ONTOLOGICAL MODELLING TO SUPPORT THE PLANNING

OF IS DEVELOPMENT PROCESSES

A Position Paper

Robert T. Hughes, Graham Winstanley and Karl Cox

School of Computing, Mathematical and Information Sciences, University of Brighton

Watts Building, Lewes Road, Moulsecoomb, Brighton, BN2 4GJ, U.K.

Keywords: Project planning, Cognitive causal mapping, Process modelling, Method engineering.

Abstract: IT projects are known for the high rate at which they fail. Past work by the authors has investigated the

building of cognitive causal maps to find and represent what the participants in a project feel are factors that

lead to project success or failure. It was found that while agreement can often be reached on the broad

causes of failure, there tended to be differences about the precise nature of the identified factors (for

example the exact meaning of 'inadequate resources'). The position paper proposes the use of ontological

models to enrich and clarify causal maps with information about the classes of object in the real world to

which they refer. This would facilitate more effective planning of new projects. An aspiration of the authors

is to use the information generated by ontology-enriched causal maps to provide guidance on the tailoring of

methodologies, particularly Agile ones, for specific projects.

1 INTRODUCTION

1.1 Context and Motivation

This position paper presents a proposed programme

of work that seeks to enrich the understanding of

participants in IT projects of the factors that affect

project success and failure and the likely

effectiveness of actions aimed at promoting success.

It proposes to extend work which captures the

perceptions of stakeholders by means of cognitive

causal maps by adding ontological models to

provide details of ‘real world’ entities and their

attributes implied by the factors that have been

identified. This will support project planners by

clarifying the implications of decisions, particularly

their alignment with the outcomes sought for the

project.

There is justified dissatisfaction with the failings

of IT development projects. For example, a report on

the UK health service’s national programme for

information technology (NPfIT) noted a 30% failure

rate for IT projects. One such IT project failure in

the Wessex Regional Health Authority led to a loss

of £43millions of public money (Hendy et al., 2005).

Reducing the proportion of such failures would

clearly generate financial savings and increase the

benefits of successfully implemented projects.

One response to project failures from the

project management community has been an

emphasis on risk management – see, for example,

Boehm (1991). Risks may be the result of

uncertainty or of ignorance. In some cases, it is

difficult, if not impossible, to predict what occurs –

for example that a fire at a supplier will prevent

specialist IT equipment being delivered. However,

in many cases the risk is caused by a lack of

knowledge. An enterprise with little IT experience

might misguidedly acquire an order processing

system unsuitable for their business. The knowledge

that might have avoided this existed in the world,

but had not been disseminated to those who needed

it. It is this type of risk that is our focus.

It is acknowledged by most writers on the

management of IT projects that it requires attention

to both technical and social issues. For example,

Winter et al. (2006) recognised the inevitably

instrumental nature of much of project management

– the execution of planned activities to achieve

physical outcomes – but emphasised the need to

develop a richer understanding of the ‘concepts and

images which focus on social interaction between

people, illuminating the flux of events and human

action…’. Major projects have a range of

319

T. Hughes R., Winstanley G. and Cox K..

ONTOLOGICAL MODELLING TO SUPPORT THE PLANNING OF IS DEVELOPMENT PROCESSES - A Position Paper.

DOI: 10.5220/0003118003190324

In Proceedings of the International Conference on Knowledge Engineering and Ontology Development (KEOD-2010), pages 319-324

ISBN: 978-989-8425-29-4

 2010 SCITEPRESS (Science and Technology Publications, Lda.)

stakeholders with distinct interests and developers

with differing expertise, each of whom is aware of

different types of risk. Sharing of viewpoints and

agreement on joint action to reduce risks is needed.

The identification of risks is futile unless projects

are flexible enough to adapt to risks. Agile

approaches address shortcomings in IT development

by measures to promote flexibility, for example,

delivering projects in smaller units with more

interaction between developers and users. However,

we believe that while Agile approaches hold great

promise as an effective way of organising

development projects, they demand a greater

common understanding of the nature and context of

the project by project participants. Our proposed

research attempts to support this understanding.

