
Knowledge management is a process for optimizing the effective application of
intellectual assets to achieve organizational objectives (Pollock, 2001). It is “the disci-
pline of creating a thriving work and learning environment that fosters the continuous
creation, aggregation, use and re-use of both organizational and personal knowledge in
the pursuit of new business value” (Cross, 1998, as cited in Quintas 2005). Knowledge,
the sets of processed information in a relevant context ready for understanding and
action (Turban and Aronson, 2001, as cited in Mezher et al. 2005), is an organization-
specific resource that is indispensable to create value for the organization (de Geytere,
2007). Knowledge differs from data and information in that it is “a function of a
particular stance, perspective, or intention” and involves an action “to some end;”
whereas information and data are the “necessary medium or materials for eliciting and
constructing knowledge.” Knowledge, data, and information however are alike in that
they are “context specific” and “relational” (Nonaka and Takeuchi, 1995, p. 58). In
decision making terms, knowledge contains the least “noise” whereas data contains the
most noise pertaining to a decision making process (Sheehan et al., 2005). Business
organizations that manage the creation and the application of knowledge more effec-
tively will be more competitive than others. Organizations that better manage and
share knowledge from past experiences of their peers will more effectively utilize their
resources to fulfill their mission. In the construction industry, which is very competi-
tive with low profit margins, although each project is invariably to some extent unique,
there are common processes which require engineers to seek out knowledgeable persons
“who know what” and to learn the “lessons learned” from them in a timely fashion
to minimize cost and stay competitive. This competitive environment has made KM
attractive in engineering consultant firms (Carrillo and Chinowsky, 2006; Mezher et al.,
2005). KM helps an engineering consulting firm to leverage their technical skills and
store them in a manner to speed up the work, and compete with others in such a way
that projects are produced with high quality in less time (Mezher et al., 2005). As the
world enters the “knowledge society,” the basic economic resource is no longer capital,

4

natural resources, or labor, but is and will be knowledge; and “knowledge workers”
will play an increasingly central role in society (Drucker, 1993; Nonaka and Takeuchi,
1995).

2.1.2 Classification of Knowledge
Knowledge can be classified into two kinds, namely, explicit knowledge and tacit
knowledge. Explicit knowledge can be articulated in formal language such as grammat-
ical statements, mathematical expressions, and manuals, which can be processed by a
computer with relative ease. Tacit knowledge on the other hand consists of personal
knowledge embedded in individual experience and involves intangible factors such as
personal beliefs, perspectives, and value systems, which are difficult to articulate com-
pletely in formal language. Because of its subjective and intuitive nature, it is difficult
to process or transmit acquired tacit knowledge in a systematic or logical manner (Non-
aka and Takeuchi, 1995). A large portion of human knowledge resides in the form of
tacit knowledge, and often tacit knowledge is regarded as the more important kind of
knowledge (Nonaka and Takeuchi, 1995, p. viii), as may be illustrated by the quotes
“If NASA wanted to go to the moon again, it would have to start from scratch, hav-
ing lost not the data, but the human expertise that took it there last time” (Brown
and Duguid, 2000, as cited in Quintas 2005) and “A master craftsman ... develops a
wealth of expertise at his fingertips after years of experience. But he is often unable
to articulate the scientific or technical principles behind what he knows” (Nonaka and
Takeuchi, 1995, p. 8).

2.1.3 Theory of Organizational Knowledge Creation
Knowledge is dynamic in nature. Nonaka and Takeuchi (1995) proposed the
theory of organizational knowledge creation which postulates four temporal modes of
conversion among tacit and explicit knowledge, namely, (1) socialization from tacit
knowledge to tacit knowledge, (2) externalization from tacit knowledge to explicit
knowledge, (3) combination from explicit knowledge to explicit knowledge, and (4) in-
ternalization from explicit to tacit knowledge, which constitute the “knowledge spiral”
that drives innovation in an organization.
Socialization, the process of “sharing experiences and thereby creating tacit
knowledge” such as shared mental models and technical skills, helps an individual
to acquire tacit knowledge directly from others through observation, imitation, and
practice in a field of interaction. Externalization, the process of “articulating tacit
knowledge into explicit concepts,” is the definitive process through which an individ-
ual attempts to conceptualize a mental image linguistically by metaphors, analogies,
concepts, hypotheses, or models. It is typically used in the process of concept cre-
ations and is “triggered” by meaningful dialog or collective reflection. Combination,
the process of “systemizing concepts into a knowledge system,” involves “networking”
different bodies of explicit knowledge in media such as documents or computerized
communication channels so that they are “crystallized” into a new piece of knowledge,
product, service, or methodology. Formal education and training in schools usually
takes the combination form of knowledge conversion. Internalization, the process of
“embodying explicit knowledge into tacit knowledge,” is the process through which
experiences of individuals, obtained through socialization, externalization, and combi-
nation, are internalized into their tacit knowledge bases in the form of shared mental
models or technical know-how. Internalization is closely related to “learning by do-

5

ing.” Documentation helps individuals to internalize what they experienced and also
facilitates the transfer of explicit knowledge to other people.
Each mode of knowledge conversion naturally produces different knowledge
types: socialization produces sympathized knowledge; externalization produces con-
ceptual knowledge; combination produces systemic knowledge; and internalization pro-
duces operational knowledge. These knowledge types interact with each other in a spiral
fashion: sympathized knowledge may become explicit conceptual knowledge through
socialization and externalization. The conceptual knowledge produced then becomes a
guideline for creating systemic knowledge through combination. The systemic knowl-
edge is next converted to operational knowledge through internalization. When an
individual socializes with colleagues and externalizes his knowledge, his operational
knowledge consequently triggers a new cycle of knowledge creation. Knowledge conver-
sion is a social process between individuals and is not confined within an individual. As
human cognition is the deductive process of an individual, who is never isolated from so-
cial interaction when things are perceived, the quality and quantity of tacit and explicit
knowledge are constantly “amplified” through shifts between these social conversion
processes. The four temporal modes of knowledge conversion among individual staff
members, departments, and divisions in an organization therefore constitute a “knowl-
edge spiral” that drives innovation in the organization. This knowledge creation model,
based on the theory of organizational knowledge creation, is sometimes called the
Socialization-Externalization-Combination-Internalization (SECI) model (de Geytere,
2007).

2.1.4 Knowledge Management Process and Tools
The KM process has been discovered rather than invented. Human activity is
inconceivable without knowledge. Even though the phrase “knowledge management”
came into common usage in the mid-1990s, the management of knowledge processes
began long before the term was coined, without the KM label (Quintas, 2005). KM
is different in each organization. There is no one right way and organizations develop
different approaches according to their values and objectives (Sheehan et al., 2005).
Some KM processes that function well may not have the KM label affixed. In contrast
“many so-called KM initiatives and tools that emerged in the late 1990s were less
concerned with addressing real knowledge issues than informal or existing processes
that are not so labeled” (Quintas, 2005). The focus of these initiatives and tools
was often on capturing explicit knowledge, rather than facilitating the conversion and
sharing of tacit and explicit knowledge, as in the SECI model. Their scope can thus
more properly be described as limited to “information management.”

2.2 Knowledge Management in Engineering Design Firms
Structural engineers often work for engineering design and consultant firms
rather than for construction managers or contractors. The nature of their work differs
from that in construction management, in which project planning and scheduling, re-
source allocation, procurement and document management are of more concern. Multi-
ple KM initiatives and tools for engineering design firms are available in the literature,
cf. Mezher et al. (2005); Jonathan Cohen & Associates (2004); Al-Ghassani et al.
(2005); Carrillo and Chinowsky (2006). Most of them, however, rely on proprietary
technologies that do not conform to Internet standards other than basic HTTP and

6

HTML for Web publication. Cross-system interoperability and advanced applications
of Web technologies, such as the Semantic Web (Berners-Lee et al., 2001) and Web feeds
(Wikipedia, 2007e) cannot therefore be implemented without major modifications of
these approaches. For illustrative purposes, the intranet knowledge management sys-
tems at the consultant firms DAR, ADD Inc. and Ove Arup and Partners are presented.
Weblogs and Digital Libraries, recent Internet technologies which overcome these diffi-
culties and can be integrated to form part of a more effective knowledge management
system, are introduced in the next section.

2.2.1 DAR Knowledge Management System
Mezher et al. (2005) developed a KM system at the DAR consulting firm. The
system was developed based on the KM cycle outlined by Turban and Aronson (2001)
which consists of six processes, namely, (1) the creation of knowledge from design
procedures prepared by engineers, lessons learned from construction sites and known
design hints and shortcuts, (2) the capturing of explicit and tacit knowledge created
by the first process, (3) the refining of knowledge through the review and approval
by group leaders and directors, (4) the storage of useful knowledge in a repository
for access by others in the organization, (5) the management of knowledge to main-
tain relevance, accuracy and currency, and (6) the dissemination of knowledge in a
format useful to users. The system consists of shared folders on the department’s file
server repository at the head office and a Web-based user interface that helps users
to access Word, Excel or AutoCAD documents and relevant e-mails in file server hi-
erarchies. Folders on the file server include design standards and manuals, archives of
completed project drawings and documents, pending project drawings and documents,
department-specific software and worksheets, libraries of AutoCAD shortcut menus
and programs as well as document templates and forms, all of which go through the
six KM cycle processes above. Access to these folders is restricted to relevant personnel.
The system is accessible from branch offices through the company’s intranet.
The system provides engineers with ready access to proper design templates and
procedures, as well as past site lessons. It prevents unnecessary re-work and allows en-
gineers to work both faster and more efficiently. Although the knowledge retrieval
process is assisted by a Web interface, it is not clear whether the knowledge creation
and review workflow and repositorial publication, which are part of the externalization
process, are computerized or automated. Metadata, as well as keyword and full-text
search features, which are information and tools necessary for precise retrieval of knowl-
edge from large knowledge bases or nested folder hierarchies, are not mentioned. The
system is therefore likely to not be scalable. Users will also find it difficult to review,
publish and precisely locate pieces of knowledge as the size of the knowledge base grows.
In addition, the system does not support communication means other than e-mails to
enable peer discussion and collaboration, leaving the socialization aspect of the SECI
model relatively unfacilitated.

2.2.2 ADD and Arup Intranet Systems
Jonathan Cohen & Associates (2004) developed intranet KM systems for ADD
Inc. and Ove Arup and Partners. ADD is a medium-sized multidisciplinary design firm
with 150 employees and three offices in Cambridge, Massachusetts, San Francisco and
Miami; whereas Arup is a very large engineering consultant firm with 5,000 employees
in 60 offices and 40 countries.

7

The ADD intranet is a unified Web portal to internal firm information. ADD
staff are directed to a “today” homepage when they enter the site where they are
presented with updated events, news and announcements stored on a shared calendar
database. From the homepage, they can navigate to information about firm standard
procedures, document templates and CAD details. Meetings can be scheduled using
the calendar function; and pictures and news items can be shared among colleagues
through personal pages assigned to each user. These personal pages help to create a
social network across the three offices. A list of experts within the firm is also posted
on the intranet so that they can be promptly consulted should the need arise.
The Arup intranet is an attempt to create a road map to the knowledge within
the firm, which is so large that it is difficult to keep track of all the expertise it possesses.
The Arup intranet is a central repository of the firm’s expertise that provides firm-
wide access to specifications in Word and PDF formats as well as CAD details with a
database that tracks how specific details have been implemented in projects. Arup’s
engineers can annotate their experience with the details and specifications, such as how
they worked in one project and failed in another, and so capture the feedback loop
between design and implementation. These experience records and the feedback loop
are invaluable information and practice that help distinguish Arup in the marketplace.
Employees of Arup are also allowed to “volunteer” information through personal Web
pages where they can quickly and easily share their interests and expertise with the
rest of the firm (Sheehan et al., 2005). In this way individuals can socialize and join
the SECI knowledge spiral which drives innovation at the firm.
The ADD and Arup intranet systems are examples of systems that conform to
KM best practice but rely on proprietary techniques. Cross-system interoperability
and advanced applications of recent Web technologies, which can further improve the
productivity of engineers, cannot thus be implemented easily.

2.3 Weblogs and Digital Libraries
2.3.1 Weblogs
A weblog, or a blog, is a Web-based journal publication which consists primarily
of periodic articles, normally in reverse chronological order (Wikipedia, 2007b). Each
article in a blog is called a blog entry. A person who publishes to a blog is called a
blogger . To publish to a blog is called to blog, the gerund form of which is blogging.
The totality of all blogs on the Internet is called the blogosphere (Wikipedia, 2007c;
Crystal, 2006).
A blog entry usually consists of (1) the title, (2) the content, (3) the categories,
tag names, or keywords of the content, (4) the date and time of publication, (5) the
comment section where readers and the blogger can discuss and exchange experiences
and opinions about issues raised in the blog entry, and (6) the trackback (Trott and
Trott, 2002) hyperlink where the readers who blog about the blog entry can notify
the original blogger about the existence and the content of their blogs. Hyperlinks,
comments, trackbacks, and citations in blogs create online social networks and promote
the externalization, socialization, and combinations of knowledge among the bloggers,
according to the SECI model.
Blogging is as easy as composing an e-mail. A blogger can blog about anything,
ranging from news items, personal opinions, pictures, hobbies, problems at work, to
random thoughts. A blog entry usually contains some hyperlinks to other Web docu-

8

ments which are the subjects of discussion. To publish a blog entry, a blogger logs on
to his account on the blog service provider Website, cf. Blogger (2007) and WordPress
(2007), clicks the compose button, types the content into a Web form, enters the title
and keywords for the content. Multimedia contents such as pictures, audio/video clips,
as well as Word and Excel files can be uploaded and included as attachments. When a
publish button is clicked, the blog entry is instantly published to the Internet and can
be accessed by friends, colleagues, and the public through regular Web browsers, Web
search engines, and dedicated blog reader software which conforms to the Really Sim-
ple Syndication (RSS) (RSS Advisory Board, 2006) or Atom (Nottingham and Sayre,
2005) Web syndication (a.k.a. Web feed) standards. A blogger can also send e-mails
to a secret e-mail address to publish the contents and picture attachments directly
from an e-mail client software on a laptop computer or a mobile phone. Blogging thus
allows people to share information on the Internet with very few barriers. Technorati,
Inc., a Web tracking company, reported that as of April 2007 there were 70 millions
blogs on the Internet, with 120,000 new blogs added each day, and 1.5 million posts per
day. The blogosphere grew from 35 to 70 million blogs in 320 days. Japanese, English,
Chinese, and Italian are the four major blogging languages—37 percent of the blogs
are in Japanese, 33 percent in English, 8 percent in Chinese, and 3 percent in Italian
(Sifry, 2007).

2.3.2 Digital Libraries
A digital library is an integrated set of services for capturing, cataloging, stor-
ing, searching, protecting, and retrieving information (Reddy et al., 1999). It comprises
focused collections of digital objects, including text, video, and audio, along with meth-
ods for access and retrieval, and for selection, organization, and maintenance of the
collections (Witten and Bainbridge, 2003), all of which support life-long learning, re-
search, scholarly communication and preservation (Wikipedia, 2007d). Digital libraries
were originally used to archive the digitized copies of rare documents, books, and the
pictures of historical objects so that they can be studied by people of later generations.
They have also recently been used as central repositories to preserve the works of in-
dividuals in an organization, so that they do not vanish with time and technological
obsolescence (DSpace, 2007b).
A digital library system is often developed such that it is accessible on the
Web, with the user interface resembling that of a Website. However, not all Websites,
even the ones that offer focused collections of well-organized material and appropriate
methods of access and retrieval, can be regarded as digital libraries, unless metadata
(data about data), which are important information to precisely catalog, locate and
retrieve pieces of information, are stored along with each object in the collections,
and the access, retrieval, and modification of metadata, as well as the retrieval of
objects based on them, are facilitated by the system (Witten and Bainbridge, 2003).
The metadata typically used in digital libraries include bibliographic information and
subject keywords, much as the indexing information used in physical libraries.
The users of a digital library system have one of two roles, namely, the publisher
and the reader. In some digital library implementations, such as the Greenstone Dig-
ital Library software (Witten et al., 2006), a librarian takes the sole publisher role to
capture (e.g., by scanning printed articles or typing in a word-processor), create cata-
loging metadata, store, manage, and publish the collections of digital objects; and the
general users take the reader role to browse, search, and consume information in the

9

Service
UDDI
Description
Service
WSDL
Registry

Pu
d

bli
Fin

sh
Service Service
Bind
Consumer Provider
SOAP Web Service

Figure 2.1: Conceptual Web Services Model

collections. In other digital library implementations, such as DSpace (MIT Libraries
and Hewlett-Packard Company, 2007) which is based on an institutional repository
concept, select users may also publish to some collections in the repository, thus taking
both the publisher and the reader roles. Computers and users can locate objects across
heterogeneous digital library systems by using standard query protocols, such as the
Open Archives Initiative Protocol for Metadata Harvesting (OAI-PMH) (Lagoze et al.,
2004). When a digital library system is deployed in an organization, it functions as a
knowledge base which stores explicit and externalized knowledge from the users. The
metadata-rich search feature of the system also allows users to precisely and promptly
locate pieces of knowledge which can be used to solve their specific problems in a timely
fashion. In this sense, a digital library system can thus promote the combination and
internalization of knowledge in the SECI model.

