A COMPARATIVE SURVEY OF ACTIVITY-BASED METHODS

FOR INFORMATION SYSTEMS DEVELOPMENT

Amanda Quek, Hanifa Shah

Faculty of Computing, Engineering and Technology, Staffordshire University, Beaconside, Stafford, United Kingdom.

Keywords: Activity, Activity theory, Methods, Information systems, Survey

Abstract: The role of human factors and the importance of sociocultural and contextual issues in information systems

(IS) development has long been recognised. However, these ‘soft’ details remain elusive and difficult to

capture. Activity theory (AT) provides a framework with which to analyse and understand human behaviour

in context. AT-based methods for IS development may therefore be a way forward. This paper presents a

comparative survey of five AT-based methods. Each method is described, and its strengths and weaknesses

briefly identified. The methods are then compared along nine key dimensions. As part of the findings, it is

determined that most of the methods are selective in their use of AT, and are not sufficiently validated.

Several correlations have also been noted across dimensions. Observations are presented on the limitations

of existing methods, and suggestions are then made on possible ways forward.

1 INTRODUCTION & CONTEXT

This paper describes five activity theory (AT)-based

methods for information systems (IS) development

and makes a detailed comparison between them. The

purpose of the comparison is to review the state of

the art in AT-based methods, and to identify

potential areas for improvement.

In the initial sections of this paper, a brief

context of the problem is given, and an overview of

AT is presented. Following that, five methods based

on AT are described, identifying the general

strengths and weaknesses of each. A detailed

comparison along nine dimensions is made, and the

results of the comparison are presented. Finally,

suggestions for improvement are offered, and

conclusions are drawn on the state of AT-based

methods for IS development.

The importance of considering contextual and

social issues in IS development has been well

documented, for example Mumford and Ward

(1968), Bignell and Fortune (1984), and Constantine

(2001). However, the difficulties of capturing

contextual information (Brown and Duguid, 1994),

and the paucity of practical methods available for

contextual analysis remains an ongoing concern in

IS research. This has resulted in researchers turning

to other fields of study in efforts to discover

knowledge, appropriate skills, or develop methods

that afford the capture of contextual and social

factors that can contribute to IS development.

One example of an alternative approach to

contextual analysis is AT. AT has been

recommended for use in IS development as a

guiding framework to elicit, analyse, understand and

incorporate contextual features (Bødker, 1991,

Kaptelinin, 1992, Kuutti, 1996, Hasan, 2001), and

has been applied to various types of computer-based

systems development.

Several AT-based methods have emerged for

use in IS development. These methods are new, still

evolving, and have emerged from diverse origins;

hence as a way forward, this paper makes a detailed

comparison between them and offers suggestions for

possible areas of improvement.

We now look at a definition of ‘method’, and

the reasoning behind the choice to use ‘method’ in

this paper. Avison and Fitzgerald (1995) define

method as a “recommended series of steps and

procedures to be followed in the course of

developing an IS”. The major difference between a

method and a methodology is that a methodology

has an explicit underlying philosophy, whereas a

method does not discuss philosophical aspects (ibid).

Therefore, we choose to use the term ‘method’

throughout this paper, as the underlying philosophy

is not deemed an informative point of comparison.

This is because the ‘soft’, human-focused nature of

221

Quek A. and Shah H. (2004).

A COMPARATIVE SURVEY OF ACTIVITY-BASED METHODS FOR INFORMATION SYSTEMS DEVELOPMENT.

In Proceedings of the Sixth International Conference on Enterprise Information Systems, pages 221-229

DOI: 10.5220/0002621002210229

 SciTePress

AT predisposes all AT-based work to have a similar

philosophy in mind.

The following section is a brief introduction to

AT and its body of principles and concepts.