1.2 Cognitive Causal Maps (CCMs)

This new research direction has grown out of

previous work by the authors and their colleagues

which investigated ways of improving the

management of risk in IT developments – see Al-

Shehab (2007), Al-Shehab et al (2005, 2006),

Hughes et al. (2006). This work used cognitive

causal maps (CCMs) to diagnose the causes of IT

project failure. CCMs are diagrams consisting of

nodes representing the outcomes of a particular

course of action and connectors between the nodes

which indicate where some factors influence others

– see Figure 1. The technique is supported by a

research tradition starting with Axelrod’s seminal

work (Axelrod 1976). In the UK CCMs are

particularly associated with Eden and his colleagues

– see, for example, Eden (2004).

When applying CCMs to the causes of project

success and failure, some nodes represent the

desired outcomes of the project. Other nodes

represent policies, the means by which the desired

outcomes are to be achieved. Further factors are

environmental relating to conditions assisting or

hindering the achievement of the desired outcomes.

The factors are presented as ‘concept variables’

which for a context take a value within a range

bounded by two opposing poles, for example, a

range varying from an abundance of required skills

on the one hand to a severe skills shortage on the

other. Connectors are drawn between these nodes

indicating the influence – positive or negative – of

the factors on one other.

Figure 1 below describes a situation where a

shortage of staff with the skills needed for a task

means that resources available to a project do not

have the expertise envisaged when the project was

planned. (The minus sign indicates that the skill

shortage influences the second pole of

adequate...inadequate resources, i.e. makes

resources inadequate). This leads to the planned

project duration being exceeded. A policy of

additional training can, to a certain extent, offset the

problems caused by the skills shortage.

Figure 1: Notations used in a primitive Cognitive Causal

Map, following Eden 2004.

As part of our previous work, collaborative sessions

with stakeholders involved in problematic projects

were used to generate relevant CCMs. Two

problems emerged from this:

1) There might be agreement on the importance of

a factor such as ‘inexperienced staff’ or ‘poor

project management’, but with differences of

opinion about what the participants meant by

these;

2) Little guidance was generated on the practical

steps needed for success on new projects.

1.3 Adding Ontological Support to

CCMs

CCMs need details of their context to be more

effective. Recent work by Chauvin and colleagues

(2007) shows the scope for ontological models that

describe the context of CCMs. We intend to assess

this approach to ontological modelling as a basis for

enriching the data gathered from the collaborative

creation of CCMs by project stakeholders. In Figure

1, the classes of interest in a project environment

might include:

• Developers who carry out the work

• The technologies with which they are

familiar

• The activities to be carried out

• The technologies that the activities will use.

A collection of such descriptions can form the basis

KEOD 2010 - International Conference on Knowledge Engineering and Ontology Development

320

of a project management ontology. If the

participants in a project agree on this ontology, then

the potential for agreement on the meaning of

higher-level terms such as ‘adequate resources’ is

increased. This facilitates the selection of actions to

reduce or mitigate obstacles to project success.

Technology

type

Developer

Personal

expertise

Project

Actviity

Expertise

requirement

Expertise

match

Figure 2: A fragment of an ERD to support the concept

variable adequate...inadequate resources.

A fragment of an entity relationship diagram (ERD)

which might support the concept variable

adequate...inadequate resources is illustrated in

Figure 2. A developer could be skilled in a number

of technologies which could include particular

programming languages such as Java. A project

consists of a number of activities, each of which

requires the application of one or more technologies.

The relevant expertise of a developer will depend

not just on their innate qualities but the demands of

the new project. Someone who has been highly

regarded as knowledgeable may have their expertise

reduced by the introduction of new technologies.

The classes in the project ontology can be

mapped to instances in the project plan. Thus a three

level structure of concept, class and instance can be

envisaged – see Figure 3. In this context ‘concept’

has a broader meaning than class and can be

composed of a number of classes and relationships.