2.4 Web Services
Web Services (W3C, 2004d) is a modern paradigm of distributed computing
that uses the Internet as the communication medium. Its fundamental concept is to
build computer software by making use of remote procedure calls (RPC) over the In-
ternet or an intranet. It differs from other distributed computing technologies, for
example CORBA (OMG, 2005) and DCOM (Microsoft, 1996), in the use of platform-
independent standards, based on HTTP (Fielding et al., 1999) and Extensible Markup
Language (XML) (Fallside, 2001), which allows service providers to hide the imple-
mentation details from clients. The clients need to know the URLs (Berners-Lee et al.,
1994) of the services and the data types for method calls, but not the details of the
service implementation in order to use them (Basha et al., 2002). Web Services is dif-
ferent from the common term “web services” (spelled in lower case) adopted by some
researchers, which simply refers to the “ability to access many services through the
Internet” (Chen et al., 2006) via an arbitrary communication protocol without regard
to the specific World Wide Web Consortium (W3C) standards that are introduced in
this section.
The typical architecture of Web Services is shown in Figure 2.1. Three roles,
namely, a service provider, a service consumer, and a service registry, and three oper-
ations, namely, publishing, finding, and binding, are involved.
A service provider is an entity that creates Web services. Typically, it exposes

10

certain functionality in their organization as a Web service that is made available for
any other organization to invoke. To be effective, a Web service needs to perform
two tasks. First, it needs to describe the Web service in a standard format that
is understandable by all organizations that will be using that Web service. Next it
needs to publish the details about its Web service in a central registry that is publicly
accessible.
A service consumer is any organization that uses Web services provided by
a service provider. It knows the functionality of a particular Web service from the
description made available by the service provider. To retrieve those details, a service
consumer searches the registry where the service provider has published its Web service
description. The service consumer can then obtain the description of the required
mechanism, bind to the service provider’s Web service and invoke that Web service.
A service registry is a central location where service providers list their Web
services, and where service consumers search for them. Service providers usually pub-
lish their Web service capabilities in the service registry for service consumers to find
and later, if interested in the provided service, bind to them. Examples of information
typically stored in the service registry include the detailed description of an organiza-
tion, the Web services that it provides, and a written description of each Web service
as well as the input and output formats for its invocation.
Web services take multiple roles on different occasions. A Web service takes
the “provider” role when it accepts requests from consumers to be processed. It takes
the “consumer” role when it delegates its tasks to other Web services. A Web service
provider takes the special role of “service registry” if it helps a service consumer to
discover a service provider for later binding. A client program that interacts only
with end users takes the consumer role when it requests a service from a Web service
provider.
The Web Services architecture aims to achieve interapplication communication,
irrespective of the programming language that the application is written in or the
platform that the application is running on, by three fundamental operations which
are “finding,” “binding,” and “publishing.” To make this happen, standards for each
of these three operations and a standard way for a service provider to describe their
Web services are needed.
The Web Service Description Language (WSDL) (W3C, 2001) is such a standard
that uses the XML data format to describe Web services. The WSDL document for a
Web service defines the methods that are present in the Web service, the input/output
parameters for each method, the data types, the network transport protocol used, and
the URLs where requests for services are to be submitted and from which the results
can be received.
The Universal Description, Discovery, and Integration (UDDI) standard (OA-
SIS, 2002) provides a way for service providers to publish details about their organi-
zation and the Web services that they provide to a central registry. It also provides a
standard for service consumers to find service providers and details about their Web
services. Publication of the details is the “description” part of UDDI and the location
of such details is the “discovery” part of it. Publication of these details is typically
achieved by advertising hyperlinks to their WSDL documents on UDDI service reg-
istries.
Communication between service consumers, service providers and service reg-
istries is through the Simple Object Access Protocol (SOAP) (W3C, 2000), which is a

11

lightweight XML communication mechanism to exchange information between appli-
cations, regardless of the operating systems, programming languages, or object models
employed in their development. SOAP may be visualized as a “carrier” of data encoded
in XML format.
Web Services technology has been involved in many areas of scientific comput-
ing, ranging from computational infrastructure developments (Chiu et al., 2002; van
Engelen, 2003) to finite-element analysis of a coupled fluid, thermal, and mechani-
cal fracture problem (Chew et al., 2003). One significant effort in this area is the
standardization attempt by the Global Grid Forum (GGF, 2003) to create the Open
Grid Service Architecture specification (OGSA) (Foster and Gannon, 2003) which is
the global standard for interoperation in the grid computing community (Foster, 2002;
Foster et al., 2001) and is the specification in which the SOAP and WSDL standards,
two important components of Web Services architecture, have been adopted. As there
are already a significant number of scientific computing projects that rely on grid com-
puting infrastructure, such as NASA’s Information Power Grid (IPG, 2003) and the
TeraGrid (Catlett, 2002), imposition of such a standard is likely to increase the number
of scientific computing applications that rely on Web Services technology significantly.

2.5 The Semantic Web
2.5.1 Background
The Semantic Web (Berners-Lee et al., 2001) is a vision for the next generation
of the Web on which information will be useful and meaningful not only for people
but also for computers. Computers will be able to understand pieces of information
on Web pages rather than merely presenting them to users, and will thus be able to
autonomously assist users in manipulating the information.
The Semantic Web is not a separate Web. Rather, it is an evolution of the
present Web into one in which information is given well-defined meaning (Berners-Lee
et al., 2001). Documents on the present Web can be transformed into Semantic Web
documents by augmenting them with metadata aimed at computers. Metadata is data
about other data that enables computers to determine the meaning of information
in a Web document by following hyperlinks to definitions of key terms and rules for
reasoning about them logically.
The Semantic Web relies on two existing technologies which are XML and the
Resource Description Framework (RDF) (Berners-Lee et al., 2001).
XML (Fallside, 2001), a generalization of HTML (W3C, 1999a), has been pro-
posed as the language for the publication of Web documents. XML allows Web pub-
lishers to add structure to their documents by using tags, which are hidden labels that
annotate Web pages or sections of text on a page. Computer programs can make use of
these tags in many ways, from rendering XML documents in a human-friendly format
on Web browsers, to mapping XML documents as data sources. However, XML tags
are not uniformly created and they can be used by different groups of people. There-
fore, their intended meaning must be precisely comprehended by the programmers in
order to develop programs that properly manipulate the XML tags. The intended
meaning of an XML tag will be understood if the name of the tag is drawn from a
common dictionary.
RDF (W3C, 1999b) is a language for describing information about resources
on the Web. Typically encoded in XML format itself, the RDF language allows Web

12

publishers to assert that particular things (Web pages, videos, etc.) have properties
(such as “is authored by” or “is a kind of”) with certain values (such as a thing,
a person, etc.), in a grammatical form similar to the subject, verb, and object of
an elementary sentence (Berners-Lee et al., 2001). An assertion that the Web page
“www.asce.org” “is authored by” “ASCE” is called an RDF statement, or a triple. A
set of triples with the same term or concept in a domain forms a definition of the term.
A collection of the definitions and the relations among terms in a domain constitutes
an ontology of the domain. In the context of the Semantic Web, an ontology is a
document or file that formally defines terms and their relations (Berners-Lee et al.,
2001). It can serve as a common dictionary from which XML tag names, and RDF
property names and values can be drawn.
Subjects, verbs, and objects in RDF statements are uniquely identified by Uni-
versal Resource Identifiers (URIs). The Uniform Resource Locators (URLs) that link
Web pages are the most common type of URI. URIs ensure that terms or concepts are
not just words in a document but are tied to a unique definition that anyone can find
on the Web (Berners-Lee et al., 2001). In addition, the object of an RDF statement
can be another RDF statement, so that complex RDF statements can have other RDF
statements nested inside in multiple layers. Figure 2.2 illustrates the RDF statements
serving as metadata about the Web page http://ce-res.blogspot.com/2005/10/
kmitlsut-reinforced-concrete-design.html. It asserts that the page “has title”
“Reinforced Concrete Design Worksheets,” is categorized in the “subjects” “RC De-
sign,” “Working Stress Design,” etc., and “is authored by” the person identified by the
URI http://stweb.ait.ac.th/~ st029284/foaf.rdf#thitiv.
Ontologies and RDF statements are typically stored on Web servers for later
retrieval by other parties in the same way as the publication of traditional Web doc-
uments. The statements about the author of the Reinforced Concrete Design Work-
sheet can be retrieved for display on a Web browser by accessing the URI http:
//stweb.ait.ac.th/~ st029284/foaf.rdf#thitiv. It is also possible to embed an
RDF statement that employs terms from specific ontologies inside hidden tags of Web
documents themselves (Palmer, 2002). Software tools such as the Jena Semantic Web
Framework (Jena, 2007) provide a means to retrieve ontologies and RDF statements
from Web sources and store them in a local knowledge base. The set of inference
rules that correspond to the semantics of the RDF constructs (W3C, 2004c) is prebuilt
into Jena. Answers to queries such as “give me a list of Web pages authored by
http://stweb.ait.ac.th/~ st029284/foaf.rdf#thitiv” can be obtained through
Jena’s Application Programming Interfaces (APIs) (Wikipedia, 2007a).

2.5.2 How the Semantic Web Works
The following examples illustrate how Web documents, XML, ontologies, and
RDF statements work on the Semantic Web.
Example 1
John Doe is at the Reinforced Concrete Design Worksheet Web page on the CE
Resources Website (Vacharasintopchai, 2005). He is interested in the link mentioned on
the page and wants to contact the author for more information about it. The author’s
e-mail address or contact information is not given on the Web page. How can John
find the author’s e-mail address?

Traditional Solution. On the Web page, John has to find the name of the author.

13

<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
xmlns:dc="http://purl.org/dc/elements/1.1/">
< rdf:Description
rdf:about="http://ce-res.blogspot.com/2005/10/kmitlsut-reinforced-concrete-design.html"
5 dc:title="Reinforced Concrete Design Worksheets">
<dc:subject rdf:resource="http://thitiv.blogspot.com/semblog/terms.owl#rc-design"/>
<dc:subject rdf:resource="http://thitiv.blogspot.com/semblog/terms.owl#working-stress-design"/>
<dc:subject rdf:resource="http://thitiv.blogspot.com/semblog/terms.owl#ultimate-strength-design"/>
<dc:subject rdf:resource="http://thitiv.blogspot.com/semblog/terms.owl#structural-members"/>
10 <dc:language rdf:resource="http://www.iso.org/iso/en/prods-services/iso3166ma/02iso-3166-code-
lists/list-en1.html#TH"/>
</ rdf:Description >
</rdf:RDF>
15 xmlns:dc="http://purl.org/dc/elements/1.1/">
< rdf:Description
rdf:about="http://ce-res.blogspot.com/2005/10/kmitlsut-reinforced-concrete-design.html">
<dc:author rdf:resource="http://stweb.ait.ac.th/~st029284/foaf.rdf#thitiv"/>
</ rdf:Description >
20 </rdf:RDF>

Figure 2.2: Example of RDF Metadata Embedded in Web Page

xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#"
xmlns:foaf="http://xmlns.com/foaf/0.1/">

5 <foaf:PersonalProfileDocument
rdf:about="http://stweb.ait.ac.th/~st029284/foaf.rdf">
<foaf:maker rdf:resource="#thitiv"/>
<foaf:primaryTopic rdf:resource="#thitiv"/>
</foaf:PersonalProfileDocument>
10
<foaf:Person rdf:ID="thitiv">
<foaf:givenname>Thiti</foaf:givenname>
<foaf:family_name>Vacharasintopchai</foaf:family_name>
<foaf:mbox rdf:resource="mailto:thitiv[at]gmail.com"/>
15 <foaf:homepage rdf:resource="http://thitiv.blogspot.com"/>
</foaf:Person>

</rdf:RDF>

Figure 2.3: Sample RDF Statements about Person

Then he has to go to a search engine, such as Google, and type in the author’s
name and a keyword such as “e-mail” to search for traces of an e-mail address.
If John is lucky enough, he gets the e-mail address after the ﬁrst search.

Semantic Web Solution. A set of RDF statements that describe the Web page (Fig-
ure 2.2) will usually be published in conjunction with the page content. A Se-
mantic Web-enabled Web browser will recognize the “is authored by” RDF state-
ment (the dc:author and the rdf:resource tags) and will be directed to the
RDF statements about the author. The Web browser can inspect the statements
about the author (Figure 2.3) and extract the author’s e-mail address from the
foaf:mbox statement automatically. This e-mail address will be suggested by
the Web browser as the e-mail address that John is looking for.

14

Example 2
The RC design Web page is categorized on the Website under a keyword “struc-
tural members.” If John wants to see more articles about “structural members” from
the same author, how would he do it?

Traditional Solution. On the Web page, the author provides a link to an index of
articles in the “structural members” category. If John clicks on the “structural
members” link, he is directed to a list of articles about “structural members.”
Had the author not provided such an index, John would have to search for a list
of articles with “structural members” and the author’s name as the keywords.
However, John is not likely to encounter an article about “beams,” although
obviously it is a kind of structural member, unless “beams” is also explicitly
added to the keyword list.

Semantic Web Solution. The Web browser will recognize the “subject” statements
(the dc:subject tag) in the set of RDF statements about the RC design Web
page (Figure 2.2). When requested to ﬁnd more Web pages whose dc:subject
property is http://thitiv.blogspot.com/semblog/terms.owl#structural-members
and whose dc:author property is http://stweb.ait.ac.th/~ st029284/foaf.
rdf#thitiv, the Web browser can browse through a list of RDF statements,
then select the Web pages whose dc:author and dc:subject are the same as,
or related to, the given ones, and then present the result accordingly. If the Web
browser ﬁnds an RDF statement which asserts that http://thitiv.blogspot.
com/semblog/terms.owl#beams is a “subclass of” (W3C, 2003) http://thitiv.
blogspot.com/semblog/terms.owl#structural-members, it will also include
the article about “beams” in the results, as it can be deduced by an inference
engine that beams are a type of structural member. In technical terms, we say
that the concept “structural members” subsumes the concept “beams.”

It should be noted here that the term “RDF statement,” used earlier, collectively
refers to the statements that are represented by the RDF language (W3C, 2003) as well
as the Web Ontology Language (OWL) (W3C, 2004a), which is an extension of the RDF
language for more comprehensive representation of knowledge on the Semantic Web.
OWL provides several additional constructs, such as “equivalent class,” “intersection
of,” “union of,” “inverse of,” and “disjoint with,” that more precisely capture bodies
of knowledge over those available in the RDF language. Henceforth, the term “OWL
statement” may be used when it is desirable to emphasize the employment of OWL’s
additional constructs in the statement.

2.6 Semantic Web Services
2.6.1 Background
Web Services introduced a new model of distributed computing that uses the
Internet as the communication platform. The present technology based on WSDL,
SOAP, and UDDI allows automatic publication, discovery, and execution of services at
the syntactic level. However, WSDL documents, which play a key role in the interop-
eration of Web services, are XML documents and therefore inherit the same semantic
problem of XML discussed earlier. To make proper use of the Web services advertised
by WSDL documents, programmers have to know in advance the intended meaning

15

of the custom tags that specify the input and output schemas as well as the names of
services that a Web service provides. Like other XML documents, WSDL documents
are usually created and used by different groups of people. Unless supported by on-
tologies, it would be difficult for a computer program looking for a GetSineValue Web
service that calculates the sine of a “degree” angle to autonomously execute and get
the correct sine value from an advertised TrigonometricSine Web service that takes
“radian” angles as the input, even though both of them are described as Web services
that calculate the sine values of angles.
Semantic Web Services, the idea of augmenting Web service descriptions using
Semantic Web technology (Burstein et al., 2005), were introduced to address this kind
of problem and to facilitate the autonomous publication, discovery, and execution of
services at the semantic level. Semantic Web service description languages, such as
OWL-S (OWL–S, 2003) and Web Service Modeling Ontology (WSMO) (Roman et al.,
2005), were proposed as abstractions of syntactic Web service description languages
such as WSDL, which are used in current Web Services technology. They are meant
to be used with semantic matchmakers, which are the software agents that accept and
keep track of the descriptions of available services from providers and match them
against the requirements from service consumers (Burstein et al., 2005). Matchmakers
enhance the role of UDDI service registries in the Semantic Web Services architecture.
OWL-S is among the most widely used semantic Web service description lan-
guages and was submitted to W3C for possible standardization (Martin et al., 2004).
It describes the categories, the inputs, the outputs, and the consequences of Web ser-
vices in terms of concepts defined in OWL ontologies. It also provides the grounding
constructs for specialization into WSDL constructs for compatibility with existing Web
services, which are described by WSDL documents. OWL-S is used as the semantic
Web service description language in this research.

2.6.2 How Semantic Web Services Work
To illustrate how Semantic Web Services can help a program looking for the
GetSineValue Web service to get the correct value of sine from the TrigonometricSine
Web service, consider the OWL-S description of GetSineValue in Figure 2.4. The
“service profile” description, which is the proposed information for use by matchmakers
(Sycara et al., 2003), is presented. Figure 2.4 shows that:

1. The GetSineValueService (which is the service that the program is looking for)
is a kind of SineFunctionEvaluator defined in the profileHierarchy taxon-
omy;

2. The service takes one input parameter “Angular Degree” the definition of which is
available at http://std.swscm.org/200407/quantities.owl#AngularDegree,
and gives one output the definition of which is available at http://std.swscm.
org/200407/quantities.owl#UnaryConstantFunctionQuantity; and

3. The effect of this service is defined by http://std.swscm.org/200407/Math-
OperationEffects.owl#TrigonometricFunctionValueEffect.

