INTELLIGENT AUTHORING TOOLS FOR ENHANCING MASS

CUSTOMIZATION OF e-SERVICES

The smarTag Framework

Panagiotis Germanakos

Department of Management and MIS, University of Nicosia, 46 Makedonitissas Ave., P.O.Box 24005, 1700 Nicosia, Cyprus

Nikos Tsianos, Zacharias Lekkas, Costas Mourlas

Faculty of Communication and Media Studies,

National & Kapodistrian University of Athens, 5 Stadiou Str, GR 105-62, Athens, Hellas, Greece

Mario Belk, George Samaras

Computer Science Department, University of Cyprus, CY-1678 Nicosia, Cyprus

Keywords: Mass Customization, Adaptation, Personalization, eServices, Web Authoring tools, Cognitive Factors.

Abstract: Mass customization should be more than just configuring a specific component (hardware or software), but

should be seen as the co-design of an entire system, including services, experiences and human satisfaction

at the individual as well as at the community level. The main objective of this paper is to introduce a

framework for the automatic reconstruction of Web content based on human factors. Human factors and

users’ characteristics play the most important role during the entire design and implementation of the

framework which has the inherent ability to interact with its environment and the user and transparently

adapt its behaviour using intelligent techniques, reaching high levels of usability, user satisfaction,

effectiveness and quality of service presentation. The initial results of the evaluation have proven that the

proposed framework do not degrade the efficiency (in terms of speed and accuracy) during the Web content

adaptation process.

1 INTRODUCTION

Peoples’ lives today are more turbulent and

diversified. The “one size fits all” (Stonebraker M.,

Çetintemel U., 2008, Brown D., 2004) model could

be considered out-of-date. People now want to be

seen and treated as individuals and many are

prepared to pay for this. They are better educated

and informed; able and willing to make their own

decisions. Mass customization moves towards this

direction and it aims to replace mass production,

which is no longer suitable for today’s chaotic

markets, growing product variety, and opportunities

for eCommerce. Mass customization is a broad term

and could be characterized as a working and

profitable business model. There is a whole

spectrum of ways that mass customization

methodologies can benefit companies. At the most

visible end of the spectrum, companies can mass

customize products and services for individual

customers.

Regarding the convergence of Internet and mass

customization, there is still the dispute whether

could be proved successful, and perhaps more to the

point: is it actually happening? (cyLEDGE Media,

2008).

For the scope of our current research, we

perceive mass customization along with

personalization, as a combination that together tend

to change the business information systems offering

personalized service relationships as a way of

connecting with customers over a number of

platforms and of differentiating their services from

those of competitors. Mass customization should be

more than just configuring a specific component

Germanakos P., Tsianos N., Lekkas Z., Mourlas C., Belk M. and Samaras G. (2009).

INTELLIGENT AUTHORING TOOLS FOR ENHANCING MASS CUSTOMIZATION OF e-SERVICES - The smarTag Framework.

In Proceedings of the 11th International Conference on Enterprise Information Systems - Human-Computer Interaction, pages 91-96

DOI: 10.5220/0001997200910096

 SciTePress

(hardware or software), but should be seen as the co-

design of an entire system, including services,

experiences and human satisfaction at the individual

as well as at the community level.

The research that is described in this paper

focuses on incorporating theories of individual

differences in information processing within the

context of Web-sites. Since the WWW is by

definition a huge resource of information, it would

make much sense that individuals’ information

processing characteristics should be taken into

consideration. As part of our previous research, it

has been demonstrated that the incorporation of

related human factors in eLearning and eServices

environments leads to better comprehension on

behalf of the users (Germanakos et al., 2008a,

Tsianos et al. 2008). This has been achieved with the

reconstruction and enhancement of the quality of

information presentation and users’ interactions in

the Web with techniques and tools aligned with their

specific needs and preferences. To that direction, our

efforts are focused on improving the effectiveness of

Web-sites by employing methods of personalization.

