IEEE --- 2005 International Conference on Emnerging Technologies

September 17-18, Islaimabad

Web Engineering: A New Emerging

Discipline1
Rashid Ahmad2, Zhang Li, Farooque Azam
Department oJ Computer Science
Beihang University of Aeronautics & Astronautics (BUAA), Beijing, China
tr.ahmad, lily, firooque [email protected]

Abstract 'CERN' in 1990. Mr. Tim Berners-Lee with a

background in text processing and real-time
Web Engineerfing, an emerging new discipline, communications, wanted to create an information
advocates a process and a systematic approach to system in which researchers could exchange
development of high quiality Web based systems. In information during the course of a project. He came up
contrast the commercial practice still remains ad-ho0C. with a method of linking documents together using
Although the web based systems has becomne hypertext technology. Since then this emerging
increasingly complex, the development process is still discipline went through many mile stones. Table no 1
uin-engineered. "There are very ftw standard methods mentions some of the major mile stones.
Jbr the Web developers to uise. IHIence, there is a strong
need to uinderstand and undertake Web Engineering Table 1; Journey of Web Engineering
[1] This paper is a result of ouJr extensive survey of 1990 * World Wide Web originates- 1990
* .Hypertext Markup Language written- 1991
literatuire, ouir work and interaction with developers. 1995 * Text-based browser opens WWW for use- 1992
This paper gives an introductory overview on Web *New Web terms: HTTP and URL-1992
Engineering. It assesses similarities and difrrences * Graphical user interface, Mosaic, developed- 1993
between development of traditional software and Web * HTML introduced as code for web design- 1993
based systems, and reviews some of the ongoing work * World Wide Web Consortium,W3C founded- 1994
in this area. We discuss the need/fbr development of 1996 * First Paper about Web Engineering related
process modeljbr web based applications (WBA), sutch - problems appear at the 5"' WWW Conference in
as is caviailable for conventional softwvare, with an 2001 France- 1996
overview of our work on dlevelopnment of such a * Microsoft releases Explorer 3.0 Web browse- 1996
process framework [2]. This paper also attempts to * Web Engineering Homepage published
highlight the areas that need further stuidy. WebEngineering.org -1997
* First workshop on Web Engineering organized at
1. Introduction and history the 7th WWW Conference in Australia- 1998
* First International Conference on Web Engineering
A very simple but inclusive definition of Web (ICWE)-200 1
2002 * For the first time Web Engineering became part of
application amongst the many available we found We - the main program-2002
define a WBA as "any software application that 2005 * First international Journal on Web Engineering
depends on the Web fbr its correct execution" [3] [3] published by Rinton Press, USA-2002
Web Engineering is defined [4], by experienced * Web Engineering became an area of the refereed
researchers in the field, as: "The application of paper track at the WWW conference, demonstrating
systematic, disciplined and quantifiable approaches to the emerging need for a discipline on development,
development, operation, and maintenance of Web- operation, and evolution of Web applications-2003
based Injormation Systems (WIS) . W3C describes principles of Web architecture-2004
The World Wide Web (WWW) originated at the . W3C actively busy in standardizing protocols and
European Particle Physics Laboratory known as web technologies-2005

This work is supported by National University of Science aand Technology (NUST). Rawalpinidi, Pakistan
Corresponding Author: Email: r.ahrmad(0&buaa.edu.cn

0-7803-9247-7/05/$20.00 ©2005 IEEE 445

IEEE --- 2005 International Conference on E1nerging Technologies
Septemttber 17-18, Islamiiabad