2 ACTIVITY THEORY (AT)

Activity theory (AT) is a complex, abstract,

theoretical framework that provides concepts and a

vocabulary with which to analyse and understand

human activity in context. The concept of ‘activity’

is an entire field of study (Wertsch, 1979), which

cannot be detailed here. However, it can be said that

a major tenet of AT is that human activity can only

be understood within its sociocultural and historical

context.

AT stems from the work of Russian

psychologists Vygotsky and Leont’ev, in particular

Vygotsky’s (1934, 1978) theory of mediated

activity, which argues that all purposeful human

activity is accomplished through the use of physical

and/or psychological tools. After Vygotsky,

Leont’ev (1977, 1978) pioneered the concept of the

hierarchical levels in activity, and explained that

activity is always collective, never individual. AT

was developed further by Engeström (1987), who

produced the diagram of the activity system (Fig 1).

Figure 1: The Activity System (Engeström, 1987)

The activity system diagram consists of six

interlinked components – the subject, the tools, the

object (see section 2.2 for an explanation of the

term), the community, the rules, and the division of

labour. The subject refers to the actor or group of

actors that is carrying out the activity. The subject

uses tools to achieve an object, thereby turning the

object into an outcome (fig. 1). A subject can carry

out many activities, and each of the activities will

have a different object. Activities can be

distinguished from one another according to their

different objects.

The tools refer to both the physical and

psychological tools that are used by the subject in

order to achieve the object. The rules refer to both

the written and unwritten regulations which govern

and constrain the subject’s behaviour. The

community encompasses all the actors who work

with the subject and share the same object, and the

division of labour refers to the way work is divided

up between the subject and the community

members.

Figure 2: The Activity Network (Engeström, 1987)

The concept of ‘contradiction’ is an important

one in AT. Engeström (1987) explains

‘contradiction’ as a conflict or clash within and

between the components of the activity system, as

well as conflict within and between activity systems.

By identifying contradictions in an activity, we can

identify areas where improvements can be made to

work processes, and build these into the IS design.

Engeström (1987) also produced the activity

network diagram (Fig 2), which illustrates the wider

network of activities. In addition to the activity

system and activity network diagrams, there are six

guiding principles behind AT (Kaptelinin, 1992,

Wertsch, 1979). Each principle is described briefly

in the following sections, together with the way in

which the principle can contribute to IS

development.

2.1 Unity of consciousness and

activity

This principle stresses that humans learn by doing,

and that the human consciousness is formed by

interaction with the external world (Kaptelinin,

1992). External activity is constantly informing and

building the mental world, and vice versa – there is

no separating the two. The implications for IS

development are that in order to understand or

influence human thinking and behaviour, we must

look at the interaction between the person and their

activity in context.

2.2 Object-oriented activity

The term ‘object’ as used in AT refers to both the

‘objective’ of the activity as a mental construct, and

the physical or conceptual ‘thing’ that is worked on.

Both of these evolve as the activity is carried out.

Activity is described as being ‘object-oriented’ (not

Object Subject

Division of Labour

Community

Outcome

Tool -

Producing

Activity

Subject -

Producing

Activity

Rule-

Producing

Activity

Object’s

Activity

Culturally More

Advanced Central

Activity

Central Activity

ICEIS 2004 - HUMAN-COMPUTER INTERACTION

222

to be confused with the software engineering term),

which means that every activity has a purpose (or

object), and is carried out in order to achieve some

outcome. Thus, it is essential to determine what the

object of each activity is in order to design systems

that support users in achieving what they need to do.

2.3 Mediation

The principle of mediation states that in every

purposeful activity there will be tools involved, both

physical (e.g. hammer and nails) and psychological

(e.g. language and mathematics). A computer is a

hybrid of the two: it is a physical tool that is used for

psychological processing. AT views an IS as a tool

for people who are carrying out some higher

activity. This viewpoint shifts the emphasis from

being technically oriented to user oriented. In order

to design systems that support the work of humans,

we need to understand what tools are currently used,

how they have evolved, and how their functions

could be supported by the system.