In the Semantic Web Services architecture, when a service consumer needs a
service from a provider, it creates an ideal service profile, such as the one in Figure 2.4,
and submits it to a matchmaker as a request for recommendation (Sycara et al., 2003).

16

<rdf:RDF>
<service:Service rdf:ID="GetSineValueService">
<service:presents rdf:resource="#GetSineValueServiceProfile"/>
</service:Service>
5
<profileHierarchy:SineFunctionEvaluator rdf:ID="GetSineValueServiceProfile">
<service:isPresentedBy rdf:resource="#GetSineValueService"/>

<profile:hasInput>
10 <process:Input>
<process:parameterType
rdf:resource ="http://std.swscm.org/200407/quantities.owl#AngularDegree"/>
</process:Input>
</profile:hasInput>
15
<profile:hasEffect rdf:resource=
"http://std.swscm.org/200407/MathOperationEffects.owl#TrigonometricFunctionValueEffect"/>

<profile:hasOutput>
20 <process:UnConditionalOutput>
<process:parameterType
rdf:resource ="http://std.swscm.org/200407/quantities.owl#UnaryConstantFunctionQuantity "/>
</process:UnConditionalOutput >
</profile:hasOutput>
25
</profileHierarchy:SineFunctionEvaluator >
</rdf:RDF>

Figure 2.4: OWL-S Service Profile Description of GetSineValue Web Service

From the profiles registered by various service providers, the matchmaker selects the
one that most closely matches the ideal profile requested, and recommends it to the
consumer for further binding between the two parties.
As the matchmaker processes the request, suppose that it had searched the list
of service profiles advertised by service providers but could not find any service that
matched perfectly. However, there is a TrigonometricSine Web service that advertises
itself as a kind of SineFunctionEvaluator which gives a UnaryConstantFunction-
Quantity output and the TrigonometricFunctionValueEffect effect. The service
takes the input parameter defined by http://www.math.org/terms.owl#RadianAngle
rather than the http://std.swscm.org/200407/quantities.owl#AngularDegree. If
there are some assertions that the “radian angle” and the “angular degree” concepts
are related in some ways, for example:

1. Both “radian angle” and “angular degree” are subclasses of http://www.math.
org/terms.owl#PlaneAngle; and

2. The “angular degree” “has conversion factor” of π/180 and ”has reference unit
of measurement” of http://www.math.org/terms.owl#RadianAngle,

by populating these assertions into the knowledge base of an inference engine, and
by querying whether the pairs of service types, input parameters, output parameters,
and effects, subsume each other (Sycara et al., 2003), the matchmaker will be able to
determine that the two Web services near perfectly match, and could thus recommend
this TrigonometricSine Web service to the consumer. The consumer will be able to
use the “has conversion factor” assertion to convert its angular degree into the radian
angle unit expected by the service provider, and will be able to initiate a request to
the TrigonometricSine Web service and finally get the correct sine function value.

17

CHAPTER 3
DEVELOPMENT OF A SUPPORT SYSTEM FOR
STRUCTURAL ENGINEERING

3.1 Structural Engineers and Construction Projects
Construction projects start when owners decide to have buildings constructed to
satisfy their needs. Through bidding processes or fast track approaches, an architect is
contracted, the owner’s requirements collected and a preliminary design agreed upon.
A series of drawings and specifications which consist primarily of architectural and
structural work are indispensable for the materialization of a project. Tender drawings
and specifications are required for bidding to select a general contractor for the project.
Detailed drawings from designers and shop drawings from field engineers are necessary
for work at the construction site to proceed.
Structural engineers play a critical role in all these construction project phases.
At the outset, architects need a structural engineers’ opinion about viable structural
systems that are appropriate for preliminary designs. Structural engineers are con-
sulted for the estimated size of columns, beams and slabs before architectural drawings
are developed for tender. Detailed foundation designs are often the first construction
drawings requested from structural engineers, even though the information for them
is last obtained in the analysis and design process. In addition, field engineers need
prompt advice from structural designers when unexpected difficulties occur at the con-
struction site to minimize expenses from work delays.
Sound structural analysis and design is central to the safety and economy of a
construction project; yet structural engineers are often on the critical path with tight
schedules and are allocated very limited time for their tasks. Two obstacles that hinder
the productivity of structural engineers in dealing with their tasks are limitations in
available computing tools and limited access to relevant knowledge. A support system
that improves the productivity of structural engineers during the engineering process
is therefore highly advantageous, as postulated in the first chapter.

3.2 Design Principles
Structural engineering departments typically have both senior engineers with
extensive practical experience and junior engineers (or fresh graduates) with little pro-
fessional experience. Many kinds of tools are used in the engineering process. Junior
engineers are often more acquainted with relatively new technologies, whereas senior
engineers have a wealth of practical insight and often know useful shortcuts. Some
commercial software packages are licensed and many departments have special-purpose
software tools that have been individually developed in-house. Commercial software
and in-house tools are usually shared among engineers in the department. The output
from structural engineers is commonly in the form of hand sketches or CAD drawings,
supported by calculation sheets. On completion of projects, relevant documents are
collected on file servers or in cabinets for future reference.
To improve the productivity of structural engineers in dealing with their tasks,
good support systems should at least have the following features:

1. References and Premises Engineering is an applied science by nature. It deals
with the application of pure scientific knowledge to solve practical problems.
When structural engineers are assigned tasks at the launch of new projects, ref-

18

erence documents, such as design manuals, theoretical notes and data sheets, as
well as examples from similar past projects, are the first information needed to
come up with solution strategies. Good support systems for structural engineers
should thus facilitate the efficient storage, as well as fast and precise retrieval, of
reference documents and premises to minimize the engineers’ start-up time.

2. Knowledge and Tools Sharing Engineers study project requirements and pre-
dict assumptions to simulate the behavior of structures under critical usage con-
ditions. Based on the mechanical properties of construction materials, structural
members are designed such that they can withstand severe conditions at reason-
able levels of safety. To maximize the safety and economy of structural engineer-
ing products and arrive at proper rather than over- or under-designs engineers
should have access to the best possible set of knowledge and tools relevant to their
tasks. Methodologies, tips and techniques, as well as software tools licensed from
commercial vendors or developed in-house, should be accessible to as many fel-
low colleagues as possible. Good support systems for structural engineers should
second facilitate the creating, sharing and disseminating of knowledge and tools
among structural engineers so that they are well equipped for their work.

3. Computational Workflow Commercial and in-house software tools are typically
based on specific assumptions that need to be satisfied. Their input and output
data are usually based on specific conventions, formats and units of measurement
that must be strictly followed to obtain sensible results. Such software is often
chained into a computational workflow to address specific requirements unsup-
ported by general-purposes software packages. It is tedious and error-prone for
the engineers to adapt the potentially incompatible input and output data of
tools in the chain manually. It is also difficult for users in workplaces where nu-
merous software tools are shared to find the tools that fit their specific problem
best. Good support systems for structural engineers should thus thirdly facilitate
the interoperation of software tools and allow them to be chained such that the
least possible user intervention is required to use them effectively and prevent
blunders.

4. Discussion and Collaboration Engineering is collaborative in nature. Senior
engineers in structural engineering departments have a wealth of practical insight
and experience, whereas junior engineers have fewer but are more acquainted
with modern technologies and design techniques. In general, the two exchange
knowledge and learn from each other. Senior engineers give guidelines and opin-
ions about approaches to a solution. Colleagues give suggestions about tools
and techniques appropriate for specific project types. Although members of en-
gineering teams usually meet and discuss formally on a regular basis, human
knowledge is amplified during informal socialization, as postulated by the theory
of organizational knowledge creation in Chapter 2. To assist structural engineers
in exchanging and amplifying their knowledge in other than lunch and hallway
meetings, good support systems for structural engineers should fourthly facilitate
informal discussion and collaboration among peers. The systems should allow ex-
changes of audio/video clips, worksheets, pictures and drawings for effective peer
communication. To encourage user participation, the tool for discussion and col-
laboration should be easy to use. The systems should also have an archiving

19

Figure 3.1: General Architecture of Support Systems

feature to make both past projects and the group thinking processes behind the
solutions arrived at, accessible to colleague engineers in the future.

5. Accessibility and Interoperability Support systems are not useful if they are
difficult to access. They will not wide gain acceptance and adoption in the
industry if they lack interoperability with other systems. In addition to the four
features above, good support systems for structural engineers should therefore
follow open standards and be platform neutral for cross-system interoperability
and accessibility. Users on different computing platforms (such as Windows, Mac
OS and Linux) should be able to access the systems indifferently. Ubiquitous
access to the systems, especially from remote construction sites, would also be
advantageous for timely access to knowledge and tools to assist with unforeseen
problems on-site.

3.3 General System Architecture
The semantic computing technologies reviewed in Chapter 2 provide various
functionalities which can be combined to meet the five requirements for good struc-
tural engineering support systems. This section proposes a general system architecture
which utilizes semantic computing technologies to build such a system. The system
architecture consists of five components as depicted in Figure 3.1. Each component is
labeled with numbers (1), (2), . . . , (5), which correspond to item numbers of the design
principles. For example, (1) refers to References and Premises and (2) refers to Knowl-
edge and Tools Sharing. The relationships between these components and the design
principles can be explained as follows:

Digital Library The need for support systems to facilitate efficient storage and fast,
precise retrieval of premises and reference documents matches well with the meta-
data rich content archiving and retrieval features of digital libraries. Digital li-
braries are specifically developed to store large amounts of electronic documents
for archival purposes and to allow them to be precisely located and retrieved
with relative ease. Digital Library technology is therefore appropriate for the
References and Premises requirement. The Web access and conformance to stan-
dard query protocols of digital library systems make them appropriate for the
Accessibility and Interoperability requirement.

20

Weblogs The need for support systems to facilitate the creating, sharing, and dis-
seminating of knowledge and tools matches well with the publication and social
networking features of Weblogs. Blogging allows engineers to instantly document
and disseminate among peers new ideas, software tools and techniques, and to
conveniently receive focused feedback on particular topics in the form of com-
ments and trackbacks. Engineers can also blog about work plans and difficulties
arising in particular projects to solicit comments and suggestions from colleagues.
Weblogs, in this regard, are therefore appropriate both for the Knowledge and
Tools Sharing and the Discussion and Collaboration requirements. The confor-
mance of Weblog servers to Web feed standards and the convenient means of
publication offered by Weblogs make the technology especially appropriate for
support systems, from the Accessibility and Interoperability point of view. In
addition, the bibliographic information of blog entries is important metadata
which contributes to the fast and precise retrieval of knowledge in blog content,
according to the References and Premises principle.

Web Services The need for support systems to facilitate with the least user inter-
vention the interoperation and chaining of software tools matches well with the
capability of Web Services and Semantic Web Services to help unify and utilize
scattered computing resources. Web service providers and consumers include,
but are not limited to, software tools on parallel computer clusters, personal
computers, laptops and smartphone devices. With premises from service meta-
data (i.e. service descriptions) and ontologies, the artificial intelligence inherent
in Semantic Web Services technology also helps to intelligently discover and unify
heterogeneous computational services and to reduce the necessity for user inter-
vention during workflow processes. Web Services and Semantic Web Services
technologies, in this regard, are thus particularly appropriate for structural engi-
neering support systems, in terms of Computational Workflow, Accessibility and
Interoperability, as well as Reference and Premises.

3.4 Summary
This chapter has described common obstacles in the structural engineering pro-
cess and proposed five principles for the design of support systems to assist engineers
in dealing with their tasks more productively. Semantic computing technologies were
identified as enabling tools for such systems and a general architecture for the develop-
ment of structural engineering support systems has been proposed. The methodology
to create support systems from these technologies are presented in the next two chap-
ters, along with proof of concept prototypes. Metadata and ontologies as well as their
annotation in data and knowledge are central to all aspects of the development, hence
the term “semantic computing.” Their importance is illustrated throughout.

21

CHAPTER 4
A STRUCTURAL ENGINEERING SUPPORT SYSTEM USING
SEMANTIC COMPUTING

4.1 Introduction
Weblogs (a.k.a. Blogs) and Digital Libraries, information technologies deploy-
able on the Internet or an intranet, can be combined to build a new generation of
knowledge management (KM) systems, which can enhance the access to knowledge
and expertise of engineers, as previously postulated. Such systems assist them in
sharing tacit knowledge, developing concepts from many pieces of tacit knowledge,
combining various elements of explicit knowledge, and storing new knowledge in the
knowledge bases of individuals and the organization. In such a KM system, an engineer
is assigned a personal blog where he/she can store work in progress, including collec-
tions of random thoughts, ideas, documents, spreadsheets, or audio-video recordings.
Through the dynamic social networking facilities of blogs, exchanges of comments and
discussions among colleagues are electronically facilitated and individual blog entries
and discussions can be cross-referenced. After a work-in-progress is completed, related
blog entries and discussions, as well as a project summary document, can be archived
to a digital library for future reference.
This chapter builds on the general design for structural engineering support
systems described in Chapter 3 to propose a methodology by which blogs and digital
libraries can be combined into a structural engineering support system based on three
enabling modules which will be described and specified. A framework for this applica-
tion is proposed and the prototype that facilitates the management and sharing of both
personal knowledge and computing tools is developed, with the objective of improving
the overall efficiency and productivity of engineering design processes.1

4.2 Blog+DL Framework for Engineering Knowledge Management
Based on the SECI model, this section identifies the features of a KM system
that promote the conversion and interaction of tacit and explicit knowledge among
individuals. A framework for the use of Weblogs and Digital Libraries as building
blocks for a KM system is presented.

4.2.1 Technology Application Framework
Socialization and Externalization
As depicted in the top left-hand quadrant of Figure 4.1, the first part of the
spiral of knowledge creation is interpersonal. It starts when people that share the same
interest gather to interact, socialize and externalize their tacit knowledge by sharing
experiences. A good KM system should help such people to discover each other and
provide them a convenient means for socialization and externalization.
The statistics about blogs in Section 2.3 indicate that blogs can be used as an
effective KM tool in this regard. Blog publication and the social networking features
help people to discover each other and create online social networks of bloggers on which
issues of common interest are expressed and personal opinions, as well as experiences,
are exchanged. When Blogs is used as an engineering KM tool, an engineer may blog
1
The content of this chapter has been published in Vacharasintopchai et al. (2007b).

22

Figure 4.1: Blog, Digital Library, and Knowledge Conversion Modes
(adapted from Nonaka and Takeuchi 1995)

about a design problem that he is trying to solve, e.g., how to setup a proper design
criterion for a construction project. The blog entry may describe the engineer’s view of
the problem as well as the available approaches, and may also contain initial research
work, such as hyperlinks to other blog entries about “lessons learned” from similar
projects. When published to the Internet or a corporate intranet, the blog entry can
be read by many other people, some of which may share the same interest or have
experienced the same problem. A reader may assist the engineer in finding a solution
by posting his opinion or experience, such as the limitation of a selected approach,
as a comment to the blog entry. Another reader may view the problem differently
and may not agree with the first reader’s opinion. He may express his opinion on a
personal blog and notify the original blogger in the form of a trackback. The original
blogger can participate in the discussion thread created by this series of comments and
trackbacks and be able to find a more appropriate design criterion for his project. This
blog discussion thread may also benefit a public audience who may have had a similar
problem and comes across this discussion thread on a Web search engine. In this way,
a group of people who share the same interest is gathered; a field of interaction is
constituted; and the participants are equipped with a tool to socialize and exchange
experiences, i.e., pieces of tacit knowledge, in blog discussion threads.
Combination and Internalization
The second part of the spiral of knowledge creation is intrapersonal. It contin-
ues from the interpersonal part when a person synthesizes new explicit knowledge—
combining different bodies of explicit knowledge to solve an unfamiliar problem. Once

23

the new knowledge is put into practice, it becomes valuable know-how that is internal-
ized into the person’s tacit knowledge base and becomes part of his skills. An effective
KM system should help people to easily build up repositories of explicit knowledge and
allow relevant pieces of knowledge in the repositories to be conveniently and precisely
retrieved so that a more complete set of knowledge is accessible and can be combined
into the best possible new knowledge. A desirable KM system should also facilitate
internalization, i.e., help people put new knowledge into practice by, for example, pro-
viding them convenient access to relevant computational software tools to shorten the
time required for a design workflow.
Keyword tagging and the selective tag browsing features of Blogs complement
the metadata rich content archiving and retrieval features of Digital Libraries and
make them suitable as effective KM tools in this regard. Blog entries created during
socialization and externalization are typically tagged with keywords when they are
published. In addition to conventional keyword and full text searches, the blog owner
and readers can use the selective tag browsing feature available in most blog service
providers to filter and browse only through relevant blog entries, allowing them to
stay focused on the topics of most interest to them. An engineer can also archive
the content and metadata of mature blog discussion threads, which contain pieces
of externalized tacit knowledge and reflects the group thinking processes by which
solutions to important problems are arrived at, into the collections of the institutional-
repository digital library of which he has a membership. He can also archive interesting
Web pages, such as product data sheets, and digital content, such as Excel, PowerPoint,
PDF, Word, and MP3 files, into the collections. In this way, the engineer builds
up a personal portfolio of knowledge that can be shared across the organization. A
digital library, with its extensive metadata browsing and search features, as well as
full-text indexing capability, becomes a human-filtered search engine (Takeda, personal
communication, February 26, 2007) which portfolio owners and peers can use to access
a focused set of relevant knowledge without “noise.” With all the mentioned features,
Blogs and Digital Libraries can thereby enhance the efficiency and effectiveness of the
combination and internalization processes of individuals.