network/Internet) are prone to attacks by hackers

2. IWeb, Software] Engineering: Clones who are not even Cuistomers.
or Siblings?
Data released (Figure 1) by CERT® Coordination
Center under the heading "Recent CERT/CC
Pressman [5] says Web Engineering is not a Experiences" [11], gives an idea of the enormous
perfect "Clone" of software engineering but it borrows number of "Reported" attacks. Despite the fact that
many of software engineering's ftindamental concepts
CERT/CC people say the organizations do not report
and principles, emphasizing the same technical and the attacks for the fear not to loose the confidence of
management activities. .Difference between Web the customers
Engineering and Software Engineering received much We conclude this debate by calling Web
debate from the researchers [6], [7], [8], [9], [10] in the Engineering and Software Engineering as sibling.
last few years. Many researchers have identified very assuming Software Engineer and Web Engineer as
interesting differences. We just enlist the few which parents (leaving aside to know who is female!).
has impact on development process (for detail study of
100,000
their impact we refer the reader to [7]: Total incidents reported (1989-2002): 182,463
* WBA have increased emphasis on user interface 90,0001
Opcn modularized architectures 80,000 82,094 " i
* Link between business model and architecture 70,000
Rapidly changing technologies 60,000
More content-driven than software engineering. 50,000
* Increased importance of quality attributes
40,000
* Client uncertainty
30,000 21,756
* Changing business requirements

* Short time frames for initial delivery 20,0009,5

Fine-grained evolution and maintenance 10,000
* Highly competitive 0
We however identify another important difference cn
- 0 0

CO

which has not yet been discussed by others. Author [5]

says conventional software system has three "Players". Figure 1. Total Incidents reported (1988-2003): 182, 483
They are Client (owner who orders the software
system also known as customer), Developer, and User 3. Need for a Process Models
(who uses the software system). We argue that the
WBA has four "Players". They are Client (owner who The discussion above encourages us to view WBA
orders the software system also known as customer), as special kind of as software systems: and therefore,
Developer, User (who uses the software system e.g. to blindly apply existing models to the representation
webmaster etc) and Customers (people accessing of WBA persuades the developers to overlook the
websites). This increase in Players (Cufstomers) has a peculiarities of these WBA, and hence not address
great impact on the issues of Quality and Security. these peculiarities, leading to inappropriate solutions.
Because the number of the customers and their This does not imply that existing models should not be
expectations would always remain a variable and utilized; rather we need to do so with an awareness of
would not be predictable, this would always have an their limitations with respect to the aspects of WBA
impact on the Quality and Security in the following that we wish to understand and document. We also
manner: need to understand how these limitations may be
. In case of conventional software system for Quality circumvented by appropriately supplementing (or
the Developer has to satisfy Client and a known replacing where necessary) the models [7]. In this
number of Users only. Whereas incase of WBA section we discuss few more aspects which necessitate
other than these two (client, user), the system must that process models for engineered web development
serve the Quality needs of a diverse community of have to be used.
Customers which can grow enormously with a
varying intellectual level.
In case of conventional software system "security" is The CERT Coordination Center is part of the Software
the issue with in organization but WBA being Engineering Institute. The Software Engineering Institute is
network intensive (i.e. resides on the sponsored by the U.S. Department of Defense.

446
 IEEE --- 2005 International Conference oni Eiiierginzg Technologies
Septcmiber 17-18, Islamabad

3.1. Rapid shift from simplicity to comnplexitv 4. Ongoing research work

The complexity of WBA has grown significantly In this section we have discussed the current
and very rapidly. Table no 2 explains it further. [ 12] active research work which we have identified in our
thorough survey. We believe that this section is the
Table 2; A shift from simplicity to complexity strength of our paper as this gives complete overview
Past Systems Present Svstems of the present and future of the Web Engineering as a
* Simple Web pages * Complex Web pages discipline- a good food for a starter.
* Less emplhasis on aesthetics! * More emphasis on
user interface aesthetic, user interface 4.1. MDA (Model Driven Architecture) for
* Iniformation content static * Information is dynaimic
* Simple navigation * Complex navigate
Web Engineering
* Stand-alone systems * Integrated with database
and other systems Model-Driven Software Engineering (MDE) is
* High performance wasn't a * Requires high performance becoming a widely accepted approach for developing
major requirement complex distributed applications The MDA defined by
* Developed by a single * Requires a large OMG (object management group), is an approach to
individual/small team development team using models in software development, with the goals
* Used for information * Deployed in mission- of portability, interoperability and reusability through
dissemination in non core critical applications architectural separation of concerns. Software
applicationis I development in the MDA starts with a Platform-
Independent Model (PIM) of an application's business
3.2. Previous experience with WBA fuinctionality and behavior, constructed using a
miiodeling language based on OMG's MetaObject
Ad-hoc devxelopment of WBA has brought Facility (MOF). This model remains stable as
disasters to many organizations. A survey on Web- technology evolves (thus curing the problem of rapid
based project development by the Cutter Consortium change in technology peculiar to WBA), extending and
[13] highlighted problems plaguing large Web-based thereby maximizing software ROI (return on
projects: investment). Researchers are working to see how
* Delivered systems didn't meet business needs 84 MDA can be used for the design and implementation
percent of the time. of WBA. In particular, to show how it may help
* Schedule delays plagued the projects 79 percent of organize the design of an application right from its
the time. CIM (Computation Independent Model) at the higher
* Projects exceeded the budget 63 percent of the time. level of abstraction, to its implementation using
* Delivered systems didn't have the required different technologies (such as WebServices, .NET,
fLunctionality 53 percent of the time. J2EE, CORBA, etc.). Efforts are being made to
* Deliverables were of poor quality 52 percent of time. integrate existing Web technologies into the MDA
approach. For more information in this direction see
3.3. Quality and security of Software [25], [26]