2.4 Hierarchical structure

This principle states that an activity can be

decomposed into actions and operations. Each

activity can be decomposed into actions, which are

conscious steps that are taken to carry out the

activity. Actions in turn can be decomposed into

operations, which are subconscious. For example,

from the point of view of an IS developer, the

activity of ‘developing an IS’ can be decomposed

into actions such as ‘attending meetings’, ‘talking

with users’, and ‘programming’. The action of

‘programming’ can be decomposed further into

subconscious operations, such as using the keyboard

and the mouse.

These levels are constantly in a state of flux.

Conscious actions, if undertaken often enough, can

become subconscious. Similarly, a subconscious

operation can revert to a conscious action if a

problem is encountered. An IS needs to support

users’ work at all three levels concurrently, to ensure

a smooth transition through the levels.

2.5 Internalisation/externalisation

Internalisation is the process by which mental

representations are formed by carrying out external

actions. Externalisation is the opposite: where

mental representations are manifested in external

actions. Internalisation and externalisation are

closely related to the way learning takes place. The

implication of this principle is the need for the IS

itself to support users in learning to use the system,

to support the internalisation of system functionality,

and to support users who are at different levels of

expertise.

2.6 Development

This principle explains that activity can only be

understood through analysis of its developmental

transformations. All human activity is a result of

historical development, and is constantly changing

and re-forming. We need to analyse and understand

how work has developed over time in order to fully

grasp how work is done today. At the same time, the

concept of development tells us that change and

development are certain to occur whenever humans

are involved. Therefore, as far as possible, we

should ensure that the systems we design are

equipped to handle change and development in work

practice.

Within the field of IS development, AT has been

widely used in computer supported collaborative

work (Kuutti and Arvonen, 1992), computer

supported collaborative learning (Gifford and

Enyedy, 1999), and human-computer interaction

(Bødker, 1991), and has been a major contributor to

the Scandinavian tradition of systems design.

The following section describes AT-based

methods in detail.

3 AT-BASED METHODS FOR IS

Following a review of the literature, five AT-based

methods emerged that aim to operationalise the

theoretical concepts of AT to produce practical

methods for IS development. These methods are the

ActAD method, the Activity Checklist, the AODM

method, the Jonassen & Rohrer-Murphy framework,

and the Martins & Daltrini framework. A

comparison of these methods will be made from

different dimensions, and conclusions are then

drawn.

3.1 The ActAD method

The Activity Analysis and Development (ActAD)

method was first developed by Korpela (1997). It is

recommended for several purposes, namely for IS

users (to improve their own work processes), IS

developers (to analyse the IS users’ work in order to

design improved IS facilities for them), and IS

researchers (to design improved IS methods and

techniques for the developers). Korpela et al. (2000)

A COMPARATIVE SURVEY OF ACTIVITY-BASED METHODS FOR INFORMATION SYSTEMS DEVELOPMENT

223

focus in particular on the use of ActAD as a method

for IS development.

ActAD provides a framework for IS developers

to examine sociocultural features that can inform

development of an IS. The first step guides

developers to analyse the components of the activity

that is to be supported by the system, and provides

checklists of guiding questions with which to elicit

these components. The surrounding activities are

analysed in the second step. The activity system and

the activity network diagrams are derived from those

of Engeström (1987), and have been modified to

become more graphical.

Step 3 in the method focuses on analysing the

development of the central activity, which is broken

down into ‘History’, ‘Problems, and ‘Potential’. A

brief checklist of questions is provided with which

to elicit each factor.

In the fourth step, the required new tools are to

be developed, and processes are to be improved,

based on the information elicited previously. The

final step in ActAD involves disseminating the

results, evaluating the process, and starting again,

looping back to the problem analysis in Step 3.

A limitation of ActAD is the lack of a notation

with which to document the third, fourth and fifth

steps in the method. The sparseness of guidelines

provided in these steps also calls for future

development.