4.2.2 System Architecture
A system architecture that corresponds to the proposed technology application
framework is presented in Figure 4.2. Seven components, namely, the Web Browser,
the Blog Server, the Digital Library Server, the World Wide Web, the Archiver, the
Keyword Suggester, and the Web Service (WS) Portfolio Manager, are involved. The
latter three are the key components that, together with Weblog and Digital Library
technologies, enable the constitution of an effective knowledge management system.
The Web Browser is proposed as the primary user interface for convenient and
ubiquitous access to the system and also because Blogs and Digital Libraries are them-
selves Web-based. An engineer can use a Web browser to create an account on a blog
server and a digital library, and use it to log on to the blog server to socialize and
externalize his knowledge by publishing blog entries or joining blog discussion threads.
The Archiver is a special-purpose component that monitors requests from a
Web browser to take a “snapshot” of a blog discussion thread or a Web page and
archive it to a specific collection in a digital library. It is the primary component that
assists in building up personal knowledge portfolios. When a blog discussion thread has
matured, the blog owner may summarize the discussion and use a Web browser to take

24

Figure 4.2: Blog+DL Framework for Computer-Aided Engineering Design

25

a snapshot of the thread. The Web browser would prompt him to enter appropriate
metadata, such as the title, the author’s name, keywords, and a brief description of the
thread, as well as his digital library account name and the name of the collection, i.e.,
knowledge portfolio, to which he is authorized to submit. It would then submit the
collected information along with the URL of the thread to the archiver and request that
a snapshot of the thread be taken and processed accordingly. In addition, an engineer
may use the Web browser and the archiver to take snapshots of interesting Web pages,
such as discussion threads, product datasheets, or design tips, into his portfolio. During
the combination and internalization processes, when explicit knowledge from various
sources is crystallized into new explicit knowledge and put into practice, the engineer
then uses the Web browser to access relevant discussion threads in blogs and pieces of
knowledge collected in personal portfolios in the digital library.
The Keyword Suggester is a second essential component in the KM system
proposed here. Since the Blog and Digital Library are independent technologies that
do not naturally interoperate; the use of common metadata is necessary to enable
such interoperation, i.e., the keywords assigned to blog entries and digital library items
should be from the same controlled set and the bibliographic information assigned to
them should conform to the same data format (e.g., DD-MM-YYYY for dates) so that
relevant knowledge in blogs and the digital library can be retrieved simultaneously with
a common search query. The Keyword Suggester is the component that ensures the
enforcement of common metadata.
Besides exchanging and commenting on traditional documents and multimedia
content, an engineer can participate in the SECI modes of knowledge conversion by
exchanging and commenting on computational tools, such as mathematical routines,
numerical analysis modules, and design shortcuts. An engineer can develop these
computational tools as subroutines (or methods, in the object-oriented programming
paradigm) in his preferred programming languages, such as Java or VB.NET, and al-
low them to be remotely executed by other computers as Web services on the Internet.
To “publish” a Web service, i.e., to make a subroutine available for remote access, an
engineer places the code of the subroutine onto a Web service server, such as Apache
Axis for Java or Microsoft Internet Information Server for VB.NET. The server au-
tomatically generates a “Web service description document” that corresponds to the
subroutine. This document contains the instruction for computers and people on how
to execute the service. The engineer can post the Web service description document in
a blog entry, along with a brief description of the service, its theoretical explanation,
and a user’s guide. In this way, the service is advertised and made accessible to the
public. Discussions about the service, such as user experiences, can be exchanged as
blog comments and trackbacks as in the ordinary socialization and externalization use
of blogs. The Web services published through blog entries are discovered primarily
when the service advertisement entries are visited by fellow bloggers. They may also
be discovered secondarily when referred to in other discussion threads. Besides reading
the contents, e.g., descriptions and explanations, blog readers can execute the service
or build up a portfolio of computational tools, i.e., Web services.
This availability of the portfolio of Web services necessitates the third compo-
nent being suggested in the framework, namely, the Web Service Portfolio Manager.
This third key component turns the proposed framework into a fully-functioning KM
system by enabling the sharing and execution of Web services published in blog discus-
sion threads. This add-on component to the blog server detects whether the content

26

of a blog entry includes a hyperlink to the description document of a Web service. If
it does so, the Web Service Portfolio Manager will provide an “execute” button and a
“save” button in the Web page that displays the blog entry, so that the readers can
choose to execute the service or save it to their portfolio for later use. In the latter
case, the user will be asked to provide the description of the desired Web service when
a computational service is needed. The Web service portfolio manager would search
the user’s portfolio and retrieve a list of Web services that best match the request.
Once a Web service is selected by the user, the user can execute the Web service by
entering the input data and clicking the execute button. The selection of Web services
from portfolios and the execution of Web services are performed intelligently according
to the Semantic Web Services Framework for Computational Mechanics (SWSCM),
which is proposed and explained in the next chapter. The Web service portfolio man-
ager performs the consumer and the matchmaker roles in the SWSCM. Readers are
referred to the mentioned reference for in-depth information about the Semantic Web
Services technology and its application in engineering design.

4.3 A Semantic Computing-Based Support System for Structural Engi-
neers
4.3.1 Implementation
Based on the proposed Blog+DL framework for computer-aided engineering
design, a prototype system has been developed to assist engineers in socializing and
externalizing their knowledge, as well as in building up personal portfolios of knowl-
edge and computational tools. The knowledge and tools in such portfolios can be
combined and crystallized into new knowledge and can also be shared with colleagues
of the engineers to form a larger organizational KM system. The prototype system
contains structural design aids, such as properties of steel sections from various in-
dustrial standards, as well as Web services and worksheets for structural analyses and
designs. It also includes articles and discussion threads about general sciences and
applications of information and communication technologies in structural engineering.
The implementation of the prototype system follows the system architecture presented
previously which consists of six components, namely the Web Browser, the Blog server,
the Digital Library server, and the three KM enabling components: the Archiver, the
Keyword Suggester, and the Web Service (WS) Portfolio Manager. A federated search
agent was additionally introduced for even improved productivity.
The prototype system was implemented in Java and JavaScript and utilized
platform neutral open-source software tools. The server components were deployed on
a Windows XP Professional Web server at http://kb.blogdns.org:8080/blojsom.

Web Browser Mozilla Firefox (Mozilla, 2007), a standards conformant Web browser
available on major computer platforms, such as Windows, Linux and Mac OS X,
was chosen as the Web Browser component. Firefox is highly extensible and fully
supports bookmarklets—JavaScript codes embedded in Web browser bookmarks—
which are useful as custom user interface components.

Blog Server Blojsom (Czarnecki, 2007), a blog server developed in Java with ex-
tensive support for third-party add-on modules, was chosen as the Blog server.
Optional features useful for KM, such as site-wide article aggregation, selective
tag browsing and keyword tagging, as well as full-text searches, were installed

27

and customized.

Digital Library Server DSpace (DSpace, 2007a), a widely accepted large-scale Java
digital library server based on the institutional repository concept, with an exten-
sive documented set of application programming interfaces (APIs), was chosen
as the Digital Library server. Several collections, such as general “Structural
Engineering,” “Steel Design,” “Reference,” “News,” and “Tips and Techniques”
were created to store user contributed contents.

Archiver The Archiver was developed as a Java servlet that monitors requests to
archive documents at particular URLs along with the provided sets of metadata
into specific DSpace collections. An archive bookmarklet was developed as a
bridging user interface between Mozilla Firefox and the Archiver (Figures 4.7a, b).
The bookmarklet was deployed as a Save to DSpace button in the Mozilla Firefox
bookmarks toolbar. Once clicked it prompts the user to enter a brief descrip-
tion and the metadata of a Web page and forwards the request to the Archiver.
The Archiver utilizes DSpace APIs, such as the methods of classes Item and
Collection in the org.dspace.content package, and the JavaMail APIs (Sun
Microsystems, Inc., 2007) to create and store standard compliant MHTML doc-
ument archives (Palme et al., 1999) into the DSpace repository. By adopting the
MHTML Internet standard, document archives generated by the Archiver can be
readily processed by the built-in full-text indexer of DSpace and are displayable
on major Web browsers, such as Microsoft Internet Explorer, Mozilla Firefox and
Opera.

Keyword Suggester The Keyword Suggester was implemented as a servlet that pro-
vides ontology-assisted keyword suggestions. It takes input strings from HTTP
POST requests and consults a predefined set of ontologies to extract relevant key-
words from Web contents at specific URLs by using APIs from Jena (Jena, 2007),
an open-source Semantic Web programming toolkit originated from Hewlett-
Packard Laboratories, and Apache Lucene (Apache, 2007), a high-performance
text search engine toolkit from Apache Software Foundation. The bookmarklet
uses Asynchronous JavaScript and XML (AJAX) programming techniques to
consult the Keyword Suggester and retrieve a list of suggested keywords for the
Web page being saved. JavaScript codes in the bookmarklet prepopulate these
keywords into the keywords textbox to assist users in entering proper metadata.
Ontology-assisted keyword suggestion is helpful to enforce a controlled set of key-
words, which is necessary for the interoperation of Blog and Digital Library and
the effective retrieval of content in large digital library collections. The same func-
tionality is also provided in the blog entry editing module of Blojsom to suggest
proper keyword tags when entries are created or modified (Figure 4.4b). Read-
ers are referred to Jakobsen (2006) for an in-depth explanation on bookmarklet
programming.

Federated Search Agent A federated search function which allows users to locate
pieces of knowledge across blog entries and digital library collections is provided
for increased productivity. Like the Archiver, the federated search agent was
developed as a Java servlet that monitors requests to search for keywords or
metadata. A search bookmarklet was developed as a bridging user interface
between Mozilla Firefox and the Federated Search agent (Figure 4.8c). The

28

bookmarklet was deployed as a Search button in the Mozilla Firefox bookmarks
toolbar. Once clicked it prompts users to enter search keywords or metadata
and forwards requests to the Federated Search agent. The agent utilizes APIs
such as methods of classes SimpleSearchServlet and AdvancedSearchServlet
in the org.dspace.app.webui.servlet package of DSpace, methods of class
BlojsomFetcher in the org.blojsom.fetcher package of Blojsom and full-text
search functions of Apache Lucene, the underlying search engine used by Blojsom
and DSpace, to locate relevant pieces of knowledge and consolidate them into a
single Web page for convenient access by users (Figure 4.8d).

Web Service Portfolio Manager The Web service portfolio manager consists of
two important subcomponents: the Blojsom plug-in and the Web service port-
folio management servlet. The Blojsom plug-in implements the BlojsomPlugin
interface class in the org.blojsom.plugin package of Blojsom to preprocess the
contents of blog entries before they are displayed on Web browsers. By using
the regular expression string matching technique, the plug-in detects whether
blog entries contain links to Web service descriptions, such as WSDL or OWL-S
files. If they do so, (1) Save buttons are inserted into the HTML codes of blog
entries such that, once clicked, users’ log-in names and URLs of Web service de-
scriptions are forwarded to the Web service portfolio management servlet, which
maintains the registration of Web services in users’ portfolios; and (2) user inter-
face components such as input boxes and submit buttons are inserted into the
HTML codes of blog entries for immediate execution of Web services, with the
assistance of the OWL-S API (Sirin, 2004), WSDL4J (WSDL4J, 2006) toolkit
and Apache Web Service Invocation Framework (WSIF) (Apache, 2003b). The
Web service portfolio management servlet handles the registration, unregistration
and selection of services in users’ portfolios. It was developed as a Java servlet
that monitors requests to add, update or delete services in particular portfolios,
identified by log-in names. It also acts as a proxy to the Matchmaker service
described in Section 5.3 on page 51, which implements the matchmaking algo-
rithm in Section 5.2 to help users select appropriate services from many they have
collected. Java hash tables and property files were used as building blocks for
service registration databases. OWL-S API and WSDL4J toolkits were used to
extract information from Web service descriptions. WSIF was used for the execu-
tion of Web services. Jena Semantic Web toolkit and RACER Description Logic
reasoner (Haarslev and Moller, 2003) were used to provide intelligence assistance
to the matchmaker. Intelligent interoperability of Web services are explained in
detail in the next chapter.

4.3.2 A Proof of Concept Prototype
Figures 4.3 through 4.8 illustrate a typical application of the prototype, a walk-
through example of which is demonstrated as follows:
Installation
Initially, an engineer needs to create an account on the system by visiting the
homepage of the Blog and the Digital Library server. An instruction on the homepage
will direct him to the services page where his user name is verified (Figure 4.3a), URL to
personal blog assigned, membership of DSpace collections granted, and Save to DSpace

29

(a) Account Verification (b) Personalized Bookmarklets

Figure 4.3: User-specific Initialization of Prototype

and federated Search bookmarklets personalized (Figure 4.3b). Bookmarklets must be
dragged and dropped onto the bookmark toolbar of Mozilla Firefox before they can be
used.
Socialization and Externalization
In terms of publication, to externalize his knowledge and socialize, the engineer
logs on to his blog and clicks Manage to enter Blojsom administration console. To
create a new blog entry, he chooses the Add Entry command in the Entry toolbar which
directs him to the Add Blog Entry page. Here, he can start blogging about new findings,
as well as ideas and plans related to a project, in a fashion similar to composing e-
mails (Figure 4.4a). The content of his entry is automatically analyzed for important
keywords as he blogs, thanks to embedded AJAX codes, preregistered ontologies and
the Keyword Suggester. Checkboxes of suggested keywords are presented in real time
in the Smart Keyword Suggestions section where he can pick ones that apply. He may
also add invented keywords into the User’s specified textbox, although such a practice
is discouraged for vocabulary control reasons (Figure 4.4b). The composition of blog
entry completes when the Add Blog Entry button is clicked. New entries are saved into
the server’s database.
Entries by individual users are listed in a reverse chronological order on their per-
sonal blogs, accessible at the URLs http://http://kb.blogdns.org:8080/blojsom/
blog/[user_name]. Tag clouds, lists of tag names with varying font sizes, which rep-
resent different frequencies of keywords used by particular users, are also provided so
that visitors can easily navigate through various aspects of the authors’ interests (Fig-
ure 4.4c). Recent entries by all users are additionally aggregated on the prototype’s
homepage at the URL http://http://kb.blogdns.org:8080/blojsom (Figure 4.4d)
for quick overview of new contributions from all users.
In terms of consumption, the engineer can browse through blog entries by visit-
ing the prototype’s homepage as well as personal ones of acquainted colleagues. He may
check for updates by paying regular visits to respective sites. Such a practice is consid-
ered the “pull” method of blog consumption. He may also choose the “push” method
to subscribe to blogs, as well as aggregated entries on the homepage, by clicking at the
orange Web feed icon at the right end of Web browser address bars, cf. Figures 4.4c, d.
By doing so, he is next presented with a Web feed subscription screen (Figure 4.5a)
where he can allow updated contents to be regularly pushed to him via a blog reader.
Figure 4.5b illustrates how a personal blog is subscribed as a Web feed source in Mi-
crosoft Outlook 2007. Excerpts from recent entries on the blog are regularly pushed

30

into the engineer’s Inbox, along with other e-mail items (Figure 4.5c). Interesting blog
entries can be opened and read in a fashion similar to that for e-mails (Figure 4.5d).
The engineer can click the View article button to visit original blog entry pages to read
full articles and participate in discussion threads with his peers (Figure 4.6a).
Combination and Internalization
During day-to-day tasks, engineers often come across new knowledge that are
sometimes not related to present tasks but may be useful in the future. The engineer
can log on to his DSpace portfolios (collections) and collect pieces of knowledge, as
well as their metadata, into the knowledge repository. For improved productivity, Web
contents such as blog discussion threads, datasheets and design aids can be directly
archived into DSpace collections by clicking at the Save to DSpace button on Mozilla
Firefox’s bookmark toolbar. The engineer will be prompted to enter bibliographic
metadata, i.e., title, authors, keywords and brief descriptions, and to select a DSpace
collection most appropriate for the new knowledge piece (Figures 4.7a, b). Non-Web
contents such as printed documents can be digitized into electronic forms, such as PDF
and JPEG files, and are archived into the repository through DSpace’s standard Web
interface (Figure 4.7c). Figure 4.7d illustrates a personal knowledge portfolio built
upon the accumulation of knowledge pieces into DSpace collections.
Knowledge shared in blog entries as well as those accumulated in personal port-
folios are invaluable resources to promptly solve the engineer’s problems at hand. He
can also use these knowledge to help others by providing suggestions and URLs to
knowledge pieces in colleagues’ discussion threads. When working on a project simi-
lar to some in the past, archived documents in the repository can be readily accessed
thanks to DSpace’s metadata and full-text search facilities (Figure 4.8a). If he feels
he had seen design tips related to this type of projects in a colleague’s Web feed, he
can use Blojsom’s keyword and full-text search features to locate that particular blog
entry (Figure 4.8b). For improved convenience, the engineer can also use the fed-
erated Search bookmarklet to simultaneously search through discussion threads and
knowledge portfolios contributed by him and colleagues (Figure 4.8c). Search results
are summarized in single Web pages where articles are readily accessed by one mouse
clicks (Figure 4.8d). Web services published through blog entries can also be executed
right away from results pages with the assistance of the Web Service Portfolio Manager
described earlier (Figure 4.8e).
In this way, the engineer and fellow staff members are equipped with an inte-
grated set of tools to better share knowledge and experience.