The security of a software-intensive system is 4.2. Web intelligence systems

directly related to the quality of its software. This is
evident by the following data released by Software Web Intelligence is a new direction for scientific
Engineering Institute [14]: research and development that explores the
* Over 90% of software security incidents are caused
fundamental roles as well as practical impacts of
artificial intelligence and advanced information
by attackers exploiting known software defects.
technology for the next generation of Web-empowered
* Analysis of 45 e-business applications showed that
systems, services, and environments. To accurately
70% of secuLrity defects were design defects. quantify success, WBA must examine visitor profiles
* Experienced and capable software engineers inject,
and online behavior, instead of the number of page
on average, one defect every nine lines of code.
views. To support business goals, companies need a
* A one million line of code systems typically contains
robust Web intelligence model capable of
1,000-5,000 defects when shipped. accommodating new strategies and technologies. The
basis of Web intelligence is the ability to uniquely

447
IEEE --- 2005 International Conference on Emnerging Technologies
Septemitber 17-18, Islamnabad

identify and segment users. Web Intelligence is research work is addressing issues related to
regarded as the key research field for the development implementation of measurements and metrics for Web
of the Wisdom Web (including the Semantic Web). It cost estimation, Web quality measurement, Web
is one of the most important as well as promising IT testing, Web reliability measurement, Web process
research fields in the era of Web and agent improvement, WBA development methods and
intelligence. Web Intelligence (WI) exploits the methodologies and, Web usability measurement.
fundamental and practical impact that advanced Existing metrics require better interpretation (e.g.
Information Technology (IT) and innovative Artificial mapping from low level measures to high level quality
Intelligence (Al) will have on the next generation. For factors). Proposed dimensions for future research are:
more information see the resources [27], [28], [29], Age, Rate of Change, Evolution Strategy, Size,
[30] Domain, Purpose, Functionality and Technology.
Autlhors [38] have done a great job in collecting
4.3. Agent based Web Engineering references about the global research activities in this
area. Also see [39], [40], [41], [42], [43]
Agent-based approaches to systems development
are becoming increasingly popular. The rise in the 4.5. Business model integration into web based
popularity of agent-based approaches coincided with systems
the advent of the web, and was, in part, driven by it. At
this time, agent applications drive web search engines For the development of WBA to be effective, we
and underpin the semantic web initiative. Agent- need to consider two aspects of technical architecture:
oriented conceptual modeling techniques are emerging the information architecture and the fiinctional
as popular approaches to web application modeling. architecture. Web modelling languages (WML), used
Model-driven Web Information Systems (WIS) design in the development of these systems need to be able to
methodologies have been recently influenced by these capture and communicate both these aspects. It has
emi-erging technologies like the Semantic Web, been argued [44] that a number of gaps exist in
(Semantic) Web Services, and (Semantic) Web Rule contemporary modelling approaches, particularly with
Languages, which foster application interoperability. regard to the level of support for modelling of the
Semantic Web languages (like RDF(S), OWL) linkage between information and fuinctional designs,
facilitate the description of WIS models. Agent-based primarily at the architectural level; and the connections
monitoring, both of run-time code and of the entire between business models and the system designs.
application development life-cycle, is emerging as an However, following two limitations are of great
interesting proposition. Research work in this direction importance and interest: "Disconnection between
includes agent based methodologies for web business model and technical architecture" and
engineering, agent-oriented conceptual modeling for "Disconnection between fiunctional architectutre and
WBA, agent-oriented web process modeling, agent information architectuire ". Most of the approaches aim
based web information systems, semantic web agents, to facilitate the Design and Implementation of software
agent based web services, agent-based web process systems. However, Requirement Engineering and
automation etc. This field is very closely related to Business Modelling are not a focus for the majority of
"Web intelligence systems" discussed above. For the approaches. Due to the unique characteristics of
prerequisite knowledge to research in this direction see WBA, including the uncertainty in the business
[31], [32] [33] and for current work in this field see domain and the rapidly changing user requirements,
[34], [35], [36], [37] and also references in section 5.2. there is no commonly accepted and well-proven
approach that can effectively capture the business
4.4. Web measurements and metrics requirements and translate them into technical
architecture during WBA design. For current status of
Measurement can play a significant role in the work in this area see [44], [45]. and [46]
effective management and development of WBA. It
provides the scientific basis for Web engineering to 4.6. Anytime-Anywhere-Anymedia paradigm
become a true engineering discipline. As Web
engineering as a discipline is now emerging from its We are facing a new generation of web
infancy. Web measurement and metrics will receive applications being characterised by the
growing attention, acceptance, and recognition in the ani'time/anywhere/anymedia paradigm, thus providing
future, like in all other engineering disciplines. Current ubiqluitouis access to services, tuming e-conmerce into