3.2 The Activity Checklist

The Activity Checklist was developed by Kaptelinin

et al. (1999), and aims to enable researchers and

designers to identify the contextual factors that can

influence the use of computer technology in a real

life setting, and to spot potential trouble areas that

designers can address.

The checklist has two foci – design and

evaluation, and subsequently there are two slightly

different versions. Both versions of the checklist

consist of four columns, based on four of the

principles of AT. The column headings are:

Means/ends (which relates to the hierarchical

structure), Environment (related to object

orientedness), Learning / cognition / articulation

(related to internalisation / externalisation), and

Development (named after the corresponding AT

principle). The principle of mediation is said to

permeate all four columns. Within each column,

between 5 and 13 items are listed to guide the

analysis. The developer is advised to generate their

own questions based on the items listed, and a table

of sample questions is provided.

The main weakness perceived with the checklist

is the repeated usage of activity theoretical jargon.

The checklist is also purely textual, and described at

a high level of abstraction.

3.3 The AODM Method

The Activity-Oriented Design Method (AODM) was

first detailed in Mwanza (2001) and is based on the

models of Engeström (1987). AODM was developed

further in Mwanza (2002), and is intended to

contribute to the early phases of systems

development, with an aim to support requirements

capture, analysis, and design, focusing on human-

computer interaction. The method consists of 6

stages, and 4 tools.

Stage 1 analyses the situation involved. For this

stage, the first tool is provided - the ‘eight-step-

model’. This is a list of 8 questions that guide the

analysis of the activity and its components. Stage 2

involves modelling the situation, using the

information obtained in Stage 1 with the activity

system model (Engeström, 1987). Following that,

stage 3 decomposes the activity to reduce

complexity. The ‘activity notation’ tool is provided

to assist in this stage. This tool details 6 ‘sub-

triangles’ that can be analysed in order to

decompose the activity.

Stage 4 is aided by a third tool, consisting of 6

general questions, which can be used to generate a

wide range of research questions to analyse the

interaction and relationships within and between the

components of each subtriangle. This tool also

elicits the presence of conflict within and between

the components. In Stage 5, the research questions

generated are used in data gathering, e.g. in

interviews, questionnaires, or observation. Finally,

Stage 6 involves interpreting and communicating the

findings. For this a fourth tool is provided - the

diagram for mapping operational processes. This

tool presents the results of Stage 4 in illustration

form, with clear visual indications of the research

questions generated, as well as the areas of conflict

that have become apparent, facilitating

understanding of the process as well as the results.

A unique contribution of AODM is its

development of subtriangles as further units of

analysis. However, the application of AT in AODM

appears to be in differing degrees, for example, the

principle of mediation seems to be given more

prominence than development.

3.4 The Jonassen & Rohrer-Murphy

Framework

The framework of Jonassen & Rohrer-Murphy

(1999) is a method for the design of constructive

ICEIS 2004 - HUMAN-COMPUTER INTERACTION

224

learning environments (CLE). There are six steps in

the framework, each divided into a number of

substeps. Each substep provides sample questions to

be asked and actions to be taken.

The first step is primarily based on the principle

of object-orientedness, and guides the developer in

examining the subject’s (learner’s) goals to

determine the purpose of the activity that the CLE is

to support. In the second step, each component of

the activity system is examined. The third step is

chiefly based on the principle of hierarchical

structure, and leads to the decomposition of the

learner’s activities into actions and operations. The

fourth step involves elicitation of the tools and other

mediatory means that have been and could be used

in the CLE.

Step 5 of the framework analyses the context,

and the substeps within guide the analysis of the

community, rules, and division of labour present in

the activity. Finally, step 6 analyses the interaction

and rules for the relationships that exist within and

between the components of the activity system.

Each step of the framework is clearly based on

individual principles and components from AT, but

some principles influence all the sections. For

example, the principle of development permeates the

entire framework. The framework provides a large

set of questions to be answered that cover diverse

combinations of AT principles and components.