4.4 Summary
This chapter has proposed a framework for the application of Weblog, Digital
Library, and Semantic Web Services technologies to support structural engineers dur-
ing the engineering process. The framework is based on the theory of organizational
knowledge management which postulates that knowledge is created by the interactions
between tacit knowledge and explicit knowledge among individuals in an organization.
In contrast to the KM initiatives in the late 1990s, the focus of this framework is on
facilitating the conversion, interaction, and sharing of tacit and explicit knowledge,
which are the bases for creating knowledge, rather than on capturing the individual
pieces of explicit knowledge. The proposed framework also differs from many modern
KM tools for engineering design firms in the literature because it is based on several

31

Internet standards beyond HTTP and HTML and therefore promotes cross-system in-
teroperability and advanced applications of such Web technologies as Web feeds and
Web Services, which are enabling technologies for Web 2.0 (a.k.a. the social Web) and
Internet-based multiplatform distributed computing. The introduction of Semantic
Web Services technologies and the application of ontology, as a fundamental compo-
nent of the Semantic Web, further open KM systems that conform to the proposed
framework to a wide variety of Internet-based artiﬁcial intelligence technologies. The
proposed framework can improve the knowledge and productivity of engineers in to-
day’s ever more globalized and competitive world.

32

(a) Creating New Blog Entry

(b) Ontology-assisted Keyword Suggestion

Figure 4.4: Blog Publication Features of Prototype

33

(c) Tag Cloud (on right)

(d) Recent Blog Entries

Figure 4.4: Blog Publication Features of Prototype (cont’d)

34

(a) Subscribing to Web Feed

(b) Subscribed Web Feeds in Microsoft Outlook

Figure 4.5: Blog Consumption Features of Prototype

35

(c) Using Microsoft Outlook as Web Feed Reader

(d) Reading Blog Entries by Microsoft Outlook

Figure 4.5: Blog Consumption Features of Prototype (cont’d)

36

(a) Discussion among Colleagues

Figure 4.6: Blog Socialization Features of Prototype

37

(a) Archiving Discussion Thread to DSpace

(b) Archiving Webpage to DSpace

Figure 4.7: Knowledge Archiving Features of Prototype

38

(c) Archiving Non-Web Content

Figure 4.7: Knowledge Archiving Features of Prototype (cont’d)

39

(d) Personal Portfolio of Knowledge

Figure 4.7: Knowledge Archiving Features of Prototype (cont’d)

40

(a) Digital Library Search

(b) Blog Search

Figure 4.8: Knowledge Retrieval Features of Prototype

41

(c) Federated Search

(d) Search Results

Figure 4.8: Knowledge Retrieval Features of Prototype (cont’d)

42

(e) Web Service Execution

Figure 4.8: Knowledge Retrieval Features of Prototype (cont’d)

43

CHAPTER 5
ENGINEERING SOFTWARE INTEROPERABILITY

5.1 Introduction
Computers performing structural analysis using numerical methods require close
guidance from human users. Consider the case of finite element analysis, in which it is
first necessary to examine a real world problem and construct a computer model of the
problem including the problem domain, material behaviors, and boundary conditions.
It may also be necessary to consult many design codes or previously obtained experi-
mental results to assemble the proper input data for the analysis. Next, the problem
domain must be discretized, either manually or through the help of automatic meshing
software, before the element stiffness matrices can be computed and assembled, the
boundary conditions applied and a solution for nodal unknowns found. Experience
and judgment are required by the analyst to select the element types and the consti-
tutive models for the materials and physical processes being simulated, to examine the
accuracy of the analysis results and, if necessary, to repeat the entire process to study
solution convergence. For large and complex analyses, running an entire process may
take hours or days and visualization techniques, including animations or virtual real-
ity, may be needed to effectively interpret the analysis results. Such situations may be
even more complicated when incompatibilities between input and output data formats
of the software components employed exist, and are likely to be tedious and lead to
inaccurate solutions, especially if performed by inexperienced users.
Web Services (W3C, 2004d) and the Semantic Web (Berners-Lee et al., 2001),
two Internet technologies, have the potential to help unify and utilize scattered re-
sources, such as personal computers, computer clusters, supercomputers, databases
and knowledge bases to enable numerical structural or mechanical analyses to be done
in a fast, accurate and automated manner. These technologies potentially reduce the
need for analysts to be actively involved in providing all the required data and detailed
instructions. Rather, analysts can initiate the solution process by giving simple in-
structions, monitoring the computers as they work on obtaining solutions, and making
a final decision on the acceptability of the results. With today’s ubiquitous access, the
Internet is not simply a hyper-library for information searches, nor only a communica-
tion tool for sending e-mail messages. It may, in fact, be utilized as a large platform
for scientific computation with the help of intelligent software agents that collaborate
to accomplish a given task.
A framework for the joint application of Web Services, the Semantic Web and
Semantic Web Services to address software interoperability problems in structural en-
gineering is proposed in this chapter, a proof of concept prototype presented and its
potential illustrated by a case study which involves the solution of a typical problem
in structural analysis and design.1

5.2 A Semantic Web Services Framework for Computational Mechanics
As discussed earlier, Semantic Web Services have the potential to help unify
computing resources and knowledge scattered on the Internet into a large platform for
scientific computing. In this section, an architectural framework is proposed, from a
system development point of view, for the development of computational mechanics
1
The content of this chapter has been published in Vacharasintopchai et al. (2007a).

44

Service
Description
Matchmaker
APP LOGIC, OWL-S
KB, DB

Pu
d

bli
Fi n

sh
Service Service
Consumer Provider
Bind
APP LOGIC, APP LOGIC,
KB, DB SOAP KB, DB
(XML+RDF)

Figure 5.1: Semantic Web Services Model

software based on Semantic Web Services. To prevent ambiguity with terms in the
general Semantic Web Services architecture, the proposed framework is called Semantic
Web Services for Computational Mechanics (SWSCM), and a semantic Web service in
the SWSCM architectural framework is called an SWSCM Service.

5.2.1 SWSCM Model
The proposed SWSCM is an extended version of the standard Web Services
model presented in Figure 2.1 on page 10. The model of the SWSCM architecture is
illustrated in Figure 5.1. The goal in the development of the SWSCM framework is to
make it as compatible with the prevalent Web Services technology as possible, i.e., we
try to make SWSCM the semantic “shell” of Web Services. By adopting such a shell
principle, we will have access not only to a variety of software development tools for
Web Services technology, which are available in several programming languages, but
also the many Web services deployed on the Internet.
To facilitate the publication and discovery of relevant semantic Web services, the
UDDI service registry is replaced by the matchmaker service, and the WSDL service
description is replaced by the OWL-S semantic Web service description (Figures 2.1
and 5.1). Access to WSDL documents on the Internet is still necessary for service
consumers to properly ground and bind to service providers. However, the matchmaker
does not need to store the copies (or URL to the copies) of such documents locally
because only the semantic Web services description from OWL-S documents is used in
the matching process.
SOAP messages, which are used for communication between Web services, are
augmented with RDF statements so that XML data transmitted from a sender will
be meaningful to the recipient. Each parameter of the request and response mes-
sages between service consumers and providers consists of the serialized XML data
(as required by the SOAP speciﬁcation) and its corresponding RDF statements which
explain the meaning of such data, as shown in Figure 5.2. RDF statements are typi-
cally encoded in XML format and the format is termed RDF/XML according to the
RDF speciﬁcation. As RDF/XML can naturally be included as extra attachments in
SOAP messages, RDF-augmented SOAP messages will not cause a compatibility issue

45

SOAP Envelope

SOAP Body

REQUEST / RESPONSE
MESSAGE

CONSUMER

PROVIDER
Parameter 1
XML
RDF
DATA

...
Parameter n
XML
RDF
DATA

Figure 5.2: Semantic SOAP Message

KNOWLEDGE BASE APP LOGIC DATABASE
Concepts Objects Table A

Processor
Inference
Engine

Query
Assertions - fields Table B
- methods ...
Instances Table n

Figure 5.3: Components of Semantic Web Service

with non-SWSCM aware recipients because the recipients can ignore the part of an
attachment that they do not recognize.

5.2.2 Components of SWSCM Service
Each SWSCM service comprises three components, which are the application
logic component, the knowledge base component, and the optional database component,
as illustrated in Figure 5.3.
The application logic is the primary component of a SWSCM service. It per-
forms the functions advertised in its service description. It handles and processes
requests initiated by service consumers, and delegates tasks to the others by sending
out request messages. The application logic is the same component found in typical
services on the Web Services architecture.
The knowledge base provides intelligent assistance to the application logic. It
consists of an inference engine and copies of ontologies and RDF (or OWL) state-
ments downloaded from the Semantic Web. Multiple domain ontologies, e.g., a units-
of-measurement ontology, a material property ontology, and a structural engineering
ontology, are downloaded and stored locally on the knowledge base during the initial-
ization process of an SWSCM service. The knowledge base uses its inference engine
and the premises from RDF (or OWL) statements to answer queries initiated from the
application logic.
The database component provides information processing assistance to the ap-
plication logic. It is an optional component which contains a dedicated database man-
agement system. The database component is useful when an SWSCM service has to
deal with a large amount of raw data and complicated database queries, as in the

46

case of postprocessing the results from a large-scale finite-element analysis which can
involve several million degrees of freedom.

5.2.3 Communication between SWSCM Services
In standard Web Services architecture, to send a request to a service provider,
a service consumer first serializes its respective input parameters—“objects” in the
object-oriented programming paradigm or “structures” in the procedural programming
paradigm—into a structured XML representation (shown as XML data in Figure 5.2).
According to the WSDL standard by W3C, the structure of the XML representation
(termed XML schema) is agreed upon in the “Types” section of the WSDL document
published by the service provider. Next, the service consumer consecutively wraps
the input parameters with the SOAP Body and SOAP Envelope XML elements, and
finally uses the HyperText Transfer Protocol (HTTP) to transmit the SOAP envelope
to the service provider at the “rendezvous URL,” which is also specified in the WSDL
document. Once the service provider receives the SOAP message from the consumer,
it does the reverse procedure to extract the input parameters and deserialize them into
objects or structures, depending on its programming paradigm. Later, it processes the
input data, serializes the output into XML format, constructs the SOAP message, and
finally transmits the result back to the service consumer at the rendezvous.
SWSCM services communicate in the same fashion as standard Web services do,
i.e., through SOAP messages. The WSDL grounding section of an OWL-S description
grounds the generic input parameters, output parameters, and atomic processes to
specific sections of a WSDL document that specify the XML schema of the input
parameters and output parameters, and the rendezvous URL for actual invocation of
the Web service. At the syntactic level, SWSCM services refer to the XML schema
and rendezvous URL for the serialization of objects or structures and the transmission
of SOAP envelopes, respectively.
In addition, at the semantic level, the communication between SWSCM ser-
vices extends the standard SOAP messaging protocol by embedding the RDF state-
ments relevant to the objects or structures carried by SOAP messages (Figure 5.2).
It is proposed that the application logic component of each SWSCM service assigns
each object or structure a unique URI at runtime and that it populates its associated
knowledge base component with a set of assertions about the object or structure. For
example, a floating-point object Width 01 with the value of 2.00, which represents
the inch width of a rectangle, could be given the URI of http://ws.example.com/
object_ids#Width_01. If we refer to this URI as URI W01, its corresponding set of
assertions would be the RDF statements:
URI W01 “is an instance of” http://std.swscm.org/geometry.owl#Width; and
URI W01 “has unit” http://std.swscm.org/units.owl#Inch.
A pair of an object (or a structure) in the application logic and its assertions in the
knowledge base is termed a Semantic Object.
During the serialization process, when an SWSCM service consumer sends a
request message to the provider, the application logic component not only serializes an
input parameter (an object or a structure) into the XML Data, but also provides an
accompanying set of RDF/XML assertions (shown as the RDF block in Figure 5.2).
Having received and deserialized the SOAP message, the SWSCM service provider will
be able to not only reconstruct the data set of input parameters but also replicate the

47

assertions about each input parameter into its own knowledge base. In addition to
the semantic matching of Web service description profiles, which is discussed in the
section on “Semantic Web Services”, the proposed RDF-augmented SOAP messages
provides an extra cross-checking means to address the incompatibility issue between the
input and output parameters. RDF-augmented SOAP messages would help prevent
the kind of problem that causes mishandlings of physical quantities in the mission-
critical calculations, which occurs, for example, just because one service expects the
input parameters in metric units of measurement and another service provides the
input parameters in English units, and the two services by chance refer to their input
parameters by the same ontological concept. The RDF augmented SOAP messaging
in the SWSCM framework is termed the Semantic SOAP.

5.2.4 Inference Engine and Reasoning Process
The proposed SWSCM framework is independent of an inference engine. Any
inference engine can be used provided that they are equipped with the basic rules
for processing RDF and OWL statements. As discussed in the sections on “Semantic
Web” and “Semantic Web Services”, reasoning on RDF and OWL statements relies
heavily on concept subsumption, which is the determination of whether one concept is
more general than another. For instance, term X subsumes term Y if term X is more
general than term Y, and term X is the same as term Y if term X and term Y subsume
each other. Therefore, the inference engine for the knowledge base component of an
SWSCM service must satisfy the following requirements:

1. It must be capable of parsing the statements encoded in RDF and OWL lan-
guages;

2. It must natively understand or be equipped with the set of rules to understand
the semantics of RDF and OWL constructs as specified by W3C (W3C, 2004c,b);
and

3. It must at least be able to answer concept subsumption problems.

An inference engine that understands the semantics of RDF and OWL will be able to
answer queries similar to the following examples, with terms in the quotation marks
referring to standard constructs in the RDF and OWL specifications.
Example 1
Given the premise RDF statements:

term X is an “instance of” class X;

class X is a “subclass of” class Y; and

class Y is a “subclass of” class Z,

then

Query: Does class Z subsume class X? Answer : Yes, by transitivity.

Query: Is term X an “instance of” class Z? Answer : Yes, because term X is an
instance of class X which is subsumed by class Z.

48

Example 2
Given the premise RDF statements:

Triangle is a “subclass of” PlaneShape;

Triangle is a “subclass of” ClosedShape;

Triangle “has side” of Line only; and

The “number of” “has side” properties that Triangle has is 3,

then, if an object is defined by the following assertions:

The object is a “subclass of” PlaneShape;

The object is a “subclass of” ClosedShape;

The object is a “subclass of” EquilateralShape;

The object “has side” of Line A, which is an “instance of” Line;

The object “has side” of Line B, which is an “instance of” Line;

The object “has side” of Line C, which is an “instance of” Line; and

The “number of” “has side” properties that this object has is 3.

Query: Is the object a Triangle? Answer : Yes, because based on the given
premises, Triangle subsumes the definition of the object; therefore, the object is a
kind of Triangle.

Examples of inference engines that can be used in SWSCM knowledge base compo-
nents are RACER (Haarslev and Moller, 2003) and the SWI-Prolog Semantic Web
package (Wielemaker et al., 2003), which are based on Description Logics (Nardi et al.,
2003) and Prolog, respectively. RACER is an example of an inference engine that na-
tively understands the semantics of RDF and OWL constructs because its underlying
Description Logics framework was used as the basis to design the OWL language. SWI-
Prolog with its Semantic Web package is an example of an inference engine equipped
with a set of rules that enables it to understand the semantics of the RDF and OWL
constructs.
It should be noted that RDF and OWL are knowledge “representation” lan-
guages. The RDF and OWL specifications do not mandate the use of Description
Logics as the only logical framework for the Semantic Web. The knowledge inferred by
an inference engine, or the answer that it provides, is limited by its underlying logical
framework.

49

5.2.5 Matchmaking Algorithm
One of the most important operations in the Semantic Web Services architecture
is the matching of the ideal service profile of a service consumer against the service pro-
files registered by several service providers. In this section, the service profile matching
algorithm proposed for use by matchmakers in the SWSCM architecture is discussed.
The algorithm is inspired by the one proposed by Semantic Web Services researchers
at Carnegie Mellon University (Paolucci et al., 2002; Sycara et al., 2003).
An OWL-S profile description is a set of OWL-S statements that semantically
describes a service, which is either needed by a service consumer or offered from a service
provider. From the section on “Service Profiles” in the OWL-S specification (OWL–S,
2003), the elements of a profile description that are relevant to the interoperation of
SWSCM services are the taxonomic type of profile (i.e., whether a service belongs
to a certain class) and the “has input,” “has effect,” and “has output” properties.
Examples of such elements are the <profileHierarchy:SineFunctionEvaluator>,
the <profile:hasInput>, the <profile:hasEffect>, and the <profile:hasOutput>
in Figure 2.4 on page 17, respectively.
For each pair of the service profiles, the degree of match is calculated by using
the weighted average of the matching scores between the pairs of the profile types, the
input parameters, the effects of service, and the output parameters. Mathematically,
the degree of match between a pair of service profiles is

i wi di p
DS = (5.1)
i wi

where DS = degree of match between two service profiles; and wi and di p = respective
weight and the matching scores between the profile types, the input parameters, the
effects of services, and the output parameters. By default, equal weights are assigned to
the matching scores in typical matchmaking operations. Service consumers may request
higher weights to certain pairs of the profile description if compatibility between those
pairs is more important.
For each pair of the ideal concept CR and the advertised concept CA , the match-
ing score between CR and CA with respect to CR , d(CR , CA ), is defined as

1.00 if CR is the same as CA ,


0.75 if C subsumes C ,
A R
d(CR , CA ) = (5.2)
0.50 if CR subsumes CA ,


0 otherwise.