448
IEEE --- 2005 International Conference on Emerging Technologies
Septemliber 17-18, Islam1abad

m-commerce. Over the next decade, millions of have discussed these issues in [2]. We have cured it by
businesses, billions of people and trillions of devices integrating it with another well established
will be connected. The promise of this envisioned methodology "six sigma for software" in our previous
future is access at anytime, from anywhere, with work [2].
anvvmedia, to everyone and everything, thereby
boosting productivity and enabling more satisfying 5. Conclusion
ways to get things done. This will radically change the
way how people will interact, how business is Web Engineering is in its infancy but growing
conducted, and how our every day life will be very fast. As participants in WBA development, we
organized. The internet and the World Wide Web in need to take our bearings fairly frequently and identify
particular will be the technological bases for that issues and problems, both philosophical and practical.
fundamental change. Ubiquity with its This paper has attempted, in parallel with others, to
anytime/anywhere/anymedia nature requiring context- focus on some of them and tried to highlight some
aware computing calls for new engineering techniques significant areas.
supporting these kinds of services. Research is
underway to propose the notion of customization as the 6. References:
uniform mechanism to deliver ubiquitous WBA
providing adaptability with respect to a certain context. [1] Yogesh Deshpande, Martin Gaedke "Web Engineering:
Also efforts are being made to find solution for Developing Successful Web Applications In A
modeling such device independent application. For Systematic Way" 14th Intermational World Wide Web
more information on current work see [47], [48], [49] Conference, 10-14 May, 2005, Chiba, Japan
[2] R Ahmad, Z Li, F Azam "Web OPEN Integrated:
4.7. Process models for development Proposed Framework for Web Development",
ISSN: 0302-9743, pp 533- 538, volume 3579 of Lecture
Notes in Computer Science, Springer-Verlag, July 2005
From the discussion in section 3 it becomes clear [3] Gellersen, H.-W. and M. Gaedke. "Object-Oriented
that the construction and evolution of WBA requires Web Application Development", IEEE Internet
support such as is available for traditional applications Computing 3 (1), 1999, pp. 60-68
through process models, architectures, methods and [4] Deshpande Y, Murugesan S, Ginige A, Hansen S,
principles of software engineering [15]. However there Schwbe D, Gaedke M & White B "Web Engineering"
are very few potential choices of software Journal of Web Engineering, 1(1), 3-17, 2002
development process for WBA development. [1] [5] Pressman, R S 'Sofiware Engineering: a Practitioner's
Researchers have tried to extend or integrate already Approach (4th ed.)',McGraw-Hill Publishing Company
[6] S. Overmyer, "What's different about requirements
available process models for conventional software to engineering for web sites?," Requirements Engineering
apply it to WBA development. See for example [16], Jouirnal, vol. 5. no. 1, pp. 62-65. 2000.
[1 7], [18], [19], [20]. Recently in our work [2] we have [7] Lowe, D., & Henderson-Sellers. B. "'Impacts on the
proposed a process framework named "Web OPEN development process of differences between web
Integrated". OPEN (Object-oriented Process, systems and conventional software systems",
Environment, and Notation) is a process-focused International Con/erence on Advances in Infrastructure
methodological approach to software-intensive for Electronic Business, Science, and Education on the
systems development useful for both object-oriented Internet, L'Aquila, Italy, 2001
and Component Based Development (CBD). OPEN [8] T. Powell, D. Jones and D. Cutts: Web Site
was developed and is maintained by the not-for-profit Engineering: Beyond Web Page Design. Prentice Hall,
1998
OPEN Consortium, an intemational group of [9] Jesper H & Torkil C "What makes web-development
methodologists, CASE tool vendors and developers. It different?" 24th Information Systems Research Seminar
is documented in a series of books (e.g. [20] [21], [22], Norway, Aug 2001
[23], [24]) and in many journal articles particularly [10] Pressman, R S 'Can Internet-Based Applications Be
JOOP. The component-based nature of OPEN permits Engineered?, IEEE Software, 15(5), 104-110, 1998
appropriate extensions to support development in new [ 11] "CERT/CC Overview" (D 1998-2003 by Camegie
domains. One such set of extensions is those for WBA Mellon University
development, called Web OPEN (see Haire et al. [16], [12] A.Ginige & S.Murugesan, "Web Engineering: An
[ 17] for a more details). Web OPEN which specifically Introduction," IEEE MultiMedia, vol. 8, no. 1, Jan.-
Mar. 2001, pp 14-18
focuses on WBA development is handicapped with [13] Cuitter Consortium Research Briefs. 7 Nov. 2000,:
some limitations and is unable to address some of the http://www.cutter.com
peculiar issues associated with WBA development. We