However, the level of granularity to which each

question refers is not always clear. There is also

repetition across the steps of the framework, as some

questions are closely related.

3.5 The Martins & Daltrini

Framework

The framework of Martins & Daltrini (1999) is an

approach for requirements elicitation. There are

three steps in the framework. The first step is to

identify the activities to be supported by the target

system. The second step is to identify the

components of the activity system belonging to each

activity, based on Engeström’s activity system

model (fig 1). The third step is to decompose each

activity into actions and operations, based on the

principle of the hierarchical structure of activity.

This is done with the aid of a table with three

columns: activity, action and operation. The list of

actions and operations obtained can be used to

derive requirements for the target system.

This framework focuses mainly on the principle

of the hierarchical structure of activity. It provides

little support for identifying activities, and there is

little guidance provided for identifying the activity

system components.

3.6 Summary of general strengths

and weaknesses

Table 1 shows a summary of the general strengths

and weaknesses observed in each of the methods.

Table 1: Strengths and Weaknesses of AT-based Methods

Strengths Weaknesses

ActAD

Activity system model

made more graphically

oriented

Lack of a notation and

sparse guidelines, in steps 3,

4, and 5

Activity Checklist

Strong application of AT

theoretical principles

Purely textual, abstract,

high usage of AT jargon

AODM

Unique contribution in

using subtriangles as further

units of analysis

Some principles given

prominence over others e.g.

mediation

Jonassen and Rohrer-Murphy

Full application of AT

principles, extensive list of

analysis questions

Lengthy and sometimes

repetitive

Martins and Daltrini

Provides a notation for the

decomposition of actions

and operations

Does not provide guidance

for identifying activities or

components

The strengths of each method indicate the

contributions that are made, which future work

could build upon. The weaknesses could indicate

areas that need improvement. These have been

discussed in detail in the previous sections.

4 COMPARISON OF THE

METHODS

This section details the results of the comparison of

the methods. There are many dimensions of

comparison that could be taken, but due to space

constraints, this paper looks at 9 key dimensions.

We propose that a comparison of AT based methods

requires several different dimensions to those used

in studies that compare IS methods. Traditional

factors for comparison between IS methods are such

as life cycle coverage, underlying philosophy

(‘systems’ or ‘science’), ‘structuredness’, user role,

and techniques used (data flow diagrams, dialogue

design, entity modelling, etc) e.g. Tudor and Tudor

(1997). Some of these dimensions (e.g. the

development phases supported) are suitable for our

purpose. However, the main focus in comparing AT-

based methods is on the way that AT is applied and

made practical by the method. To reflect this, in

A COMPARATIVE SURVEY OF ACTIVITY-BASED METHODS FOR INFORMATION SYSTEMS DEVELOPMENT

225

addition to traditional dimensions for comparison,

we have chosen to compare the extent to which the

method uses AT visual models, the AT principles

that informed the method, the elicitation of

‘contradiction’ and historical background, and the

provision of analysis questions. Because these

methods are new, we also compare the way the

method was developed and validated.

• Development phases supported

• Area or type of system supported

• AT visual models used

• AT principles that informed the method

• Elicitation of ‘contradictions’ in the activity

• Elicitation of the historical background

• How the method was developed

• Provision of analysis questions

• Validation of the method

There are other factors that could be used for

comparison, but due to space constraints, only 9 key

dimensions are described here.

4.1 Development phases supported

Table 2 shows that each method supports different

phases of development. No method covers all phases

of development. The phases that are strongly

supported by methods are domain analysis,

requirements elicitation and design. There is less

support for the phases of evaluation and interface

design. However, this does not mean that AT is less

suitable for use in these phases, or in phases that are

as yet unsupported by methods (such as the

implementation phase). Rather, this testifies to the

need for further research into producing and testing

AT-based methods that can be used throughout the

IS development process.