If we collectively call the profile types, the input parameters, the effects of service, and
the output parameters “parameters,” the value of d(CR , CA ) = 1.00 signifies that the
ideal parameter perfectly matches the advertised parameter. The value of 0.75 signifies
that the advertised parameter is more general than the ideal parameter, and that the
advertised service is not specifically made for the consumer. The value of 0.50 signifies
that the ideal parameter is more general than the advertised parameter, and that the
advertised service may not completely fulfill the consumer’s request. The value of 0
signifies that the two parameters are incompatible and the advertised service is not
recommended for the consumer (Paolucci et al., 2002).
SWSCM services often take several input parameters. In this case the matching
score between the two sets of input parameters is the unweighted average of the match-

50

ing scores between each sequential pair of the input parameters. Mathematically, it is
defined as

jdj I
dI p = (5.3)
N

where dI p = matching score between two sets of input parameters; dj I = degree of
match between the jth pair of input parameters; and N is the total number of input
parameters. To eliminate the case of strong incompatibilities between certain pairs of
input parameters, dI p must be taken as zero if any of dj I equals zero.

5.3 Application to Structural Analysis and Design
5.3.1 Implementation
A prototype system was developed to determine the maximum stress in a rect-
angular solid object of various configurations. The object, made of a material selected
by the user, was to be loaded uniformly under a user-given load specification and
subjected to user-given support conditions. The dimensions of the object were to be
specified in different units of measurement. The specific objectives in developing the
prototype were:

1. to demonstrate the interoperability of heterogeneous software components, as
offered by ontologies, OWL statements, semantic Web service descriptions, and
the matchmaking concepts; and

2. to demonstrate the role of OWL statements, published on the Semantic Web, as
dynamic and external data sources for structural analysis software components.

The following open source and public-domain software tools were used in the
implementation:

1. Apache Axis Web Services toolkit (Apache, 2003a),

2. Apache Web Service Invocation Framework (WSIF) (Apache, 2003b),

3. JAMA matrix library (NIST, 1999),

4. Jena Semantic Web Framework (Jena, 2007),

5. University of Maryland’s OWL-S API project (OWL-S API) (Sirin, 2004),

6. Prot´g´ ontology development tool (Prot´g´, 2005), and
e e e e

7. RACER Description Logic reasoner (Haarslev and Moller, 2003).

All Java Web services utilized the APIs provided by Apache Axis and WSIF
to initiate and process SOAP requests and responses during the execution of Web
services. Custom serializer and deserializer modules were developed in Java to allow
Apache Axis to recognize and process Semantic SOAP messages. Matrix operations
in Java Web services employed the APIs by JAMA matrix library. Java software
development kit version 1.4.1 and Microsoft Visual Studio .NET 2003 were used as
the compilers for Java and C# codes. All SWSCM services employed the inference

51

ParEFG
WS
USER-INTERFACE
(PDA/Smartphone)
Mesher WS

Match
maker
Structural
Analysis WS FEM WS

Max Value
WS

Figure 5.4: System Configuration of Prototype

engine and RDF knowledge base query and management APIs provided by Jena. The
Matchmaker employed the inference engine and RDF knowledge base APIs provided
by RACER, a dedicated Description Logic reasoner, for more powerful reasoning on
complicated ontologies. The Matchmaker utilized OWL-S API to extract concepts
represented in OWL-S service profiles. SWSCM services used OWL-S API to extract
service grounding information for binding to service providers. Grounded services were
executed by means of the APIs in WSIF. The Prot´g´ ontology development tool by
e e
Stanford University was used in the development of ontologies.
The prototype system consisted of five SWSCM services, one Matchmaker ser-
vice and one User Interface, as shown in Figure 5.4. The SWSCM services consist
of the Structural Analysis (SA) service, the Mesher service, the Parallel Element-Free
Garlerkin Method (ParEFG) service, the Finite-Element Method (FEM) service and
the Max Value service. WSDL documents and OWL-S documents, as well as ontologies
and OWL statements, were published on a Web server to facilitate the operation of
SWSCM services and the Matchmaker. WSDL documents describe the syntactic oper-
ations of SWSCM services, e.g., the schemas of XML Data and the rendezvous URLs
where SOAP messages are expected. OWL-S counterparts provide semantic profile
descriptions with grounding references to specific parts of the WSDL documents. Ideal
service profiles desired by SWSCM consumers were published as nongrounded OWL-S
documents. Ontologies developed and published on the Web server include:

1. the Quantities ontology that defines physical quantities and various types of num-
bers;

2. the Material ontology that defines materials and their mechanical properties;

3. the Units ontology that defines relations between units of measurement (e.g., “an
angular degree is a subclass of plane angle, has conversion factor of π/180 with
radian angle as the reference unit of measurement”); and

4. the Structural Engineering ontology that defines general knowledge in structural
engineering (e.g. the pinned support condition).

52

The prototype system was deliberately designed to consist of modules devel-
oped on potentially incompatible platforms. As postulated in Section 5.2.1, OWL-S,
WSDL and SOAP are the common languages that SWSCM services “speak.” Ser-
vices on different computing platforms were developed in appropriate programming
languages such that they speak the same dialects and communicate by the methodol-
ogy in Section 5.2.3. Key characteristics of modules in the prototype are presented as
follows:

Java Web Services The Matchmaker, the SA service, and the Max Value service
were developed as Web services in Java. The Matchmaker employed the service
profile matching algorithm in Section 5.2.5 to handle requests for service recom-
mendation. The SA service assisted the User Interface in numerical analyses of
structures. The Max Value service employed the Quicksort algorithm to identify
maximum values from lists of double-precision numbers.

Commercial Software Package The Mesher and the FEM services were developed
in Java as Web service frontends to the ANSYS commercial FEM software (AN-
SYS, 2006). The Mesher, upon receiving Semantic SOAP request messages, de-
serializes input parameters then generates and executes ANSYS input files which
discretizes problem domains into eight-noded brick meshes. The FEM service,
in a similar manner, takes Semantic SOAP request messages then generates and
executes ANSYS input files to solve particular problems correspondingly. Results
from these services are serialized and returned to service consumers at respective
rendezvous URLs.

Parallel Cluster Computing The ParEFG service, programmed in Java, allowed
other SWSCM services to access the Element-Free Garlerkin Method (EFGM)
meshless analysis engine developed in C and deployed on a Linux parallel com-
puter cluster (Barry and Vacharasintopchai, 2001). It acts as a proxy for SWSCM
services to access the functionality of the EFGM engine which per se takes inputs,
receives commands and delivers outputs through UNIX socket communication
(Metcalfe and Gierth, 1998).

Mobile Computing The User Interface was developed for Windows Mobile personal
digital assistants (PDAs) or smartphones, using the C# programming language
and Microsoft .NET Web Services client APIs (Microsoft, 2003). It relies on the
SA service which acts as a proxy, allowing devices with lower computing power
to access the more powerful computing power provided by ParEFG, FEM and
Max Value services.

Ubiquitous Access The User-Interface was designed to communicate with the SA
service via the General Packet Radio Service (GPRS) on mobile phones (Wikipedia,
2006a) or Wi-Fi Internet connections (Wikipedia, 2006b). In this regard, the sys-
tem prototype illustrates not only the interoperability between .NET and Java
Web services but also the ubiquitous access to computational facilities made pos-
sible by Web Services technology.

5.3.2 A Proof of Concept Prototype
The following describes an example workflow of components in the system pro-
totype to solve a particular problem of finding “the maximum service stresses in an

53

Figure 5.5: Windows Mobile User Interface

ASTM A36 steel plate with the dimensions of 1.00–m wide by 200–cm long with
0.25–in thickness, cantilever-supported along the longer edge, and subjected to the
floor load for libraries specified in the Thai Building Code.”
The workflow started when the user entered the input data into the User In-
terface shown in Figure 5.5. From the left of the figure, the geometry of the problem
was entered at the Geometry screen, and the “Generic ASTM A36 Steel” material,
the “Library Floor” load condition, the “TH-2527” building code, the “Service Load”
load factor, and the “Longer-edge Cantilever” support condition were selected at the
Parameter screen. The Summary screen allowed the user to verify the input data, and
to submit the request to the SA service. Except for the numerical values, the input
parameters were selected through the “Select a Resource” screen which retrieved pa-
rameter choices from an external XML data file. The preferred unit of stress can be
selected on the configuration menu and in this example the unit of stress was set to
pascals (Pa). The content of the SOAP body submitted by the User Interface is shown
in Figure 5.6. The width, length, and thickness of the object are accompanied by re-
spective URIs to the units of measurement. The material type, the support condition,
the building code, the loading condition, and the load factors were specified as URIs
of concepts. The preferred unit of stress was specified as the URI to the concept “Pa.”
The input parameters were deserialized by the SA service into semantic objects
as soon as the request message arrived. Preconfigured to work in the SI units, e.g.,
lengths in meters and stresses in pascals, the application logic of the SA service queried
the knowledge base for conversion factors, and the width, length, and thickness of ob-
jects are converted into the meters unit at the time they were constructed. The seman-
tic objects that represented the modulus of elasticity, density, Poisson’s ratio, tensile
proportional limit, and ultimate strength of the ASTM A36 steel were created using
the assertions about http://www.swscm.net/material-properties.owl#ASTM A36
queried from the knowledge base. Inside the knowledge base component, the material-
properties.owl file, shown in Figure 5.7, was downloaded from www.swscm.net and
OWL statements with the “hasTensileModulusOfElasticity,” “hasDensity,” “hasPoissons-
Ratio,” “hasTensileProportionalLimit,” and “hasTensileStrength” as verbs were re-
trieved. All material property objects were normalized into the SI units in the same
fashion as the length objects.
The SA service next requested the Matchmaker for a service that matches its
ideal mesher profile. The URI to the OWL-S description of the Mesher service was
suggested. The SA service then retrieved the WSDL grounding information from the
OWL-S description. The Semantic SOAP message that contained the problem geom-

54

<analysisRequest xmlns="uri:ppc-client"
xml:base="uri:ppc-client">
<getMaxStress>
<geometry>
5 <width unit="http://std.swscm.org/units.owl#m">
1.00</width>
<length unit="http://std.swscm.org/units.owl#cm">
200</length>
<thickness unit="http://std.swscm.org/units.owl#
10 inch">0.25</thickness>
</geometry>
<material>http://www.swscm.net/material-
properties.owl#ASTM_A36</material>
<support>
15 <condition>http://std.swscm.org/structures.owl#
LongerEdgeCantileverSupport</condition>
</support>
<loading>
<code>http://std.swscm.org/structures.owl#
20 TH2527</code>
<condition>http://std.swscm.org/structures.owl#
LibraryFloor</condition>
<loadFactor>http://std.swscm.org/structures.owl#
ServiceLF</loadFactor>
25 </loading>
<resultUnit> http://std.swscm.org/units.owl#Pa</resultUnit>
</getMaxStress>
</analysisRequest>

Figure 5.6: Sample Request Message from PDA Client

etry was transmitted to the rendezvous URL of Mesher, and the eight-noded brick
element mesh was constructed.
After this, the SA service handled the boundary conditions (BCs). For displace-
ment BCs, it created a semantic array object of prescribed displacements using asser-
tions from the definition of the concept http://std.swscm.org/structures.owl#-
LongerEdgeCantileverSupport, which specified nodal displacements in terms of a set
of nodes at a specific location (e.g., the set of nodes that “has X coordinate equal to
zero”) using the “pinned” node and the “locked” displacement concepts. The OWL
statements that define such concepts are presented in Figure 5.8. The definition of the
pinned node asserted that the displacements ux , uy , and uz for each node be specified
as locked. The locked displacement was asserted as the “Scalar” that “has value” of
0.00 and “has unit” of meters. The force BCs were handled similarly.
Later, the SA service requested the Matchmaker for a service that matches its
ideal structural analysis profile. The ideal input parameters comprised a mesh, the force
and displacement boundary conditions, and the material properties. The ideal output
parameter was an array of stress tensors. Each parameter was supported by an onto-
logical concept. The ideal effect of service was http://std.swscm.org/Structural-
AnalysisEffects.owl#SmoothStressAtIntegrationPoints. The same input and
output parameters that match the ideal sets were described in the ParEFG and FEM
service. However, the FEM service gave the http://std.swscm.org/Structural-
AnalysisEffects.owl#StressAtIntegrationPoints effect, whereas the SmoothStress-
AtIntegrationPoints, a subclass of StressAtIntegrationPoints, was given in the
ParEFG case. The matchmaking algorithm thus recommended the ParEFG service.
The Semantic SOAP message was submitted to the rendezvous URL and the array
of stress tensors at integration points were computed. The Von Mises stresses at the

55

<rdf:RDF>
...
<mat:DuctileMaterial rdf:ID="ASTM_A36">
<rdfs:label>ASTM A36 steel</rdfs:label>
5
<mat:hasDensity rdf:resource="#Steel_Density"/>
<mat:hasPoissonsRatio
rdf:resource ="#Steel_PoissonsRatio"/>
<mat:hasTensileModulusOfElasticity
10 rdf:resource ="#Steel_TensileModulusOfElasticity"/>

<mat:hasTensileProportionalLimit>
<mat:TensileProportionalLimit
rdf:ID="ASTM_A36_TensileProportionalLimit">
15 <mat:representedBy>
<mat:Scalar>
<mat:hasValue rdf:datatype="&double;">36.0
</mat:hasValue>
</mat:Scalar>
20 </mat:representedBy>
<mat:hasUnit rdf:resource="&unit;#ksi"/>
</mat:TensileProportionalLimit>
</mat:hasTensileProportionalLimit>

25 <mat:hasTensileStrength>
<mat:TensileStrength
rdf:ID="ASTM_A36_UltimateStrength ">
<mat:representedBy>
<mat:Scalar>
30 <mat:hasValue rdf:datatype="&double;">58.0
</mat:hasValue>
</mat:Scalar>
</mat:representedBy>
<mat:hasUnit rdf:resource="&unit;#ksi"/>
35 </mat:TensileStrength>
</mat:hasTensileStrength >

</mat:DuctileMaterial>
...
40 </rdf:RDF>

Figure 5.7: OWL Definition of ASTM A36 Steel

integration points were calculated by the application logic of the SA and the Max Value
service was consulted to find the maximum stress. Finally, the SA service identified
the point where the maximum stress occurred, and the stress value and the location
was returned to the User Interface for presentation to the user, as shown in Figure 5.9.
From the analysis, the maximum Von Mises stress in the plate was 0.254 × 109 Pa.
It exceeded the yield stress of the specified ASTM A36 material, which, according to
Figure 5.7, equals 36 ksi (0.249 × 109 Pa). Therefore, the configuration of the structure
needed to be revised for improved safety.

5.4 Summary
This chapter has presented an architecture framework, from the system devel-
opment point of view, for the application of Semantic Web Services technology to
address software interoperability problems in structural engineering. It provides the
implementational details which were identified but not yet addressed by general re-
searchers that are necessary for the practical application of Semantic Web Services in
scientific computing, particularly in computational mechanics.
Based on several software tools, which are products from ongoing research on

56

<rdf:RDF>

<swscm:DisplacementSpecification rdf:ID="Locked">
<mat:representedBy>
5 <mat:Scalar>
<mat:hasValue
rdf:datatype ="&double;">0.00</mat:hasValue>
</mat:Scalar>
</mat:representedBy>
10 <mat:hasUnit rdf:resource="&unit;#m"/>
</swscm:DisplacementSpecification >

<swscm:Support rdf:ID="Pinned">
<rdfs:label>Simple support condition</rdfs:label>
15
<swscm:hasDisplacementSpecification >
<swscm:onComponent rdf:resource="#ux"/>
<swscm:hasSpecification rdf:resource="#Locked"/>
</swscm:hasDisplacementSpecification>
20
<swscm:onComponent rdf:resource="#uy"/>
<swscm:hasSpecification rdf:resouce="#Locked"/>
25
<swscm:onComponent rdf:resource="#uz"/>
<swscm:hasSpecification rdf:resouce="#Locked"/>
30
</swscm:Support>
</rdf:RDF>

Figure 5.8: OWL Deﬁnition of Pinned Node

Figure 5.9: Output Screen on PDA Client

57

the Semantic Web and Semantic Web Services, a system prototype was developed
to demonstrate how the proposed framework can help integrate multiple structural
engineering software modules—potentially developed by various groups of programmers
and deployed on heterogeneous platforms—with the element-free Garlerkin method
software on a parallel computer cluster and the ANSYS finite element analysis software
package on a Windows PC as particular examples. In addition, the system prototype
also demonstrates how ubiquitous access to computational facilities can be achieved
through personal digital assistant and smartphone devices that support Web Services
technology. In this aspect, the proposed framework could be especially useful for
engineers at construction sites.
The proposed framework will become more useful as Web Services technol-
ogy advances. As a platform-independent technology, Web Services allows convenient
access to commercial software and potentially is an effective means both to reduce
practitioners’ cost (through a pay-per-use model) and to prevent software piracy (as
executable codes of Web services are located on remote servers and so are hidden from
users). Thus, it is envisaged that general computational mechanics and structural
engineering software developers will gradually adopt Web Services as their architec-
tural platform, and the proposed framework will be a practical tool to assist in the
integration of Web service software components.
Despite the advantages of the proposed framework and the example implemen-
tation, there are some issues that have not been addressed. In particular, issues related
to trust and security have not been taken into account explicitly. More research work
is also required to ensure that premises from ontological sources and results from ser-
vice providers are credible and can be used in mission-critical environments. Financial
aspects, such as the potential implementation of a pay-per-use model, were not specifi-
cally implemented. However, concepts from e-commerce can be adopted and integrated
into the framework. Last but not least, as the framework depends primarily on on-
tologies, a more complete body of knowledge related to computational mechanics and
structural engineering needs to be captured and published on the Web, and the wide
adoption of such ontologies from software developers are essential for the framework
to work at its full potential.