449
IEEE --- 2005 International Conference on Emnerging Technologies
Septenmber 17-18, Islanmabad

[14] Software Engineering Institute official website: [35] Web Ontology Language (OWL):
http://www.sei.cmu.edu/tsp/tsp-security.html http://Www.w3.org/2004/OWL/
[15] A Ginige, S Murugesan "Guest Editors' Introduction: [36] Rule Language Standardization Report from the W3C:
The Essence of Web Engineering-Managing the http://www.w3.org/2004/1 2/rules-ws/renort/
Diversity and Complexity of Web Application [37] Web Services Description Language (WSDL), W3C
Development," IEEE MultiMedia, vol 08, no.2, pp. 22- Working Draft 10 May 2005:
25, April-June 2001 http://www.w3.org/TR/2005/WD-wsdl2O-primer-
[16] Henderson-Sellers, B., Lowe, D., & Haire, B. (2002). 20050510/
OPEN Process Support for Web Development. Annals [38] Anita C & Chris N "Web Metrics Bibliography Version
of Software Engineering, 13, 163-201,2002 2" Educational Impact and Evaluation Standing
[17] B. Haire, B. Henderson-Sellers, and D. Lowe, Committee, National Science Digital Library Sept 2004
"Supporting web development in the open process: [39] ISO ISO/IEC 14143-3: 2003, Software Engineering-
additional tasks" International Computer Software and Functional Size Measurement-Part 3: Verification of
Applications Conference,2001, Chicago, USA Functional Size Measurement Methods, International
[18] Lowe, D and W. Hall, Hypermedia and the Web: An Organization for Standardization - ISO, Geneva. 2003.
Engineering Approach, Wiley, 1999. [40] SOFTWARE MEASUREMENT GUIDEBOOK June
[19] Gellersen, H., and Gaedke, An Object-Oriented Model 1995 by National Aeronautics anid Space
for the Web Application Development Process, IEEE Administration (NASA)
Internet Computing, January - Feb 1999 [41] Will Weidman "Website Metrics" WC Journal,
[20] D. Firesmith, B. Henderson-Sellers, The OPEN Process Weidman Consulting Inc, 2003,
Framework. An Introduction, Addison-Wesley, 2001 http://www.weidmanconsulting.com
[21] 1. Graham, B. Henderson-Sellers, and H. Younessi, The [42] Silvia Abrahao and Oscar Pastor, "Measuring the
OPEN Process Specification, Addison-Wesley, 1997 functional size of web applications", Int. J. of Web
[22] B. Henderson-Sellers, A. Simons, and H. Younessi, The Engineering and Technology, Vol. 1, No. 1, 2003
OPEN Toolbox of Techniques, Addison-Wesley, 1998 [43] COSMIC The Common Software Measurement
[23] D. Firesmith, G. Hendley, S. Krutsch, and M. Stowe, International Consortium http:/Hwww.cosmicon.coml