Previous work on ontological modelling has

been mainly associated with artificial intelligence,

seminal work in the field being that of Gruber

(1995). The drive to create a ‘semantic web’

enabling more meaningful and trustworthy

information retrieval has led to a wider interest in

ontology creation.

Eden (2004) and Marshall (2009) have used

CCMs to analyse existing organisational contexts.

The proposed work will also use CCMs as an

investigative tool, but will in addition follow some

of the relatively few examples (most prominently

Abdel-Hamid 1988) where CCMs have been used to

Figure 3: Relationships between a CCM, a Project world

ontology and a project plan.

construct decision-making tools. Techniques have

also been developed which trace linkages between

business objectives and the IS/IT developments

needed to support them (Bleistein et al. 2004, 2006a,

2006b, Babar et al. 2008). These use a CCM-like

goal modelling notation, but generate guidance on

what is to be developed whereas the current work is

concerned with how.

Some ontological modelling techniques are

similar to those of database specialists. Methods in

use in the construction of object-oriented software

are applicable to ontology construction (De Nicola et

al 2009). Conventional data modelling may be

adequate, but a pragmatic advantage of a specifically

ontological approach is that it facilitates the

incorporation of existing project-related ontologies,

for example, PROMONT (Abels et al. 2006),

PLANET (Gil and Blythe 2000) and KANAL (Kim

and Gil 2001), into new ontologies either directly or

in a modified form.

Some research has developed more rigorous

forms of CCMs where the strengths of factors and

the links between them are expressed numerically

and the models executed to produce predictions

(Abdel-Hamid 1988). One middle way uses a fuzzy

quantitative approach (Stach and Kurgan 2004). Our

own work has made extensive use of fuzzy

quantitative mechanisms developed by Montibeller

et al. (2007) to describe, through reasoning maps,

expert decision-making processes. The emphasis on

the new direction of our research is to make CCMs

created in conjunction with stakeholders more useful

by increasing their semantic richness, rather than

simply attempting to quantify factors which may in

any case lack clear definition.

Agile project management is a good fit with the

aspirations of the new work as it emphasises the

tailoring of project processes to fit the project’s

context. The focus will be on DSDM/Atern (DSDM

ONTOLOGICAL MODELLING TO SUPPORT THE PLANNING OF IS DEVELOPMENT PROCESSES - A Position

Paper

321

2007) because: it is applicable to a wider span of

IS/IT development activities than approaches such

as XP (Beck and Andres 2005) which focuses only

on software code production; it has a relatively wide

user base in the UK and, pragmatically, the project

team have links with DSDM/Atern practitioners.

DSDM/Atern is characterised by: (i) The initial

formulation of requirements in terms of the business

objectives to be met by a development; (ii) the

division of the product delivery into increments,

each of which should achieve the deployment of

usable system components generating user benefits

and (iii) the prioritisation and allocation of

requirements to increments with fixed deadlines.

The second Agile approach to be considered will be

Scrum (Schwaber 2007), which, like DSDM/Atern,

is not focussed exclusively on software code

development. Scrum appears to emphasise the

behavioural aspects of projects, compared to the

DSDM/Atern concern with process. The

examination of more than one approach encourages

the development of more generic and robust research

outcomes.

2 RESEARCH QUESTIONS

We propose that CCMs enriched by ontological

modelling can support more effective decision-

making by those planning and managing IT projects,

particularly projects that use an Agile approach. To

confirm this, we believe the following research

questions will need to be addressed:

R1. How can project processes be most

effectively represented as ontologies?

R2. Can ontological models effectively capture

cause-and-effect relationships?

R3. Can ontological models be created which

clarify the perceptions of a project captured by a

CCM?

R4. To what degree can ontological models

usefully support tools and operations at the level of

practical project planning and management?

These four questions will be discussed in turn in

the next four sections.

2.1 How can Project Processes be Most

Effectively Represented as

Ontologies?

A methodology contains a set of recommended

(sometimes mandatory) steps for carrying out a

procedure. It may be a codification of good practice

or be mandated by an authority (as with the

PRINCE2 project management standard in UK

government projects). Any method will inevitably

have to be ‘fine-tuned’ for local use (Fitzgerald et al.

