supporting personalization and customization in a collaborative setting

Supporting personalization and customizationin a collaborative setting$

David Hicks*

Department of Computer Science, Aalborg University Esbjerg, Niels Bohrs Vej 8, 6700 Esbjerg, Denmark

Abstract

Knowledge workers today face an information world with challenging characteristics. Supporting the personalization and

customization process is an important tool in helping them to cope with these challenges. It can be an important tool in helping

them to find, access, and work with the information they require to perform their tasks. This is especially true in a collaborative

setting. This paper describes a customization architecture specifically designed for knowledge workers. It examines the current

status of the architecture and a prototype system based upon it, with an emphasis on the support it provides for collaboration.

Ongoing and future research for the project is also described, to examine how the architecture is being expanded to increase

support for additional modes of collaboration.

# 2003 Elsevier B.V. All rights reserved.

Keywords: Personalization; Collaboration; Customization; Digital library

1. Introduction

Knowledge workers today find themselves in an

information world that is increasingly diverse, dis-

tributed, and fragmented. As the amount of informa-

tion available in digital form continues to increase, so

does the number of systems that are employed to make

that information available. Examples include commer-

cial web sites, private corporate information systems,

and official government databases and archives. The

result is that those working in information intensive

occupations must routinely interact with a range of

different systems in order to find and access the

information they need in performing their tasks. Infor-

mation items managed by remote systems distributed

across networks, as well as those available from local

sources, are becoming an increasingly important part

of the overall information needs of a user.

For example, an individual preparing an environ-

mental regulation compliance report for a company

might need to interact with several systems to gather

the required information. These might include a geo-

graphic information server at one site where the

individual can obtain maps and other geographic data,

a governmental database located at another site that

contains the relevant environmental regulations, and a

number of corporate sites to gather information con-

cerning the company’s recent compliance with envir-

onmental regulations.

The ability of knowledge workers to customize their

information space, to personalize the information

items with which they work, is an important capabil-

ity, one that is valuable in the performance of complex

tasks [2]. The ability to make customizations to an

information space can be especially important in

Computers in Industry 52 (2003) 71–79

$ This paper is an expanded version of a paper that appeared

in [1].* Tel.: þ45-791-27623; fax: þ45-754-53643.

E-mail address: [email protected] (D. Hicks).

0166-3615/$ – see front matter # 2003 Elsevier B.V. All rights reserved.

doi:10.1016/S0166-3615(03)00070-8

collaborative work [1]. For example, the user

described above might be working with a group of

collaborators to prepare the environmental compli-

ance report. It might be convenient for the user to

attach annotations to the environmental regulations or

other artifacts being used in the preparation of the

report, in order to convey observations to others

collaborating on the project. In this way, the user

would be able to share knowledge, at an informal

level, about the artifact or the process of preparing the

report.

The importance of user customization and perso-

nalization capabilities has often been noted in the

literature. Nurnberg et al. [3] point out the need to

allow easy personalization of the information accessed

by web client applications. They also point out that the

new digital processes that will characterize future

information systems such as agents, and user profiling

will likely require even greater personalization and

customization functionality. Marshall [4] notes the

importance of supporting personal annotation func-

tionality in digital libraries. According to Roescheisen

et al. [5], value is added to an information space

through the process of personalization.

Though the ability to customize an information

space is important, the characteristics of the informa-

tion world of today as outlined above make it a

difficult capability to provide. This paper examines

the information customization issue. It does so with a

particular emphasis on how customization can be

supported within collaborative work environments.

In Section 2 an architecture for supporting the custo-

mization process is described. The description con-

tinues in Section 3 with an examination of the support

provided in the architecture for supporting collabora-

tion. A collaborative work scenario is used to help

describe the capabilities of the architecture. The next

section provides a brief look at a prototype imple-

mentation based upon the architecture. Next, a look at

ongoing and future research for the architecture pro-

vides a look at how it is being enhanced to provide

support for additional modes of collaboration.

2. A customization architecture

This section describes the personal adaptable digital

library environment (PADDLE) customization archi-

tecture. The PADDLE architecture was designed to

create a personalization environment for knowledge

workers, especially those with diverse and distributed

information needs. As detailed in the description

below, many of the capabilities of the architecture

are useful in a collaborative setting.