Object-Oriented Development Using OPEN: A [44] F Azam, Z Li, R Ahmad "'Towards higher-level
Complete Java Application, Addison-Wesley 1998 information aspects of Web systems: An MDA
[24] B. Henderson-Sellers and B. Unhelkar, OPEN Modeling Approach", ISSN: 0302-9743, pp 472-- 477, volume
with UML, Addison-Wesley, UK, 2000 3579 of Lecture Notes in Computer Science, Springer-
[25] OMG's Official website, Verlag, July 2005
http://www.omg.org/mda/specs.htm#MDAGuide [45] Tongrungrojana, R, Lowe, D: "'WIED: A Web
[26] Nathalie M, Antonio V "Using the MDAt for Modelling Language for Modelling Architectural-Level
Designing and Implementing Web-based Applications", Information Flows", Journal of Digital Information,
Tutorial, The 4th International Conference on Web Volume 5 Issue 2, Article No. 283, 2004-08-10
Engineering (ICWE 2004), Munich, Germany [46] Gu, A., B. Henderson-Sellers, et al, "Web Modelling
[27] Web Intelligence Consortium, http://wi-consortium.org/ Languages: The Gap between Requirements and
[28] Zhong, Ning; Liu, Jiming, "Intelligent Technologies for Current. Exemplars", AusWebO2, Gold Coast, Australia
Information Analysis", 2004, ISBN: 3-540-40677-8 [47] Marta Kwiatkowska and Vladimiro Sassone "'Science
[29] Web Intelligence and Agent Systems (WIAS), journal for global ubiquitous computing" Report on Grand
of Web Intelligence Consortium Challenges in Computing Research published by The
[30] J. Liu, N. Zhong, Y.Y. Yao, Z. Ras: "The Wisdom Web: British Computer Society 2004. p-9
New Challenges for Web Intelligence (WI)", Journal of [48] Kappel G, P Birgit, R Wemer, S Wieland:
Intelligent Infonnation Systems, 20 (1) 2003 "Customisation for Ubiquitous Web Applications - A
[31] Heterogeneous Agent Systems 1st edition (640 pages) Comparison of Approaches", International Journal of
V. S. Subrahmanian, Piero Bonatti, Jurgen Dix, Thomas Web Engineering and Technology (IJWET), Jan 2003
Eiter and Fatma Ozcan MIT Press, 2000. ISBN: 0-262- [49] Distributed Software Engineering Group Research into
19436-8 Ubiquitous Computing http://www-
[32] Constructing Intelligent Agents Using Java: dse.doc.ic.ac.uk/Research/ubicomp2.htmI
Professional Developer's Guide 2nd edition (432 pages)
Joseph P. Bigus, Jennifer Bigus John Wiley and Sons,
2001. ISBN: 0-471-39601-X
[33] Erich N, Claudia N, Michael F, Fraunhofer Institute
IPSI, Darmstadt Germany "Semantic Web Application
Models" Invited Talk at 22nd International Conference
on Conceptual Modeling, Chicago, Illinois Oct, 2003
[34] Resource Description Framework (RDF):
http://www.w3.oryz/RDF/

450