2003). We have already established the feasibility of

representing project processes in an ontological

model (Hughes 2010). We now need to find a way

of not only modelling core work practices in an

ontology, but also variations on that core.

We believe that Höfferer (2007) provides

valuable guidance on deriving ontologies for

processes. Processes can be modelled for an instance

of a project (level M

), e.g. a plan. At the next

highest level (M

), activities and other project

characteristics can be generalised to cover a class of

projects – e.g. as a software development lifecycle.

A metamodel (M

) can further generalise to

superclasses – for example, ‘design’, ‘build’ and

‘test’ can be generalised as instances of ‘activity’.

This procedure will be applied to the two Agile

process models DSDM/Atern and Scrum, with

separate level M

models for the two approaches,

and then an attempt at a level M

model of a general

Agile project model. The resulting models will be

validated by populating them with sample instance

(level M

) data from specific project scenarios.

These procedures implement the process to class

mapping transition from process model at the

Diagram/Concept level in Figure 3 to the Process

Model Ontology at the Classes level.

2.2 Can Ontological Models Effectively

Capture Cause-and-effect

Relationships?

Although we often generalise the idea of causality,

different cause-and-effect relationships involve

different processes. Some are physical processes,

e.g. ‘fire destroys data centre’, others matters of

human motivation, e.g. ‘better job opportunities

cause staff departures’. These differences are

noticeable when trying to combine the different

influences on a particular concept variable: in some

cases both A and B are needed to cause an effect

(e.g. a reliance on external resources and a shortage

of those resources), in other cases either A or B (e.g.

an increase in costs or a decrease in income).

Our experience with CCMs leads us to believe

that the types of causality identified in CCMs need

to be more carefully analysed. Some guidance on

this may clearly come from the literature on

causality – one obvious source is Pearl (2000). Our

aspiration is to create a causal taxonomy to be

incorporated into the process-oriented ontologies

identified above. This is part of the concept to class

transition from CCMs to Cognitive Causal ontology

in Figure 3, but relates only to the categorisation of

KEOD 2010 - International Conference on Knowledge Engineering and Ontology Development

322

linkages at CCM Concept level and equivalent

properties at Class level.

2.3 Can Ontological Models be Created

which Clarify the Perceptions of a

Project Captured by a CCM?

The successful construction of the envisaged

ontologies depends on effective modelling of

stakeholder perceptions. Existing ontological

modelling tools are usually not designed primarily

for easy communication with subject specialists. We

need a way of identifying and eliciting the

information needed to obtain a clear understanding

of the influences on a project. Ultimately some form

of online collaborative development of ontologically

enhanced CCMs would be desirable, but to achieve

this we need ways of establishing ontological

commitment, that is, agreement on the terminology

to be used and the way that a project is to be viewed.

Established methods (see Al-Shehab et al 2005)

will be used to create preliminary CCMs capturing

overall project objectives, environmental factors and

policies. A ‘straw man’ Project World ontology will

be created from existing sources such as PROMONT

(Abels et al 2006) which will describe the generic

classes of object that would be expected in the

context of a project. Classes in this ontology will be

mapped to concepts in the CCM ontology. Chauvin

et al (2007) describe a way of enriching CCMs with

contextual information which provides a basis for

this work. The initial prototype will be evaluated and

modified through a series of test-analyse-modify

iterations. Once again populating the resulting model

with data from real world projects will form the

basis of the validation. This would implement the

class to instance transition in Figure 3.

2.4 To what Degree can Ontological

Models usefully Support Tools and

Operations at the Level of Practical

Project Planning and

Management?

The research questions above address the feasibility

of representing processes and what are effectively

ontologically-enhanced causal maps as ontological

models. It remains to establish the actual usefulness

of these. We need to know if it is possible to adapt

existing tools and techniques, or develop new ones,

that use knowledge captured in the ontologies to

support typical project planning and control tasks.

The extent to which guidance on decision-making –

for example the selection of methods to execute the

project – can be provided needs to be assessed.