In our more recent work, a framework for

achieving mass customization on the Web based on

human factors and the automatic reconstruction of

the Web content has been developed and evaluated,

called smarTag. The smarTag is an easy to use tool

that enables any Web designer and developer to

enhance their Web-site (technology and language

independent) with adaptive Web objects that adjust

according to the users’ cognitive factors.

The main objective of this paper is to present the

smarTag architecture and its components. Towards

this direction a short review of current Web

authoring tools as well as the description of the

proposed framework is outlined. Finally, an

evaluation of the smarTag System concludes the

paper with the initial results being really

encouraging for the future of our work.

2 WEB SERVICES TENDENCIES

BASED ON HUMAN FACTORS

An effective and comprehensive mass production

technique could be to devise a cognitive framework

that could assist providers to develop Web-sites that

will embrace intrinsic values of customers, tailoring

their services accordingly.

We have seen that cognitive factors have an

important role in user satisfaction in identifying the

products that are interested in (Germanakos et al.,

2008a, Tsianos et al. 2008). However, the way

cognitive factors used today in order to design and

develop Web services is considered unwisely and

principally based on provider’s perception, without

following particular rules that could achieve the

appropriate mapping with selected content meta-

characteristics; thus reconstructing any content to

the benefit of the users.

Consequently, our research interest is whether

we could develop a smart Web authoring framework

that will dynamically alter a section of a Web-site by

personalizing the content and the structure to

specific users’ cognitive preferences. This could be

achieved by enriching the existing Web structures

and raw content (provider’s original content) with

further design enhancements and specific content

transformations based on the adaptation mapping

rules derived from selected cognitive factors (i.e.,

show a more diagrammatical representation of the

content in case of an Imager user, as well as provide

the user with extra navigation support tools). In the

event that this would be proven successful and

meaningful, individuals would learn better the

information that is important to them.

Fig. 1 shows the possible Web content

transformations / enhancements based on the

mapping process that takes place during adaptation

process, the influence of human factors and the

theory of individual differences. Based on the figure

above, the meta-characteristics of a user profile are

deterministic (at most 3); Imager or Verbalizer,

Analyst or Wholist and Working Memory level

(considered only when low).

Figure 1: Web content transformations / enhancments.

To our knowledge, there has not been a Web

Development Editor that takes into consideration a

combination of the above issues for mass

customizing Web products and services.

2.1 Current Web-based Authoring

Tools

Web-based authoring tools are becoming standard

ICEIS 2009 - International Conference on Enterprise Information Systems

issue in modern content management systems. They

range from simple text editors to high powered

graphical authoring tools and content management

systems.

This section outlines some noteworthy research

oriented and commercial Web authoring tools. Many

of the editors listed below are “What You See Is

What You Get” (WYSIWYG) HTML editors, some

of which have the option to view the HTML source

code. These are quite popular due to the low

learning curve, yet it is important to get some

understanding of HTML since WYSIWYG HTML

editors can be limiting.

Such systems, mostly commercial, are amongst

others the EditOnPro (http://www.realobjects.com/),

Cute Editor (http://cutesoft.net/), TinyMCE Editor

(http://tinymce.moxiecode.com/) and JXHTMLEdit

(http://www.tecnick.com/).

Other, more research oriented Web authoring

tools that were inspected are: Protégé (Noy et al.,

2001), Swoop (Kalyanpur et al., 2005a, Kalyanpur

et al., 2005b), and Ontostudio (http://semanticweb

.org/wiki/OntoStudio).

3 THE smarTag FRAMEWORK

The smarTag Editor (see Fig. 2) is a Web

Development tool enabling a content provider to

create smart objects. It is composed of interrelated

components, each one representing a stand-alone

Web-based system.

A smart object under the smarTag Framework is

conceptually similar to the traditional XML objects:

they both consist of attributes and content. The

content can either be in a textual or diagrammatical

form in case of a Verbalizer and Imager user

respectively. The smarTag attributes are special

meta-characteristics (Germanakos et al., 2008b)

describing the possible behavior the object can

perform in its environment. All the objects are stored

on the smarTag Server which are used in the

mapping process of a user’s profile (Tsianos et al.,

2008), as well as the provider’s external Web-page.