Table 2: Development phases supported

Domai

Analy

sis

Requirem

ents

Elicitation

Desi

Interfa

Desig

Evaluati

ActA

Activi

Check

list

AOD

Jonass

partial

Martin

s &

Daltri

4.2 Area or type of system supported

AT-based approaches have been proposed for use in

different areas of computing, and for several types of

systems design. Table 3 details the area of study and

the type of system supported by each method. Two

of the methods were designed for specific areas of

application, namely, the AODM method, which is

aimed specifically towards human computer

interaction, and the Jonassen & Rohrer-Murphy

framework, which is oriented towards designing

computer supported collaborative learning. The

other three methods are for general systems

development.

Table 3: Area or type of system supported

Area or type of system

supported

ActAD General

Activity Checklist General

AODM Human Computer Interaction

Jonassen Computer Supported

Collaborative Learning

Martins & Daltrini General

4.3 AT visual models used

Two of the visual models provided by AT are the

activity system and activity network (Engeström,

1987). Table 4 shows that three of the methods use

the activity system model, and one of these three

also uses the activity network.

Table 4: AT visual models used

Activi

Syste

Activi

Netwo

Sub-

triangl

Table of

activities,

actions, and

operations

ActAD Yes Yes No No

Activity

Checklist

No No No No

AODM Yes No Yes No

Jonassen No No No No

Martins

Daltrini

Yes No No Yes

Two methods provide their own visual models–

the AODM method provides a graphical

representation of the subtriangles within the activity

system, and the Martins & Daltrini method provides

a table to record the breakdown of activities, actions

and operations. Two methods, namely the Activity

Checklist and the Jonassen & Rohrer-Murphy

framework, do not use any visual models.

ICEIS 2004 - HUMAN-COMPUTER INTERACTION

226

4.4 AT principles applied

Table 5 shows the AT principles that are explicitly

applied by each method. Upon comparison, the

Activity Checklist and the Jonassen & Rohrer-

Murphy methods produce the broadest application of

the AT principles, that is 5 for the Checklist and 6

for the Jonassen & Rohrer-Murphy framework. The

Martins and Daltrini framework only applies one

principle- the ‘hierarchical structure’. The only

method that explicitly applies the principle of the

‘unity of consciousness and activity’ is the Jonassen

and Rohrer-Murphy framework. The choice of

which principles to apply seems to depend on the

author of the method, and for the most part it is

unclear how those choices were made.

Table 5: AT principles applied

Unity of

C. and A.

O-O Med. Hier.

Str.

I / E Dev

ActAD

Activity

Checklist

9 9 9 9 9

AODM

9 9

Jonassen

9 9 9 9 9 9

Martins

Daltrini

Key to AT Principles:

Unity of C. and A. – Unity of Consciousness and Activity

O-O

– Object-Orientedness

Med.

– Mediation

Hier. Str.

– Hierarchical Structure

I / E

– Internalisation/Externalisation

Dev. – Development

4.5 The elicitation of ‘contradictions’

Identifying contradictions produces a significant

contribution towards designing systems that improve

the way work is done. Table 6 shows that more than

half of the methods carry out some form of analysis

of the contradictions existing within the work

activity.

Table 6: Elicitation of ‘contradictions’

Elicitation of ‘contradictions’

ActAD Yes

Activity Checklist No

AODM Yes

Jonassen Yes

Martins & Daltrini No

4.6 The elicitation of the historical

background of the activity

Understanding the historical background is

important if we are to understand how the activity

has developed over time. Table 7 shows that only

one method includes the analysis of the historical

aspect of activity: the ActAD method.

Table 7: Elicitation of historical background

Elicitation of historical

background

ActAD Yes

Activity Checklist No

AODM No

Jonassen No

Martins & Daltrini No

4.7 Analysis questions provided

AT can be seen as a framework for analysis of

activity. In order to carry out an AT analysis,

information has to be elicited from the practitioners

of the activity. Therefore, specific leads or questions

for analysis are required.