The starting point for this is the identification of

tasks and key decisions involved in planning a

project, initially by examining existing practitioner

and academic literature. Recognised guidelines

supporting project method configuration need to be

analysed to assess the extent to which the knowledge

held in the project ontologies, in conjunction with

heuristics, can offer useful advice the optimal

configuration of an Agile method for a particular

project. Karlsson’s and Ågerfalk’s (2009) work

which uses goal-modelling as a basis for Agile

method configuration provides a guidance on a

possible way forward. We need however to be aware

that each development project that claims to be

guided by a particular process model, may interpret

the model in a different way or may modify it to fit a

particular context.

3 CONCLUSIONS

This proposal is not just an engineering project to

create a prototype artefact that might then be

‘commercialised’. Rather it is the development and

integration of a set of interlocking techniques –

many of which already exist although some are in a

rudimentary or fragmentary form – and, where

appropriate, supporting prototype tools.

It can be envisaged that a potential user might

have difficulty simply plugging in such a tool and

then trying to use it effectively. The use of the

techniques and tools by a wider community would

depend on the acquisition of means-end reasoning

skills. It could be argued that the promotion of such

skills would itself be beneficial in many fields

including planning.

Some researchers will find that the ontological

modelling of processes by itself can be usefully

applied to a broad range of process scenarios beyond

the Agile development methods that are the subject

of the proposed work. The practical application of

contextualised CCMs is of use to researchers in

broad range of social science and other applications

who wish to document their theories in a structured

way which is sensitive to context. Software

engineers may be particularly interested in the

greater opportunities for informed method tailoring

that this work promises.

ACKNOWLEDGEMENTS

The development of many of the ideas in this paper

were developed as part of the University of Brighton

ProposalNet initiative. We would like to thank

ONTOLOGICAL MODELLING TO SUPPORT THE PLANNING OF IS DEVELOPMENT PROCESSES - A Position

Paper

323

fellow participants and the facilitation of the events

by Professors Jacqueline Reilly and Howard Rush.

REFERENCES

Abdel-Hamid T. K. 1988. Understanding the '90% syndrome'

in software project management: a simulation-based case

study, in Journal of Systems and Software. vol. 8, pp.

319-330.

Abels S., Ahlemann F., Hausman K., and. Strickman J. 2006.

PROMONT - A project management ontology as a

reference for virtual project organizations," in OTM

Workshops, pp. 813-823.

Al-Shehab A 2007 Causal and cognitive mapping methods

for the identification of risk in information development

projects: PhD dissertation. Brighton: University of

Brighton.

Al-Shehab A., Hughes R. T., and Winstanley G. 2005

Facilitating organisational learning through causal

mapping techniques in IS/IT project risk management .

Lecture Notes in Computer Science Lecture vol.

3782/2005 145-154,Springer-Berlin, pp. 145-154.

Al-Shehab A., Hughes R. T., and Winstanley G 2006

CorMod: a causal mapping approach to identifying

project development risk. Proceedings of European and

Mediterranean Conference on Information Systems

(EMCIS).

Axelrod R. 1976 Structure of decisions: the cognitive maps of

political elites. Princeton NJ: Princeton University Press,.

Fitzgerald B., Russo N. L., and O'Kane T. 2003 Software

development method tailoring at Motorola, in

Communications of the ACM. ACM, vol. 46: pp. 65-70.

Babar A., Cox K., Tosic V., Bleistein S. and Verner, J. 2008

Integrating B-SCP and MAP to manage the evolution of

strategic IT requirements. Information and Software

Technology. Elsevier. vol. 50. pp. 815-831.

Beck K. and Andres C. 2005. Extreme programming

explained: embrace change 2nd ed. Boston, MA, USA:

Addison-Wesley

Bleistein S., Aurum A., Cox K., and Ray P. (2004). Strategy-

oriented alignment in requirements engineering: linking

business strategy to requirements of e-business using the

SOARE approach. Journal of Research and Practice in

Information Technology. vol. 36. no. 4. pp. 259-76.