The term personalization sometimes refers to the

customization or modification of a digital library

system, rather than the customization of information

objects. For example, user profiles are often used in

order to allow a digital library system to be configured

or modified to adapt to the needs of a specific user [6].

Adapting systems to user needs is an important cap-

ability. However, in the PADDLE project, the empha-

sis has instead been placed on providing the ability for

a user or group of users to customize or modify the

information objects with which they work. The goal is

to enable users to personalize information in a way

that reflects personal or group preferences, and facil-

itates their ability to perform knowledge related tasks.

As described earlier, the information world of today

is an increasingly distributed and heterogeneous one.

This often requires knowledge workers to interact with

a variety of different systems in order to obtain the

information they require. An important goal of the

PADDLE architecture is to support personalization

for all of the information objects with which knowl-

edge workers interact, regardless of where the infor-

mation is stored or by which system it is managed or

controlled. This goal has significantly shaped the

design of the architecture and lead to two of the

primary characteristics of its approach for supporting

personalization: that it is decentralized and that it is

metadata based.

The approach is decentralized in that the information

(as well as the functionality) required to create and

represent personalizations for users is not centrally

stored within information repositories. Network based

information repositories can have a large if not unlim-

ited user base. A strategy that centralizes personaliza-

tion functionality at the information repository would

be a difficult, if not impossible, one to realize. This is

reflected by the relatively small number of network

based information systems that support individual per-

sonalization capabilities [12]. Few of the developers of

these systems have the resources or incentive to provide

a centralized personalization capability for their users.

The PADDLE architecture instead uses an approach

72 D. Hicks / Computers in Industry 52 (2003) 71–79

that captures personalization information locally (with

respect to the user) as users interact with and perso-

nalize information items and then maintains this infor-

mation in a decentralized way.

The PADDLE approach is metadata based in that

metadata serves as the mechanism for capturing and

maintaining personalizations that are made to infor-

mation items. In its most basic form, metadata is

simply data about data. The most common use of

metadata is as a mechanism for describing information

resources. For example, the metadata descriptions

contained in digital catalogue systems describe infor-

mation resources in a way that enables users to

determine if a particular resource is likely to be

relevant for their task at hand. When metadata is used

this way, the descriptions encoded within it need to be

general enough to be appropriate for the variety of

users that will consult the digital catalogue system [8].

The role of metadata in the PADDLE architecture is

a somewhat unconventional one [7]. Instead of being

used to describe information resources in a general

way, such as the descriptions contained in a digital

catalogue, metadata is used at a much finer level of

granularity. It serves as the basis for creating custo-

mized descriptions (or personalizations) of informa-

tion items. These customized descriptions can be used

to support a range of different forms of personaliza-

tions. For example, a user working with an environ-

mental regulation document might wish to create a

comment to record their observations about its rele-

vance and how useful it was for a specific task.

Alternatively, a user working with a satellite image

object might wish to customize it by highlighting a

certain portion of the image to indicate it is relevant for

a particular task. The PADDLE approach employs

personalized metadata to capture and preserve perso-

nalizations such as these.

An overview of the PADDLE architecture is illu-

strated in Fig. 1. The shaded part of the figure repre-

sents a user’s local computing environment. Client

applications in the figure correspond to the tools that

are used by knowledge workers to access information.

Example client applications include a web browser, a

database front end, a graphical image viewer, or any

tool used for information access. The information

resources shown in Fig. 1 represent the information

objects used by knowledge workers as they perform

their tasks. Examples include documents, images,

database records, or any information object with

which knowledge workers interact. As indicated in

the figure, information resources in the architecture

are not restricted to the local computing environment.

They can be located at any place in a knowledge

worker’s networked information world.

Fig. 1. The PADDLE customization architecture.

D. Hicks / Computers in Industry 52 (2003) 71–79 73

The primary functional component of the architec-

ture is the customization metadata manager (CMDM).

As illustrated in Fig. 1, the CMDM is positioned

between the client applications and the information

they access. It is a server process that performs a range

of functions in response to client application requests.

The most important functionality provided by the

CMDM is the creation and maintenance of metadata

to capture and preserve personalizations made to

information items.