Since all the smart objects will be embedded as

enhancements in an external Web-site, our main

concern is to ensure openness and interoperability

between the system’s components and any external

Web-site, as well as to ensure the Web security

policies. In order to achieve this, the smart objects

must be easily extendible and easy to handle. Using

XML for implementing the smart objects’ structure

seems to be the best way to achieve this. Indeed

XML (http://www.w3.org/XML/) enables the

extendibility we need and enhances interoperability

and integration among systems’ components. We

have designed a Web Service (a software system

designed to support interoperable Machine to

Machine interaction over a network) for retrieving

the smart objects. A more comprehensive

description on this concept will take place in section

3.2.

Figure 2: The smarTag Framework.

3.1 Enhancing any Web-site with

smarTag Editor

Our main concern was to create an easy to use Web

enhancement tool that enables any Web developer /

designer to enrich divisions of his / her Web-site

with mass customization and personalization

techniques. More specifically, the traditional

methods of Web Development will take place in the

process; based on the main requirements of the end-

users of the Web-site and mainly on the “design

taste” of the Web Designer / Developer.

Based on the Traditional Web-site development

process, depending on the Web-site requirements

and specifications, all the needed information (text,

data, graphics and pictures) of the Web-site is

collected. The Web-site’s layout and Navigation is

then designed by the Web Designer and all the

collected information is implemented in the Web-

site.

Predominantly, smarTag Editor is used in the

Web Programming and Customization phase. The

Web Developer will define specific divisions in the

Web-site that will adapt according to individual

characteristics (cognitive styles).

For a better understanding of how smarTag

Editor works in practice, Fig. 3 shows a quick step

process diagram for enhancing a Web-site with

smart objects. An authorized Web Developer will

create a new adaptive Web object by storing the

object’s actual content (text or image) and will

INTELLIGENT AUTHORING TOOLS FOR ENHANCING MASS CUSTOMIZATION OF e-SERVICES - The smarTag

Framework

characterize it based on the smarTag framework. A

unique identifier will be assigned to this object and

stored on the smarTag server. Based on the unique

identifier, the Web Developer will map the

corresponding object with a specific division in the

Web-site created so far. The smarTag editor will

then generate a JavaScript file based on the

provider’s preferences and will be embedded in the

Web-site.

Figure 3: Enhancing Web-sites using smarTag Editor.

This JavaScript file is the core element for

communication establishment between the smarTag

Web Service (description follows in section 3.2) and

the provider’s external Web-site.

3.2 Adaptation and Mapping Process

To get a better insight of the adaptation process and

how data flows, we hereafter discuss how the

personalized content interacts with the user profile,

using specific mapping rules under the smarTag

framework. For this purpose, we have designed an

experimental setting in the application field of

eCommerce, by authoring smart Web objects and

enhancing an existing commercial Web-site. The

eCommerce (Web) environment that has been

developed used the design and information content

of an existing commercial Web-site of IBM

(http://www.ibm.com, date extracted November 25,

2008). This Web-site provides products’

specifications of the IBM Company. We have

developed an exact replica of the IBM System

Servers section in IBM.com using smart objects.

Initially, the user’s comprehensive profile is

retrieved. In this particular example, the user

happens to be an Imager / Analyst with regards to

the Cognitive Style, has an average knowledge on

the subject (computer knowledge) based on his

traditional characteristics, and he has a low Working

Memory Span (weighting 2/7).

Figure 4: Content adaptation according to user’s

comprehensive profile.

Furthermore, every Web-page is detached into

standalone objects, each one having special

characteristics. In our example, the user visits the

“WebPage_Y” Web-page. First, the main HTML

document of this Web-page is retrieved which

contains all the needed information for building the

Web-page; that is, (i) the HTML Web-page itself

which is a predefined HTML document (designed by

the provider) keeping information of specified

divisions / frames in the page for positioning each

object, (ii) all objects (text, image, audio, video etc.)

that comprise the content of the Web-page, and (iii)

a generated JavaScript file from smarTag that is

responsible for the proper integration of the smart

objects into the divisions’ Web-page.