Bleistein S., Cox K., and Verner J. 2006a Validating strategic

alignment of organizational IT requirements using goal

modelling and problem diagrams. Journal of Systems and

Software. vol. 79. no. 3. pp. 362-78.

Bleistein S., Cox K., Verner J., and Phalp K. 2006b B-SCP: a

requirements analysis framework for validating strategic

alignment of organizational IT based on strategy, context

and process, in Information and Software Technology.

vol. 48. no. 9. pp. 846-868.

Boehm, B. W. 1991. Software risk management:

principles and practices. IEEE Software. : IEEE vol. 1 pp.

32-41.

Chauvin L., Genest D., Loiseau S.2007 Le modèle des cartes

cognitives contextuelles Annals du LAMSADE no 8

pp 285-292

De Nicola A., Missikoff M., and Navigli R.2009 A software

engineering approach to ontology building. Information

Systems. vol. 34: Elsevier, pp. 258-275.

DSDM, 2007. DSDM Atern Pocket Book Ashford: DSDM

Consortium

Eden C. 2004 Analyzing cognitive maps to help structure

issues or problems. European Journal of Operational

Research. vol. 159, pp. 673-686.

Gil Y. and Blythe J. 2000 PLANET: A shareable and

reusable ontology for representing plans. Proceedings of

AAAI Workshop on representational issues for real-world

planning systems at 17th National Conference on

Artificial Intelligence Austin, Texas, USA.

Gruber T. R. 1995. Toward principles for the design of

ontologies used in knowledge sharing. Int. J. Human-

Computer Studies. Academic Press, vol. 43 pp. 907-928.

Hendy, J., Reeves B. C., Fulop N., Hutchings A., and

Masseria C. 2005. Challenges to implementing the

national programme for information technology (NPfIT):

a qualitative study.British Medical Journal. Britisth

Medical Association vol. 331 pp. 331-4.

Höfferer, P. 2007 Achieving business process interoperability

using metamodels and ontologies , in H. Sterle, J. Schelp,

R. Winter, eds Proceedings of the 15

European

conference on information systems (ECIS2007) pp 1620-

31 Univerity of St. Gallen, Switzerland

Hughes R. T., Al-Shehab A., and Winstanley G.

2006.Obstacles to the modelling of the causes of project

success and failure systems. European Conference on

Research Methods in Business and Management. Trinity

College, Dublin: MCIL,

Hughes, R.T. 2010. Project management process ontologies: a

proof of concept. Proceedings of the 15

annual

conference of the UK Academy of Information Systems,

Oriel College, Oxford 23-24 March

Karlsson F. and Ågerfalk P. 2009 Exploring agile values in

method configuration . European Journal of Information

Systems. vol. 18, Palgrave Macmillan,.pp. 300-316

Kim J. and Gil Y. 2001. Knowledge analysis on process

models. Proceedings of International Joint Conference

on Artificial Intelligence pp. 935-942.

Marshall N. 2009. Cognitive and practice-based theories of

organisational knowledge and learning: incompatible or

complementary? Management Learning. Sage

Publications. vol. 39. no. 4. pp. 413-435.

Montibeller G., Belton V., Ackermann F., and. Ensslin L.

2007. Reasoning maps for decision aid: an integrated

approach for problem-structuring and multi-criteria

evaluation Journal of the Operational Research Society.

vol. 59, pp. 575-89.

Pearl, J. 2000 Causality: Models, Reasoning and Inference.

Cambridge, Cambridge University Press.

Schwaber K. 2007 The Enterprise and Scrum. Redmond,

Washington: Microsoft Press.

Stach W. and Kurgan L. 2004.Modelling software

development projects using fuzzy cognitive maps.

Proceedings of 4th ASERC Workshop on quantitative and

software engineering Banff AB

Winter M., Smith C., Morris P., and Cicmil S. 2006.

Directions for future research in project management: the

main findings of a UK government-funded research

network, International Journal of Project Management.

Elsevier, vol. 24 pp. 638-649.

KEOD 2010 - International Conference on Knowledge Engineering and Ontology Development

324