Also shown in Fig. 1 is the customization metadata

store. This facility provides persistence for per-

sonalizations that have been defined for information

items. This enables personalizations to persist across

user sessions with the environment. Personalizations

recorded within the customization metadata store are

automatically applied to information items as they

are accessed by client applications within the per-

sonalization environment. Note, the information

items themselves are not stored in the customization

metadata store, it only contains personalizations.

The actual information items continue to reside in

the information systems where they were originally

located.

An example is helpful to demonstrate the interac-

tions between the various components of the PADDLE

architecture. Consider an image browsing tool being

used by a knowledge worker to access a remotely

located satellite image. If integrated into the PADDLE

environment, the image browser tool could be a client

application in the arrangement shown in Fig. 1. The

system where the satellite image is actually located

would correspond to an information resource in Fig. 1

that is being accessed remotely. In order to access an

image, the browser tool can issue a request for the

CMDM to retrieve the image. The CMDM would

contact the appropriate remote information system

to retrieve the image, and then check its customization

metadata store to determine if any personalizations

have been defined for the image. If no personalizations

have been defined for it, the image would simply be

passed along directly to the browser for display to the

user. If personalizations have been defined for the

image, the CMDM would apply them before passing

the image along to the browser.

Note that the interactions between client applica-

tions that have been integrated into the environment,

such as the image browsing tool, and the components

of the customization architecture, such as the CMDM,

are transparent to the user. For example, customiza-

tions defined for objects like the satellite image are

automatically applied to the objects as they are

accessed, without the requirement for user interaction.

In this way, the object has been effectively customized

or personalized for the user within the PADDLE

environment.

While examining and working with the satellite

image, a user might decide to somehow personalize

it, such as by adding an annotation, or perhaps chan-

ging an existing one. The browser tool could support

such personalizations by requesting the CMDM to

create customization metadata records to capture it.

The records are stored in the customization metadata

store and will be automatically applied the next time

the image is accessed within this particular context.

3. Customization in a collaborativeenvironment

As has been noted in the literature, individualized

customization (or personalization) of an information

space is an important capability, one that can signifi-

cantly enhance the ability of an individual in perform-

ing knowledge related tasks [3,5]. However, many

knowledge workers today spend a significant portion

of their time working in collaboration with others [9].

Just as with the individual work case, customization

can play an important role in helping users who work

in collaborative settings [10]. These observations have

had a significant influence on the PADDLE architec-

ture. Specifically, they motivated the design goal for

the architecture to be capable of supporting multiple

users, working individually, collaboratively, or in

some combination of these modes, possibly needing

to switch between modes.

3.1. Contexts

In the PADDLE architecture, the customization

metadata store is structured into contexts, which are

collections of related customizations. Contexts pro-

vide a mechanism to logical partition and group

customizations in the customization metadata store

according to individual users, user groups, or some

other relevant criterion. When working individually,


each user in the customization environment can define

customizations within their own private context, pre-

venting the personalizations made by one user from

overlapping or interfering with those made by others.

When necessary, a user or group of users can define

more than one context, in order to organize persona-

lizations according to the multiple tasks being worked

upon, or by some other criterion.

Contexts are a useful organizational tool for custo-

mizations. In addition to serving as an organizational

tool for customizations made by individuals, contexts

provide a mechanism to flexibly define customization

spaces to support the various collaborative activities

that occur within an organization. Contexts can be

particularly helpful in the sharing of informal knowl-

edge within a group as it develops during the course of

performing a collaborative task.

A collaborative work scenario can help to demon-

strate their usefulness. Consider the example from

Section 1, in which a group of users are working to

prepare an environmental regulation compliance

report for a company. A context could be defined to

organize the customizations made by group members

in the process of preparing the report. If the report was

a large one, multiple contexts could be defined to

support the division of the task into subgroups.

Regardless of how the task is structured, customi-

zations made to information objects during the process

of preparing the report could be placed into the

corresponding group context. Users working within

the context would automatically be presented with the

group-defined customized view when accessing

images, reports, or other artifacts.

Example of group customizations include an anno-

tation that is created by one of the group members and

attached to a government document describing an

environmental regulation. The annotation might be

used to communicate specialized knowledge that the

group member has, and allow that expertise to be

shared with the rest of the group. Another user might

decide to customize a group of information items, such

as a collection of satellite images, by adding a meta-

data field named ‘‘relevance’’ to the images. This new

metadata field might be used to rank the usefulness of

each of the images for the preparation of the report.