58

CHAPTER 6
CONCLUSIONS

6.1 Conclusion and Discussion
This dissertation has proposed a structural engineering support system and two
frameworks for the application of semantic computing—by means of metadata and
ontologies—to improve the productivity of structural engineers on engineering design
projects. The first system architecture framework, the Blog and Digital Library Frame-
work for Engineering Knowledge Management (Blog+DL), is based on three software
components that enable a methodology which uses Weblogs and Digital Libraries as
building blocks for a collaborative engineering knowledge management system to assist
engineers in sharing, exchanging, and improving their knowledge, skills and computing
resources. The second system architecture framework proposed is the Semantic Web
Services Framework for Computational Mechanics (SWSCM). The software compo-
nents required for it are specified and the methodology by which they interact to unify
and utilize the scattered wealth of computing resources available is described.
The Blog+DL and SWSCM complement each other in the engineering process.
Blog+DL relies on SWSCM to integrate computing resources shared by individual
engineers. SWSCM can use Blog+DL to disseminate and discover shared computing
resources, as well as to educate and exchange peer opinions about their underlying the-
ories and user experiences. Two prototype systems were developed to demonstrate the
proposed frameworks. The first system illustrates the joint application of Blog+DL
and SWSCM to build support systems that facilitate the computationally oriented
KM process in structural engineering, and the second system illustrates the full po-
tential of SWSCM in facilitating a computationally intensive workflow that involves
heterogeneous software components.
To the best of the author’s knowledge, there have been no attempts to simultane-
ously address the computational and knowledge management aspects of the engineering
process. Existing work in engineering literature that is closely related to this research
consists of Liu et al. (2005), Peng and Law (2004), El-Tayeh and Gil (2007), Mezher
et al. (2005), and Rezgui (2006). The first two address computational aspects of the
engineering process and the latter three address knowledge management. A summary
comparison of their features with the ones of SWSCM and Blog+DL is provided in
Table 6.1.
Peng and Law (2004) propose an Internet-based framework which facilitates the
composition of distributed Web services into a finite-element analysis (FEA) program.
By adopting a service-oriented system architecture (He, 2003), as SWSCM does, the
framework allows new and legacy codes from heterogeneous sites to be dynamically
discovered, bound, and executed by the FEA program, and support for mobile device
clients, such as Palm OS and Windows Mobile PDAs or smartphones, is technically
possible. Their approach differs from the one in this research in that it is based on
the Java Remote Method Invocation (RMI) interoperation technology which predates
the World Wide Web Consortium (W3C) Web Services standard; limiting interoper-
ability with Web services on the Internet, and restricting the supported programming
language to Java. The dynamic service composition and data interoperability issues
were addressed by Liu et al. (2005), a group affiliated with Peng and Law, when they
proposed the Flow-based Infrastructure for Composing Autonomous Services (FICAS),
an infrastructure for the composition of loosely coupled software components with an

59

Table 6.1: Features Comparison with Existing Work
Peng
Liu El-Tayeh Mezher
Blog+DL & and Rezgui
et al. and Gil et al.
SWSCM Law (2006)
(2005) (2007) (2005)
(2004)
KNOWLEDGE SHARING
Documents1 Y – – Y2 Y Y
Computing tools
- Worksheets Y – – – Y –
- Individual software Y – – – Y –
- Subroutines Y – – – – –
Multimedia content Y – – – Y3 Y3
Socialization Y – – Y – –
KNOWLEDGE ARCHIVING
Documents1 Y – – Y2 Y Y
Individual software Y – – – Y –
Multimedia contents Y – – – Y3 –
Web content Y – – – – –
Repository Y – – Y Y Y
Automated review workﬂow Y – – – – –
KNOWLEDGE RETRIEVAL
Ontology-assisted Y – – Y – Y
Keyword tagging Y – – Y – Y
Smart keyword suggestion Y – – – – Y
Metadata search Y – – – – –
Full-text search Y – – – – Y
Categorized browsing Y – – Y Y Y
USER INTERFACE & INTEROPERABILITY
User interface Web – – Web Web Web
Support for Social Network Analysis Y – – Y – –

URL, Windows
Web
System interoperability standard(s) RSS/Atom, – – Proprietary ﬁles
Services4
OAI-MHP sharing
COMPUTATIONAL FEATURES
Support for mobile device clients Y Y Y3 – – –
Service-oriented architecture Y Y Y – – –
Support for legacy codes Y Y3 Y – – –
Heterogeneous platform interoperability Y Y Y – – –
Dynamic service discovery and binding Y Y Y – – –
Dynamic service composition Y Y – – – –
Dynamic data interoperability Y Y – – – –
OWL
Knowledge-base data source – – – – –
ontologies

Web Services
Java
Interoperability standards & Semantic Proprietary – – –
RMI
Web5

Java & major
Supported programming language Java Java – – –
languages6

Compatible with existing Web Services
Y – – – – –
technology
1 including AutoCAD, Excel, PDF, PowerPoint, and Word
2 plain-text only
3 technically possible, although not explicitly discussed by the authors
4 SOAP, WSDL, UDDI
5 HTTP, SOAP, WSDL, RDF, OWL, and OWL-S
6 including C++, C#, Perl, PHP, Python, and VB.NET

60

active data mediation facility. Despite the reported superior performance over SOAP
in high data volume engineering applications, the communication protocol and the ser-
vice description scheme in FICAS are proprietary and not compatible with the W3C
standards, making them incompatible with the emerging development tools in many
programming languages that conform to the W3C specifications. Ontologies and se-
mantic descriptions of terms and Web services were not taken into account in these
proposals, so that Internet-based artificial intelligence technologies that allow more
seamless integration of software components from the Web, cannot be readily incorpo-
rated.
Mezher et al. (2005) present a methodology to build a knowledge management
system in an international engineering consultant firm by using the Microsoft Windows
file-sharing system on the company’s intranet, covering three main offices in different
countries. Each office hosts a file-repository server that stores standard design pro-
cedures, templates, forms, and computer programs, as well as lessons from sites and
project documents. The documents include Word and Excel files as well as AutoCAD
drawings. Using the Internet Explorer Web browser as a front-end user interface, users
can retrieve pieces of knowledge from the repository by browsing through hierarchies of
folders to which they are granted access privileges. Knowledge and computer programs
are exchanged among users in different offices through e-mail but need to be autho-
rized by a superior engineer before they can be stored on a file server. Their approach
differs from the one in this research in that it lacks an integrated set of socializa-
tion tools which facilitate the sharing and developing of tacit knowledge among users.
The hierarchical browsing through explicit knowledge, without support from advanced
metadata and full-text search facilities, will make it difficult for users to precisely re-
trieve relevant knowledge as the size of the repository grows. The system also lacks
automated support for the review workflow which increases the time required for criti-
cal pieces of knowledge to gain approval for publication in the repository. The sharing
of computational software tools in the form of Web Services is also not supported.
An advanced ontology-assisted knowledge retrieval technique for the construc-
tion industry, with document summarization and full-text search capabilities, is pro-
posed in Rezgui (2006); however, review workflow, metadata search, and socialization
techniques are not included in this study. Digital socialization in a construction team
was addressed in El-Tayeh and Gil (2007); nevertheless, the sharing, archiving, and
retrieval of knowledge are limited to plain-text documents.
Like Blog+DL, Rezgui (2006) is based on widely-accepted system interoperabil-
ity standards. Their developments were however aimed at different applications. The
W3C standard compliant Web services adopted in Rezgui (2006) support interoper-
ability at the application programming interface level, which is intended specifically for
software developers; whereas the RSS/Atom and the OAI-MHP features in Blog+DL
support interoperability at the information exchange and consumption level, which is
intended both for developers and general users.

6.2 Recommendations for Further Research
The support system as well as the SWSCM and Blog+DL frameworks proposed
in this dissertation can potentially be applied not only to structural engineering but
also to other computationally-oriented application areas. The implementation of the
prototype systems has spawned many interesting research topics worth further inves-
tigation.

61

In particular, this dissertation has focused on methodologies to more seamlessly
integrate computational services owned or developed by different groups of people on
heterogeneous computing platforms via the Internet or an intranet. One assumption
for such integration is that all services are trustworthy and a second is that they are
accessible free of charge. Issues related to trust, security, and tariffs have not been
addressed by this research. Such conditions, however, are not easily achieved when a
large number of services are deployed and consumed via the Internet, as is clear from
current concerns about malicious software and Web phishing incidences. The trust,
security, and tariff issues of Web services therefore need to be further investigated, in
addition to the refinement of the matchmaking and composition algorithms.
As the framework depends primarily on ontologies, more complete bodies of
knowledge in the target application domains of computational mechanics and structural
engineering need to be captured and published on the Web. The wide adoption of
such ontologies by software developers is essential for the proposed frameworks to be
exploited to their full potential. The Blog+DL framework features an ontology-assisted
keyword suggestion component which assists users in providing proper metadata for the
effective and convergent retrieval of knowledge. A relatively simple keyword detection
algorithm has been implemented in the prototype system for illustrative purposes.
More advanced keyword detection techniques however, are available in the natural
language processing literature. Such techniques should be further explored to improve
the performance of metadata enforcement and knowledge retrieval. Last but not least,
the social networking feature of Weblogs opens Blog+DL to research in social network
analysis which holds the potential to further enrich the capabilities of the Blog+DL-
based knowledge management systems.

62

REFERENCES

Al-Ghassani, A. M., Anumba, C. J., Carrillo, P. M., and Robinson, H. S. (2005).
Tools and techniques for knowledge management. In Anumba, C. J., Egbu, C., and
Carrillo, P., editors, Knowledge Management in Construction. Blackwell Publishing,
Oxford, U.K., 2005. ISBN 1-40-512972-7.

ANSYS (2006). ANSYS Finite Element Analysis Software: ANSYS Structural. AN-
SYS, Inc., 2006. Available online: http://www.ansys.com/products/structural.
asp. [Downloaded: April 29, 2006].

Apache (2003a). Apache Axis 1.1. Apache Software Foundation, 2003. Available
online: http://ws.apache.org/axis.

Apache (2003b). Web Services Invocation Framework 2.0. Apache Software Founda-
tion, 2003. Available online: http://ws.apache.org/wsif.

Apache (2007). Apache Lucene. Apache Software Foundation, 2007. Available online:
http://lucene.apache.org/java/docs/index.html. [Downloaded: November 12,
2007].

Barry, W. and Vacharasintopchai, T. (2001). A parallel implementation of the element-
free Galerkin method. In Proceedings of the 8th East Asia-Paciﬁc Conference on
Structural Engineering and Construction (EASEC-8), page Paper No. 1068, Singa-
pore, December 5-7, 2001. Available online: http://stweb.ait.ac.th/~ st029284/
papers/parallel-efgm-easec-8.pdf. [Downloaded: March 20, 2007].

Basha, S., Cable, S., Galbraith, B., Hendricks, M., Irani, R., Milbery, J., Modi, T.,
Tost, A., and Toussaint, A. (2002). Professional Java Web Services. Wrox Press,
Ltd., 2002. ISBN 1-861003-75-7.

Berners-Lee, T., Masinter, L., and McCahill, M. (1994). RFC1738 Uniform Resource
Locators (URL). The Internet Engineering Task Force (IETF), December 1994.
Available online: http://www.ietf.org/rfc/rfc1738.txt. [Downloaded: August
29, 2003].

Berners-Lee, T., Hendler, J., and Lassila, O. (2001). The Semantic Web. Scientiﬁc
American, 284(5):34–43, May 2001.

Blogger (2007). Blogger homepage, Blogger.com – a Blog Service by Google, 2007.
Available online: http://www.blogger.com. [Downloaded: May 5, 2007].

Boakes, R. (2007). Semantic computing, August 2007. Available online: http://
semanticcomputing.org. [Downloaded: October 4, 2007].

Brown, J. S. and Duguid, P. (2000). The Social Life of Information. Harvard Business
School Press, Boston, MA, U.S.A., 2000.

Burstein, M., Bussler, C., Zaremba, M., Finin, T., Huhns, M. N., Paolucci, M., Sheth,
A. P., and Williams, S. (2005). A Semantic Web Services architecture. IEEE Internet
Computing, 9(5):72–81, September/October 2005. Available online: http://doi.
ieeecomputersociety.org/10.1109/MIC.2005.96.

63

Carrillo, P. and Chinowsky, P. (2006). Exploiting knowledge management: The en-
gineering and construction perspective. Journal of Management in Engineering, 22
(1):2–10, 2006.

Catlett, C. (2002). TeraGrid Primer. The TeraGrid Project, September 2002. Avail-
able online: http://www.teragrid.org/about/TeraGrid-Primer-Sept-02.pdf.
[Downloaded: June 10, 2003].

Chen, H.-M., Lin, Y.-C., and Chao, Y.-F. (2006). Application of web services for
structural engineering systems. Journal of Computing in Civil Engineering, 20(3):
154–164, 2006.

Chew, P., Chrisochoides, N., Gopalsamy, S., Heber, G., Ingraffea, T., Luke, E., Neto,
J., Pingali, K., Shih, A., Soni, B., Stodghill, P., Thompson, D., Vavasis, S., and
Wawrzynek, P. (2003). Computational science simulations based on Web Services.
In Proceedings of International Conference on Computational Science 2003, St. Pe-
tersburg, Russian Federation and Melbourne, Australia, June 2003. Available on-
line: http://www.cs.cornell.edu/stodghil/papers/iccs03.pdf. [Downloaded:
September 3, 2003].

Chiu, K., Govindaraju, M., and Bramley, R. (2002). Investigating the limits of SOAP
performance for scientific computing. In Proceedings of 11th IEEE International
Symposium on High Performance Distributed Computing (HPDC-11), Edinburgh,
Scotland, July 2002. Available online: http://www.extreme.indiana.edu/xgws/
papers/soap-hpdc2002/soap-hpdc2002.pdf. [Downloaded: September 3, 2003].

Cross, R. (1998). Managing for knowledge: Managing for growth. Knowledge Manage-
ment, 1(3):9–13, 1998.

Crystal, D. (2006). Language and the Internet. Cambridge University
Press, Cambridge, U.K., 2 edition, 2006. ISBN 0-52-186859-9. Avail-
able online: http://books.google.com/books?id=cnhnO0AO45AC&pg=PA15&dq=
blogosphere&sig=GTyNVN_DEzaeNCYOiTkZhPU5gUU#PPA15,M1. [Downloaded: May
4, 2007].

Czarnecki, D. (2007). Blojsom – java blog server software, 2007. Available online:
http://www.blojsom.com. [Downloaded: August 22, 2007].

de Geytere, T. (2007). A unified model of dynamic organizational knowledge cre-
ation: Explanation of SECI model of Nonaka and Takeuchi, 12manage B.V., 2007.
Available online: http://www.12manage.com/methods_nonaka_seci.html. [Down-
loaded: April 5, 2007].

Drucker, P. F. (1993). Post-Capitalist Society. HarperCollins, New York, U.S.A., 1993.
ISBN 0-88-730620-9.

DSpace (2007a). Dspace home page, DSpace Federation, 2007. Available online:
http://www.dspace.org. [Downloaded: August 22, 2007].

DSpace (2007b). MIT’s DSpace experience: A case study, MIT Libraries and Hewlett-
Packard Company, 2007. Available online: http://dspace.org/implement/
case-study.pdf. [Downloaded: May 5, 2007].

64

El-Tayeh, A. and Gil, N. (2007). Using digital socialization to support geographically
dispersed aec project teams. Journal of Construction Engineering and Management,
133(6):462–473, June 2007.

Fallside, D. C. (2001). XML Schema Part 0: Primer. World Wide Web Consor-
tium (W3C), May 2001. Available online: http://www.w3.org/TR/xmlschema-0.
[Downloaded: June 23, 2003].

Fielding, R., Gettys, J., Mogul, J., Frystyk, H., Masinter, L., Leach, P., and Berners-
Lee, T. (1999). RFC2616 Hypertext Transfer Protocol – HTTP/1.1. The Internet
Engineering Task Force (IETF), June 1999. Available online: http://www.ietf.
org/rfc/rfc2616.txt. [Downloaded: August 29, 2003].