Table 8: Analysis questions provided

Analysis questions provided

ActAD Partial

Activity Checklist Partial/Yes

AODM Yes

Jonassen Yes

Martins & Daltrini No

Table 8 shows that two of the methods provide

clear sets of analysis questions; in fact, the Jonassen

& Rohrer-Murphy framework consists entirely of

sets of questions. The AODM method also provides

questions, and has a dedicated step for generating

research questions. The Activity Checklist provides

sample questions, but encourages the analyst to

create their own. The ActAD method provides

questions for its initial three steps, but the questions

for the third step are at a general level, and are of

limited use. Finally, the Martins & Daltrini method

does not provide any questions.

4.8 Development of the method

The way the method was developed can provide us

with information on the orientation of the method –

whether it is developed theoretically or empirically.

A COMPARATIVE SURVEY OF ACTIVITY-BASED METHODS FOR INFORMATION SYSTEMS DEVELOPMENT

227

Table 9: Development of the method

Development of the method

ActAD At desk

Activity Checklist At desk

AODM Empirical

Jonassen At desk

Martins & Daltrini At desk

Table 9 shows that four of the methods were

developed ‘at desk’, in other words, based solely on

the theory. Only the AODM method was developed

empirically, in a real world commercial

environment.

4.9 Validation of the method

The importance of method validation cannot be

underestimated, particularly for AT-based methods,

which are new. Rigorous testing is required to

ensure that AT based methods are able to provide

quality of output.

Table 10: Validation of the method

Validation of the method

ActAD Two small experiments, neither real

life

Activity Checklist Self validation by the authors

AODM Two industrial organisations over

two years

Jonassen None

Martins &

Daltrini

One interview

Table 10 shows that only the AODM method

was tested in a real life setting, in two industrial

organisations. The ActAD method, the Checklist and

the Martins & Daltrini method were validated in

small experiments, and the Jonassen & Rohrer-

Murphy framework does not provide any

information that it has been validated.

4.10 Findings from the Comparison

The findings of the comparison include a number of

correlations across the dimensions that were used.

One finding is that a broad use of AT principles

seems to correlate with less use of AT visual

models, and vice versa (Sections 4.3 and 4.4). For

example, both the Activity Checklist and the

Jonassen & Rohrer-Murphy methods indicate high

usage of the principles of AT, and non-usage of

visual models, while the ActAD, AODM and

Martins & Daltrini methods indicate low usage of

the principles and the use of two out of four visual

models.

Several methods are also selective in their

application of AT principles, and it is often unclear

how these choices were made. To avoid the arbitrary

application of AT, it is important that justification be

provided in future by methods that selectively apply

AT concepts.

It is also found that methods developed ‘at desk’

were not subjected to as much validation as the

empirically developed method (Sections 4.8 and

4.9). Further empirical research is necessary in order

to validate the methods. More research is needed to

produce methods that support wider coverage of

development phases.

5 CONCLUSION

This paper has described five AT-based methods for

IS development, and has made a comparison

between them along nine dimensions. The general

strengths and weaknesses of each method were also

described, which indicate areas of strong

contribution as well as potential for improvement for

each method. As part of the findings, several

correlations have emerged between the dimensions

of comparison. It is found that within the AT-based

methods that have emerged from the survey, there is

a lack of comprehensive treatment, regarding

coverage of development phases as well as coverage

of AT concepts. It is also found that only one of the

methods has been validated in a real life systems

development.

In the light of the findings, we suggest that

existing methods need to be thoroughly tested and

documented through empirical studies in practice.

Further work is also needed, both to improve

existing methods, as well as to produce new methods

for making AT concepts applicable in practical

development scenarios.

ACKNOWLEDGEMENTS

This research is funded by the Tracker project at

Staffordshire University (EPSRC Grant No.

GR/R12183/01).

REFERENCES

Avison, D. E. and Fitzgerald, G. (1995) Information

Systems Development: Methodologies, Techniques and

Tools, McGraw Hill.