Finally, another user might wish to use a graphics

editor that has been integrated into the customization

environment to highlight a portion of a map document,

marking an area that is of particular importance for the

preparation of the report. Each of these customiza-

tions, if performed within the group context, would be

available to and automatically seen by each of the

group members when they are working within the

group context. In this way, the group context mechan-

ism helps individual users to share knowledge they

posses or have developed during the collaboration

process.

Many knowledge workers are involved with a

number of different projects. On some projects they

may work individually, on others collaboratively. To

support this common work practice, the PADDLE

architecture allows a user to define more than one

personal context when necessary. This enables them to

organize their customizations in a task oriented way,

perhaps defining a separate context for each of the

projects on which they might be working. Similarly,

users can be members of multiple group contexts, to

support their ability to participate on more than one

collaborative project. As users switch between tasks,

either individual or collaborative ones, they can switch

customization contexts accordingly.

3.2. Layered contexts

Another important consideration for customization

in a collaborative setting is to provide mechanisms that

can support the structure of the organization within

which knowledge workers perform their tasks. In the

PADDLE architecture, this is facilitated by the ability

to hierarchically relate contexts to each other. As

illustrated in Fig. 2, this provides a layering mechan-

ism for the customizations made within the environ-

ment [11]. This enables separately defined but related

changes to an information space to be combined.

When arranged in this way, multiple levels or scopes

of customization can be supported.

For example, an organization may wish to define a

corporate wide context that contains a set of custo-

mizations for information items that should be seen by

all users within the organization. An example of such

customizations might be a set of annotations marking

the sections of a collection of environmental regula-

tions that are most relevant for the company as a

whole. A particular department within the organiza-

tion might have departmental customization needs.

These could be organized into a departmental context.


An example of such customizations might be an

additional set of annotations that go beyond those

defined at the corporate context level to identify

sections of the environmental regulations relevant to

the specific tasks of a particular department within the

company. Finally, an individual member of the depart-

ment might wish to further personalize information

items through the creation of a private context contain-

ing a set of personal annotations for the environmental

regulations.

These contexts could be related hierarchically so

that when a user accesses an information item, such as

object B in Fig. 2 (which might correspond to an

environmental regulation document), any personaliza-

tions defined for it in the corporate context are first

applied, next any defined within the departmental

context are applied, and finally those from the indi-

vidual context are applied. Note that when object A is

accessed in the scenario depicted in Fig. 2, the corpo-

rate view of the object would be presented to the user

since no departmental or individual level personaliza-

tions have been defined for it.

4. Prototype implementation

A prototype personalization environment has been

constructed based on the PADDLE architecture. The

two main elements of the architecture, the customiza-

tion metadata manager and the customization meta-

data store, have been implemented as individual

software components. The CMDM has been imple-

mented in Java and is based upon the Netscape Fast-

track Web server. A Microsoft Access database

currently provides the functionality of the customiza-

tion metadata store. The software components com-

municate using a range of standard protocols.

Communication between the CMDM and the meta-

data store takes place using Java RMI, to facilitate

distribution. Communication between the CMDM and

external or remote information systems is flexibly

defined using abstract Java classes so that a range

of different protocols can be accommodated. Further

details of the base implementation can be found in [7].

A client application has been implemented and

integrated into the personalization environment that

enables users to access information objects from

remote sources. The client application interacts with

the CMDM to enable users to view information

objects as well as perform customizations on those

objects. Currently, three different information systems

have been integrated into the prototype environment,

each of which contains information from professional

content providers. The first information system con-

tains a collection of over 2000 Microsoft Office

documents, the second one consists of over 100,000

HTML documents, and the third one is the electronic

theses archive of a German University. Each of these

information systems provides metadata descriptions

of the resources they contain.

The prototype environment currently supports the

personalization of these metadata descriptions by

users of the client application. The types of persona-

lizations permitted on the metadata descriptions

Fig. 2. Personalization contexts related hierarchically.


include: the ability to change the value of a metadata

field, the ability to hide or delete a metadata field, and

the ability to define a new metadata field and specify a

value for it.