Based on the abovementioned information

(user’s comprehensive profile and the content

description of the particular Web-page) the

correlation rules are created. These rules are

responsible for mapping the particular implications

with the Web-page’s content, and consequently

assembling the final adapted version of the

provider’s content.

The interpretation of this data-implications

correlation results in the following conclusions

ICEIS 2009 - International Conference on Enterprise Information Systems

while the user interacts with the eCommerce

environment: (a) the provision of visual information

(diagrammatical representation since he is an

Imager), (b) extra navigation support tools (menu

tabs) since he is an analyst, and (c) a “myNotepad”

tool is used; temporary memory buffer for storing

sections’ summaries due to his/her low working

memory span (see Fig. 4b).

3.3 Viewing the Adapted Content

The AdaptiveWeb User Interface (Germanakos et

al., 2008a, Germanakos et al., 2007), namely

AdaptiveInteliWeb (see Fig. 4a,b,c), is a Web

application used for displaying the raw and/or

personalized and adapted content on the user’s

device. This can be a home desktop, laptop or a

mobile device.

The main concept of this component is to

provide a framework where all personalized Web-

sites can be navigated. Using this interface a user

interacts with the provider’s content and based on

his profile and cumulative characteristics the

adjusted content and particular supportive

navigational tools are displayed. Fig. 4a depicts an

exact replica of the IBM Web-site without any

personalization made, while Fig. 4b and Fig. 4c

shows the same Web-site after the personalization

and adaptation process has been initiated, with the

content to be adapted according to the user’s

comprehensive profile. As we can easily observe,

the original environment has been altered based on

rules that define the typologies of the users in terms

of content reconstruction and supportive tools.

For example, in case a user is identified as

“Wholist-Verbalizer” the content will be

automatically reconstructed as in Fig. 4c, where a

floating menu with anchors (Wholist) have been

added so to guide the users on specific parts into the

content while interacting. In this case no

diagrammatical presentation will be used because

the user is a Verbalizer.

4 TECHNICAL EVALUATION OF

smarTag - SYSTEM’S

PERFORMANCE

To measure the performance, functional behavior

and efficiency of our system we have run two

different simulations with 100 threads (users) each:

(a) users retrieving raw content without any

personalization and adaptation taking place and (b)

users interacting with adapted and personalized

content. In the second scenario, there is a significant

increase of functions and modules ran, compared to

the first one (raw content scenario), like user profile

retrieval, dynamic content adaptation, learner control

dynamic tools, navigational support etc.

Based on the simulations made (see Fig. 5) we

assume the following:

(i) Deviation for raw content is 67ms and for

personalized content 98ms. This difference is

expected since the system uses more functional

components in the case of personalized content like

profile loading, dynamic content, etc. Thus, this

consumes more network resources causing the

deviation of our average to be greater than that of

the raw content test. The deviation is not considered

to be significantly greater and thus this metric result

is proving the system to be stable and efficient;

Figure 5: Raw Content and Adapted Content Scenarios.

(ii) the throughput for the raw content scenario

was 144Kb/min while for the personalized content

was 179Kb/min. Based on the latter results, the

system is again considered efficient mainly due to

the fact that the difference in throughput between the

two scenarios is minimal. Taking in consideration

that major component functionality is used in the

case of personalized content this small difference

suggests the efficiency of the system; (iii) the same

arguments are true in the case of the average

response times. The average response time for the

raw content scenario was 150ms while for the

personalized content was 177ms. This difference is

again marginal that proves the efficiency of the

system.