Foster, I. and Gannon, D. (2003). The Open Grid Services Architecture Platform.
The Global Grid Forum, 2003. Available online: http://www.gridforum.org/
Meetings/ggf7/drafts/draft-ggf-ogsa-platform-2.pdf. [Downloaded: June
16, 2003].

Foster, I., Kesselman, C., and Tuecke, S. (2001). The anatomy of the Grid: En-
abling scalable virtual organizations. International Journal of Supercomputer Appli-
cations, 15(3), 2001. Available online: http://www.globus.org/research/papers/
anatomy.pdf. [Downloaded: June 6, 2003].

Foster, I. (2002). What is the Grid? A three point checklist. July 2002. Avail-
able online: http://www-fp.mcs.anl.gov/~ foster/Articles/WhatIsTheGrid.
pdf. [Downloaded: April 29, 2003].

GGF (2003). The Global Grid Forum website, The Global Grid Forum, 2003. Available
online: http://www.gridforum.org. [Downloaded: June 13, 2003].

Haarslev, V. and Moller, R. (2003). RACER User’s Guide and Reference Manual
(Version 1.7.7), November 2003. Available online: http://www.cs.concordia.ca/
~haarslev/racer/index.shtml. [Downloaded: January 30, 2004].
He, H. (2003). What is service-oriented architecture, September 2003. Available
online: http://webservices.xml.com/pub/a/ws/2003/09/30/soa.html. [Down-
loaded: September 19, 2007].

IPG (2003). What is the IPG? Information Power Grid Project, NASA, October 2003.
Available online: http://www.ipg.nasa.gov/aboutipg/what.html. [Downloaded:
June 11, 2003].

Jakobsen, T. (2006). Bookmarklets - Browser Power, April 2006. Available online:
http://subsimple.com/bookmarklets. [Downloaded: November 12, 2007].

Jena (2007). Jena – a Semantic Web framework for java, Hewlett-Packard Laboratories,
2007. Available online: http://jena.sourceforge.net. [Downloaded: November
12, 2007].

Jonathan Cohen & Associates (2004). Knowledge management for design ﬁrms: Case
studies, April 2004. Available online: http://www.jcarchitects.com/KM.html.
[Downloaded: April 27, 2007].

65

Lagoze, C., Van de Sompel, H., Nelson, M., and Warner, S. (2004). The Open
Archives Initiative Protocol for Metadata Harvesting. Open Archives Initiative
(OAI), October 2004. Available online: http://www.openarchives.org/OAI/
openarchivesprotocol.html. [Downloaded: May 11, 2007].

Liu, D., Peng, J., Law, K. H., Wiederhold, G., and Sriram, R. D. (2005). Composition
of engineering web services with distributed data-ﬂows and computations. Advanced
Engineering Informatics, 19:25–42, 2005.

Martin, D., Burstein, M., Hobbs, J., Lassila, O., McDermott, D., McIlraith, S.,
Narayanan, S., Paolucci, M., Parsia, B., Payne, T., Sirin, E., Srinivasan, N., and
Sycara, K. (2004). OWL-S: Semantic Markup for Web Services. World Wide Web
Consortium (W3C), November 2004. Available online: http://www.w3.org/
Submission/OWL-S. [Downloaded: March 15, 2007]. W3C Member Submission.

Metcalfe, V. and Gierth, A. (1998). Programming UNIX Sockets in C - Frequently
Asked Questions, January 1998. Available online: http://www.faqs.org/faqs/
unix-faq/socket. [Downloaded: November 24, 2007].

Mezher, T., Abdul-Malak, M. A., Ghosn, I., and Ajam, M. (2005). Knowledge man-
agement in mechanical and industrial engineering consulting: A case study. Journal
of Management in Engineering, 21(3):138–147, 2005.

Microsoft (1996). DCOM technical overview, November 1996. Available online:
http://msdn2.microsoft.com/en-us/library/ms809340. [Downloaded: March
13, 2007].

Microsoft (2003). Microsoft .NET, Microsoft Corporation, 2003. Available online:
http://www.microsoft.com/net. [Downloaded: November 5, 2003].

MIT Libraries and Hewlett-Packard Company (2007). Introducing DSpace. DSpace
Federation, 2007. Available online: http://dspace.org/introduction/index.
html. [Downloaded: May 11, 2007].

Mozilla (2007). Mozilla ﬁrefox home page, Mozilla Corporation, 2007. Available online:
http://www.mozilla.com. [Downloaded: August 22, 2007].

Nardi, D., Brachman, R. J., Baader, F., Nutt, W., Donini, F. M., Sattler, U., Calvanese,
D., Molitor, R., De Giacomo, G., Kusters, R., Wolter, F., McGuinness, D. L., Patel-
Schneider, P. F., Moller, R., Haarslev, V., Horrocks, I., Borgida, A., Welty, C.,
Franconi, A. R. E., Lenzerini, M., and Rosati, R. (2003). The Description Logic
Handbook: Theory, Implementation and Applications. Cambridge University Press,
Cambridge, U.K., 2003. ISBN 0-521-78176-0.

NIST (1999). JAMA: Java Matrix Package. National Institute of Standards and
Technology, 1999. Available online: http://math.nist.gov/javanumerics/jama.

Nonaka, I. and Takeuchi, H. (1995). The Knowledge-Creating Company: How Japanese
Companies Create the Dynamics of Innovation. Oxford University Press, New York,
U.S.A., 1995. ISBN 0-19-509269-4.

66

Nottingham, M. and Sayre, R. (2005). The Atom Syndication Format. The Internet
Engineering Task Force (IETF), December 2005. Available online: http://www.
ietf.org/rfc/rfc4287. [Downloaded: May 8, 2007].
OASIS (2002). Universal Description, Discovery & Integration (UDDI) Speciﬁcation
Version 2. Organization for the Advancement of Structured Information Stan-
dards (OASIS) Consortium, 2002. Available online: http://www.oasis-open.org/
committees/uddi-spec/tcspecs.shtml#uddiv2.
OMG (2005). The Common Object Request Broker Architecture: Frequently
asked questions, 2005. Available online: http://www.omg.org/gettingstarted/
corbafaq.htm. [Downloaded: March 13, 2007].
OWL–S (2003). OWL-S: Semantic markup for Web Services. Technical report, The
OWL Services Coalition, November 2003. Available online: http://www.daml.org/
services/owl-s/1.0/owl-s.pdf. [Downloaded: April 30, 2004].
Palme, J., Hopmann, A., and Shelness, N. (1999). RFC2557 Multipurpose Internet
Mail Extension (MIME) Encapsulation of Aggregate Documents, such as HTML
(MHTML). The Internet Engineering Task Force (IETF), March 1999. Available
online: http://tools.ietf.org/html/rfc2557. [Downloaded: November 8, 2007].
Palmer, S. B. (2002). RDF in HTML: Approaches, June 2002. Available online:
http://infomesh.net/2002/rdfinhtml. [Downloaded: March 14, 2007].
Paolucci, M., Kawamura, T., Payne, T. R., and Sycara, K. (2002). Semantic match-
ing of web services capabilities. In Proceedings of the First International Seman-
tic Web Conference (ISWC2002), pages 333–347, Sardinia, Italy, 2002. Available
online: http://www-2.cs.cmu.edu/~ softagents/papers/ISWC2002.pdf. [Down-
loaded: October 31, 2003].
Peng, J. and Law, K. H. (2004). Building ﬁnite element analysis programs in distributed
services environment. Computers and Structures, 82:1813–1833, 2004.
Pollock, N. (2001). Knowledge management: Next step to competitive advan-
tage. Journal of the Defense Systems Management College, September-October
2001. Available online: http://www.providersedge.com/docs/km_articles/KM_
-_Next_Step_to_Competitive_Advantage.pdf. [Downloaded: April 5, 2007].
Prot´g´ (2005). The Prot´g´ Ontology Editor and Knowledge Acquisition System. Stan-
e e e e
ford Medical Informatics Department, Stanford University, Stanford, CA, U.S.A.,
2005. Available online: http://protege.stanford.edu. [Downloaded: March 20,
2006].
Quintas, P. (2005). The nature and dimensions of knowledge management. In Anumba,
C. J., Egbu, C., and Carrillo, P., editors, Knowledge Management in Construction.
Blackwell Publishing, Oxford, U.K., 2005. ISBN 1-40-512972-7.
Reddy, R., Ager, T., Chellappa, R., Croft, W. B., Davis-Brown, B., Mendel, J. M.,
and Shamos, M. I. (1999). WTEC Panel report on digital information organization
in Japan. Technical report, Loyola College in Maryland, 1999. Available online:
http://www.wtec.org/loyola/digilibs/d_01.htm. [Downloaded: September 11,
2006].

67

Rezgui, Y. (2006). Ontology-centered knowledge management using information re-
trieval techniques. Journal of Computing in Civil Engineering, 20(4):261–270, July
2006.
Roman, D., Keller, U., Lausen, H., de Bruijn, J., Lara, R., Stollberg, M.,
Polleres, A., Feier, C., Bussler, C., and Fensel, D. (2005). Web ser-
vice modeling ontology. Applied Ontology, 1(1):77–106, 2005. Available
online: http://iospress.metapress.com/openurl.asp?genre=article&issn=
1570-5838&volume=1&issue=1&spage=77. [Downloaded: March 15, 2007].
RSS Advisory Board (2006). RSS 2.0 Speciﬁcation. RSS Advisory Board, August 2006.
Available online: http://www.rssboard.org/rss-specification. [Downloaded:
May 8, 2007].
Sheehan, T., Poole, D., Lyttle, I., and Egbu, C. O. (2005). Strategies and business case
for knowledge management. In Anumba, C. J., Egbu, C., and Carrillo, P., editors,
Knowledge Management in Construction. Blackwell Publishing, Oxford, U.K., 2005.
ISBN 1-40-512972-7.
Sifry, D. (2007). The state of the Live Web, April 2007, Technorati, Inc., April 2007.
Available online: http://www.sifry.com/alerts/archives/000493.html. [Down-
loaded: May 8, 2007].
Sirin, E. (2004). OWL-S API, Maryland Information and Network Dynamics Lab
Semantic Web Agents (MINDSWAP) Project, Institute for Advanced Computer
Studies, University of Maryland, 2004. Available online: http://www.mindswap.
org/2004/owl-s/api. [Downloaded: June 29, 2004].
Sun Microsystems, Inc. (2007). Javamail api, Sun Microsystems, Inc., 2007. Avail-
able online: http://java.sun.com/products/javamail. [Downloaded: November
8, 2007].
Sycara, K., Paolucci, M., Ankolekar, A., and Srinivasan, N. (2003). Automated discov-
ery, interaction and composition of Semantic Web Services. Web Semantics: Science,
Services and Agents on the World Wide Web, 1(1):27–46, December 2003. Avail-
able online: http://dx.doi.org/10.1016/j.websem.2003.07.002. [Downloaded:
March 15, 2007].
Trott, B. and Trott, M. (2002). TrackBack Technical Speciﬁcation, October 2002.
Available online: http://www.movabletype.org/docs/mttrackback.html. [Down-
loaded: May 8, 2007].
Turban, E. and Aronson, J. (2001). Decision Support Systems and Intelligent Systems.
Prentice Hall, London, U.K., 2001.
Vacharasintopchai, T. (2005). KMITL/SUT: Reinforced Concrete Design Work-
sheets. CE Resources Weblog, October 2005. Available online: http://ce-res.
blogspot.com/2005/10/kmitlsut-reinforced-concrete-design.html. [Down-
loaded: March 14, 2007].
Vacharasintopchai, T., Barry, W., Wuwongse, V., and Kanok-Nukulchai, W. (2007a).
Semantic Web Services framework for computational mechanics. Journal of Com-
puting in Civil Engineering, 21(2):65–77, March-April 2007.

68

Vacharasintopchai, T., Wuwongse, V., and Chalermsook, K. (2007b). Weblog, Digital
Library, and Semantic Web Services approach to computer-aided engineering design.
In Proceedings of the 8th Asia Pacific Conference on Industrial Engineering and
Management System (APIEMS2007), Kaohsiung, Taiwan, R.O.C., December 9-13,
2007.

van Engelen, R. A. (2003). Pushing the SOAP envelope with Web Services for scientific
computing. In Proceedings of the International Conference on Web Services (ICWS)
2003, Las Vegas, Nevada, USA, June 2003. Available online: http://www.cs.fsu.
edu/~engelen/icws03.pdf. [Downloaded: September 3, 2003].

W3C (1999a). HyperText Markup Language (HTML) 4.01 Specification.
World Wide Web Consortium (W3C), December 1999. Available online: http:
//www.w3.org/TR/html4.

W3C (1999b). Resource Description Framework (RDF) Model and Syntax Specifi-
cation. World Wide Web Consortium (W3C), February 1999. Available online:
http://www.w3.org/TR/1999/REC-rdf-syntax-19990222. [Downloaded: Septem-
ber 19, 2003].

W3C (2000). Simple Object Access Protocol (SOAP) 1.1. World Wide Web Consortium
(W3C), 2000. Available online: http://www.w3.org/TR/SOAP.

W3C (2001). Web Service Definition Language (WSDL) 1.1. World Wide Web Con-
sortium (W3C), 2001. Available online: http://www.w3.org/TR/wsdl.

W3C (2003). RDF Primer. World Wide Web Consortium (W3C), September
2003. Available online: http://www.w3.org/TR/2003/WD-rdf-primer-20030905.
[Downloaded: September 22, 2003]. W3C Working Draft.

W3C (2004a). OWL Web Ontology Language Overview. World Wide Web Consortium
(W3C), February 2004. Available online: http://www.w3.org/TR/owl-features/.
[Downloaded: March 15, 2004]. W3C Recommendation.

W3C (2004b). OWL Web Ontology Language: Semantics and Abstract Syntax. World
Wide Web Consortium (W3C), February 2004. Available online: http://www.w3.
org/TR/owl-semantics. [Downloaded: March 16, 2007].

W3C (2004c). RDF Semantics. World Wide Web Consortium (W3C), February 2004.
Available online: http://www.w3.org/TR/rdf-mt. [Downloaded: March 14, 2007].

W3C (2004d). Web Services Architecture. World Wide Web Consortium (W3C),
February 2004. Available online: http://www.w3.org/TR/ws-arch. [Downloaded:
March 13, 2007]. W3C Working Group Note.

Wielemaker, J., Schreiber, G., and Wielinga, B. (2003). Prolog-based infrastructure
for RDF: Performance and scalability. In Fensel, D., Sycara, K., and Mylopoulos, J.,
editors, Proceedings of the 2nd International Semantic Web Conference (ISWC’03),
pages 644–658, Sanibel Island, Florida, USA, October 20-23, 2003. Available
online: http://hcs.science.uva.nl/projects/SWI-Prolog/articles/iswc-03.
pdf. [Downloaded: April 25, 2006].

69

Wikipedia (2006a). General packet radio service, Wikipedia, The Free Encyclopedia,
April 13, 2006. Available online: http://en.wikipedia.org/wiki/GPRS. [Down-
loaded: April 30, 2006].

Wikipedia (2006b). Wi-ﬁ, Wikipedia, The Free Encyclopedia, April 30, 2006. Available
online: http://en.wikipedia.org/wiki/WiFi. [Downloaded: April 30, 2006].

Wikipedia (2007a). Application Programming Interface. Wikipedia, The Free En-
cyclopedia, March 2007. Available online: http://en.wikipedia.org/wiki/API.
[Downloaded: March 14, 2007].

Wikipedia (2007b). Blog. Wikipedia, The Free Encyclopedia, May 2007. Available
online: http://en.wikipedia.org/wiki/Blog. [Downloaded: May 2, 2007].

Wikipedia (2007c). Blogosphere. Wikipedia, The Free Encyclopedia, April 2007. Avail-
able online: http://en.wikipedia.org/wiki/Blogosphere. [Downloaded: May 4,
2007].

Wikipedia (2007d). Digital Library. Wikipedia, The Free Encyclopedia, April
2007. Available online: http://en.wikipedia.org/wiki/Digital_library.
[Downloaded: May 2, 2007].

Wikipedia (2007e). Web Feed. Wikipedia, The Free Encyclopedia, May 2007. Available
online: http://en.wikipedia.org/wiki/Web_feed. [Downloaded: May 2, 2007].

Witten, I. H. and Bainbridge, D. (2003). How to Build a Digital Library. Morgan
Kaufmann Publishers, San Francisco, CA, U.S.A., 2003. ISBN 1-55-860790-0.

Witten, I. H., Boddie, S., and Thompson, J. (2006). Greenstone Digital Library User’s
Guide. University of Waikato, New Zealand, March 2006.

WordPress (2007). Wordpress homepage, Automattic, Inc., 2007. Available online:
http://wordpress.com. [Downloaded: May 5, 2007].

WSDL4J (2006). Web Services Description Language for Java (WSDL4J) toolkit,
November 2006. Available online: http://sourceforge.net/projects/wsdl4j.
[Downloaded: November 19, 2007].

70

A Structural Engineering Support System using Semantic Computing

More Related Content

Viewers also liked (6)

Similar to A Structural Engineering Support System using Semantic Computing (20)

More from Dr. Thiti Vacharasintopchai, ATSI-DX, CISA (18)

Recently uploaded (20)

A Structural Engineering Support System using Semantic Computing