Fig. 3 illustrates the personalization component of

the client application that enables the user to perso-

nalize metadata descriptions of information resources.

As an example personalization, a user might decide

to change the value of a field to something more

meaningful for them in order for the corresponding

record to be located easier in the future. In Fig. 3,

the ‘‘Semantic Relationship’’ field of the displayed

resource description has been updated for that pur-

pose. Alternatively, a user might wish to organize a

subset of the records of the database according to some

new dimension. This could be done by creating a new

field for the resource descriptions and assigning appro-

priate values. In Fig. 3, the ‘‘Temporal Relationship’’

field has been added for this purpose. Using this new

field, the information items examined by the user can

be classified according to the time period to which

they pertain.

As illustrated in Fig. 1, the customization environ-

ment was designed to enable the integration of almost

any application used by knowledge workers to access

information objects. Although to date only the initial

client application described above has been integra-

ted into the environment, additional integrations are

planned.

5. Discussion and future research

The initial results and use of the PADDLE custo-

mization environment are encouraging. Based on user

experiences and feedback, the functionality it provides

Fig. 3. Personalization component of client application.


appears to be effective for supporting the personaliza-

tion and customization of information objects within

the prototype environment.

However, a major goal for ongoing and future

research on the project is to increase the capabilities

of the system for supporting collaboration. Much of

the usage of the system to date has been by individuals

working alone. Though the system does provide basic

facilities for supporting groups of users, especially

those working asynchronously, the support offered

needs to be expanded to accommodate additional

modes of collaboration.

Expansion and refinement of the context mechan-

ism will be an important part of increasing the support

offered for collaboration. As described above, the

prototype system enables contexts to be used in a

range of ways to help people working together. For

example, relating contexts together hierarchically has

been an effective way in which to support organiza-

tional structures, as described earlier (Section 3.2).

However, the use of contexts to directly support

collaborative customization is more limited. Specifi-

cally it is limited primarily to supporting asynchro-

nous collaboration. The primary limitation has been

the granularity of writing for contexts. The current

prototype system supports only a basic access control

facility for contexts. It enables a context to be defined

for and shared by a group, but for only one person to

have update access for the context at any specific time.

When a context has been accessed in update mode by

one user, others can only access the context in read

only mode. While this strategy can accommodate

users working asynchronously, perhaps in a distribu-

ted setting, it is obviously insufficient for those who

need to work together more closely in a synchronous

collaborative setting.

Improving update access for contexts will therefore

be an important area of future research for the project.

Important will be the determination of the appropriate

level of granularity required to support synchronous

collaborative customization. The current level of

allowing only entire contexts to be updated is too

coarse as it can only accommodate asynchronous

operations. Alternative granularities include the infor-

mation object level, as well as the information object

attribute level. A strategy that enables adjustment to

the update access granularity for contexts might also

be appropriate, enabling a group of users to adjust

update levels as they move from one form of colla-

boration (asynchronous) to another (synchronous).

Another important area being examined is expand-

ing support within the architecture for the customiza-

tion of additional types of knowledge objects. As

described earlier, the current prototype has been inte-

grated with three information systems, and supports

the customization of the descriptions provided for the

information items they contain. In practice, knowl-

edge workers routinely interact with a wide variety of

different types of information objects. Establishing a

streamlined way in which to integrate new information

systems and support the customization of the informa-

tion objects they provide will be an important area to

be examined.

An important additional area to be examined

focuses on the customization metadata store (Fig. 1).

As described earlier, it provides persistence for the

customizations defined by individuals or group of

users. This represents a significant potential resource

that might be capable of being used to assist knowl-

edge workers within the environment. The informa-

tion it contains might be able to be mined or processed

somehow to develop user profiles, group profiles, or

inform agents, recommender systems or other auto-

mated services that can perform helpful tasks for

collaborating groups.

6. Conclusion

Support for customization is an important tool in

helping knowledge workers to cope with the challen-

ging information world of today. It can be an impor-

tant way to help them more effectively work with the

distributed and diverse types of information that they

increasing need in performing their tasks. This is

especially true in a collaborative setting, where sup-

port for customization can be especially helpful in

sharing knowledge at an informal level.