5 CONCLUSIONS

The basic objective of this paper was to present a

framework, namely smarTag, for the automatic

INTELLIGENT AUTHORING TOOLS FOR ENHANCING MASS CUSTOMIZATION OF e-SERVICES - The smarTag

Framework

reconstruction of any Web content based on human

factors for providing a comprehensive personalized

result. This approach is liable of enhancing

efficiency and effectiveness of users’ interaction

with eServices in terms of information assimilation

and accuracy of finding their cognitive targets

(products or services). Based on previous findings, it

has been proven that user’s cognitive factors have an

important impact in the information space and on

specific content meta-characteristics. Accordingly,

the smarTag system provides an easy to use

framework for enhancing any Web-site with smart

objects that take into consideration human factors

for the adaptation of the content. The initial results

of the system’s evaluation have shown that the

proposed framework do not degrade the efficiency

(in terms of speed and accuracy) in the Web content

adaptation process and could be efficiently used for

targeting the mass market by encapsulating

customers’ distinct characteristics. Such a method

could be considered nowadays fundamental for the

provision of adapted and personalized eServices, via

any medium, increasing this way one-to-one service

delivery and integrity, enabling businesses to retain

their customers and therefore to gain a substantial

competitive advantage.

ACKNOWLEDGEMENTS

The project is co-funded by the Cyprus Research

Foundation under the project EKPAIDEION

(#ΠΛΗΡΟ/0506/17) and the EU project CONET

(INFSO-ICT-224053).

REFERENCES

Blom J., 2000., Personalization – A Taxonomy, ACM

2000. ISBN:1-58113-248-4., 2000

Brown D., 2004. Wave goodbye to the era of one-size-

fitsall. http://seattlepi.nwsource.com/

cyLEDGE Media, 2008. Disentangling Web 2 and Mass

Customization

De Bra P., Aroyo L., & Chepegin V., 2004. The next big

thing: Adaptive Web-based systems, Journal of Digital

Information, Vol. 5 Issue 1, Article no. 247.

Germanakos P., Tsianos N., Lekkas Z., Mourlas C., Belk

M., & Samaras G., 2007. A Semantic Approach of an

Adaptive and Personalized Web-based Learning

Content - The case of AdaptiveWeb, Proceedings of

the 2nd International Workshop on Semantic Media

Adaptation and Personalization (SMAP 2007),

London, U.K., December 17-18, IEEE Computer

Society, pp. 68-73.

Germanakos, P., Tsianos, N., Lekkas, Z., Mourlas, C.,

andSamaras, G., 2008a. Realizing Comprehensive

User Profiling as the Core Element of Adaptive and

Personalized Communication Environments and

Systems, The Computer Journal, Special Issue on

Profiling Expertise and Behaviour, Oxford University

Press, 2008, doi:10.1016/j.chb.2007.07.010.

Germanakos P., Tsianos N., Lekkas Z., Mourlas C., Belk

M., & Samaras G., 2008b. Towards an Adaptive and

Personalized Web Interaction using Human Factors, A

chapter to appear in: Advances in Semantic Media

Adaptation and Personalization, Vol.2. M. Angelides.

Kalyanpur A., Parsia B., & Hendler J., 2005a. A Tool

forWorking with Web Ontologies, In Proceedings of

the International Journal on Semantic Web and

Information Systems, Vol.1, No.1.

Kalyanpur A., Parsia B., Sirin E., Cuenca-Grau B.,

Hendler J., 2005b. Swoop: A Web Ontology Editing

Browser, Journal of Web Semantics Vol 4(2).

Noy N., Sintek M., Decker S., Crubézy M., Fergerson R.,

& Musen M., 2001. Creating Semantic Web Contents

with Protégé-2000, Stanford University

Stonebraker M., Çetintemel U., 2008. One Size Fits All:

An Idea Whose Time Has Come and Gone

Tsianos N., Lekkas Z., Germanakos P., Mourlas C., and

Samaras G., 2008. User-centered Profiling on the

basis of Cognitive and Emotional Characteristics: An

Empirical Study. Proceedigns of the 5th International

Conference on Adaptive Hypermedia and Adaptive

Web-based Systems (AH 2008), Hannover, Germany,

July 28 - August 1, 2008, Springer-Verlag Berlin

Heidelberg, LNCS 5149, pp. 214-223.

Wang J., & Lin J., 2002. Are personalization systems

really personal? – Effects of conformity in reducing

information overload, Proceedings of the 36th Hawaii

International Conference on Systems Sciences

(HICSS’03), 0-7695- 1874-5/03.

ICEIS 2009 - International Conference on Enterprise Information Systems