Bignell, V. and Fortune, J. (1984) Understanding Systems

Failures, Manchester University Press.

ICEIS 2004 - HUMAN-COMPUTER INTERACTION

228

Bødker, S. (1991) Through the Interface - A Human

Activity Approach to User Interface Design, Lawrence

Erlbaum Associates.

Brown, J. S. and Duguid, P. (1994) Borderline Issues:

Social and Material Aspects of Design. Human-

Computer Interaction, 9, 3-36.

Constantine, L. L. (2001) The Peopleware Papers: Notes

on the Human Side of Software, Prentice Hall.

Engeström, Y. (1987) Learning by Expanding: An Activity

Theoretical Approach to Developmental Research.

Gifford, B. and Enyedy, N. (1999) Activity Centred

Design: Towards a theoretical framework for CSCL.

In 3rd International Conference on Computer Support

for Collaborative Learning.

Hasan, H. (2001) An Overview of Different Techniques

for applying Activity Theory to Information Systems.

In Information Systems and Activity Theory: Theory

and Practice(Ed, Hasan, H.) University of

Wollongong Press.

Jonassen, D. H. and Rohrer-Murphy, L. (1999) Activity

Theory as a Framework for Designing Constructivist

Learning Environments. Educational Technology

Research and Development, 47, 61-79.

Kaptelinin, V. (1992) Activity Theory: Implications for

Human-Computer Interaction. In Context and

Consciousness: Activity Theory and Human-Computer

Interaction(Ed, Nardi, B.).

Kaptelinin, V., Nardi, B. A. and Macaulay, C. (1999) The

Activity Checklist: A Tool For Representing the

"Space" of Context. Interactions.

Korpela, M. (1997) Activity Analysis and Development in

a Nutshell. http://www.uku.fi/atkk/actad/nutshell.html.

Korpela, M., Soriyan, H. A. and Olufokunbi, K. C. (2000)

Activity Analysis as a Method for Information

Systems Development. Scandinavian Journal of

Information Systems, 12, 191.

Kuutti, K. (1996) Activity Theory as a Potential

Framework for Human-Computer Interaction

Research. In Context and Consciousness: Activity

Theory and Human-Computer Interaction(Ed, Nardi,

B.) Massachusetts Institute of Technology.

Kuutti, K. and Arvonen, T. (1992) Identifying Potential

CSCW Applications by Means of Activity Theory

Concepts: A Case Example. In CSCW '92ACM.

Leont'ev, A. N. (1977) Activity and Consciousness.

http://www.marxists.org/archive/leontev/works/1977/

leon1977.htm.

Leont'ev, A. N. (1978) Activity, Consciousness, and

Personality. http://www.marxists.org/archive/leontev/

works/1978/intro.htm.

Martins, L. E. G. and Daltrini, B. M. (1999) An Approach

to Software Requirements Elicitation Using Precepts

from Activity Theory. In

14th IEEE International

Conference on Automated Software Engineering, pp.

15-23.

Mumford, E. and Ward, T. B. (1968) Computers, Planning

and People, BT Batsford Ltd, London.

Tudor, D. J. and Tudor, I. J. (1997) Systems Analysis and

Design

A Comparison of Structured Methods, Macmillan Press.

Vygotsky, L. S. (1934) Thought and Language,

Massachusetts Institute of Technology, Cambridge,

Massachusetts.

Vygotsky, L. S. (1978) Mind In Society, Harvard

University Press, Cambridge, Massachusetts.

Wertsch, J. V. (1979) The Concept of Activity in Soviet

Psychology: An Introduction. In The Concept of

Activity in Soviet Psychology(Ed, Wertsch, J. V.) M E

Sharpe, Inc.

A COMPARATIVE SURVEY OF ACTIVITY-BASED METHODS FOR INFORMATION SYSTEMS DEVELOPMENT

229