The PADDLE architecture was especially designed

to accommodate knowledge workers and the challen-

ging environment in which they must work. Its decen-

tralized and metadata-based approach for supporting

customization enables it to support a variety of types

of distributed information objects. The current proto-

type of the PADDLE architecture has been integrated

with three information systems. Though the initial


experiences with use of the system are encouraging,

additional capabilities are needed to enable the cus-

tomization support it provides to be expanded to

accommodate additional modes of collaboration. In

particular, supporting additional types of synchronous

collaboration is an important area of ongoing research

for the project.

Acknowledgements

This research was conducted as part of the PAD-

DLE project funded by the program ‘‘Verteilte Ver-

arbeitung und Vermittlung Digitaler Dokumente’’ of

the Deutsche Forshungsgemeinschaft (DFG).

References

[1] W. Shen, J.M. de Souza, Z. Lin, J.P. Barthes (Eds.), in:

Proceedings of the 7th International Conference on Computer

Supported Cooperative Work in Design, Rio de Janeiro,

Brazil, 2002.

[2] N.A. Van House, User needs assessment and evaluation for

the UC Berkeley electronic environmental library project: a

preliminary report, in: Proceedings of ACM Digital Li-

braries’95, Austin, TX, USA, 1995, pp. 71–76.

[3] P.J. Nurnberg, R. Furuta, J.J. Leggett, C.C. Marshall, F.M.

Shipman, Digital libraries: issues and architectures, in:

Proceedings of ACM Digital Libraries ’95, Austin, TX,

USA, 1995, pp. 147–153.

[4] C. Marshall, Annotation: from paper books to the digital

library, in: Proceedings of ACM Digital Libraries’97,

Philadelphia, PA, USA, 1997, pp. 131–140.

[5] M. Roescheisen, T. Winograd, A. Paepcke, Content ratings

and other third-party value-added information—defining an

enabling platform, D-Lib Magazine, August 1995.

[6] A. Smeaton, J. Callan, in: Proceedings of the Joint DELOS-

NSF Workshop on Personalisation and Recommender Sys-

tems in Digital Libraries, Dublin, Ireland, 2001.

[7] K. Tochtermann, D. Hicks, A. Kussmaul, Creating corporate

knowledge with the PADDLE system, in: Proceedings of

WebNet 2000—World Conference on the WWW and Internet,

San Antonio, TX, USA, 2000.

[8] D. Hicks, K. Tochtermann, Towards support for personaliza-

tion in distributed digital library settings, in: Proceedings of the

Joint DELOS/NSF Workshop on Personalisation and Recom-

mender Systems in Digital Libraries, Dublin, Ireland, 2001.

[9] D.L. Hicks, U.K. Wiil, 2001, Towards collaborative design in

an open hypermedia environment, in: Proceedings of the

Sixth International Conference on Computer Supported

Cooperative Work in Design, London, Ontario, Canada,

pp. 181–189.

[10] A. Frank, B. Mitschang, On sharing of objects in concurrent

design, in: Proceedings of the Sixth International Conference

on Computer Supported Cooperative Work in Design,

London, Ontario, Canada, 2001, pp. 71–76.

[11] V. Prevelakis, Versioning issues for hypertext systems, in: D.

Tsichritzis (Ed.), Object Management, Centre Universitaire

d’Informatique, Universite de Geneva, Switzerland, 1990,

pp. 89–105.

[12] K. Tochtermann, D. Hicks, A. Kussmaul, Support for

customization and personalization on the web, in: Proceed-

ings of WebNet 1999—World Conference on the WWW and

Internet, Hawaii, USA, 1999.

David Hicks is an Associate Professor

in the Department of Computer Science

at Aalborg University’s Esbjerg campus

in Denmark. He holds a PhD in

computer science from Texas A&M

University. He worked for 2 years as a

visiting researcher at the GMD-IPSI

Research Institute in Darmstadt, Ger-

many focusing mainly on the investiga-

tion of hypermedia infrastructure issues.

He was then employed with Knowledge

Systems in the US as a senior research scientist. Later he joined the

FAW Research Institute in Ulm, Germany as a research associate

focusing primarily on the investigation of digital libraries, before

joining the faculty at Aalborg University Esbjerg. His current

research interests include hypertext/hypermedia, digital libraries,

computer supported cooperative work, and software engineering.


supporting personalization and customization in a collaborative setting

Documents