IMPROVING REQUIREMENTS ANALYSIS

Rigourous Problem Analysis and Formulation with Coloured Cognitive Maps

John R. Venable

School of Information Systems, Curtin University of Technolgoy, Perth, Australia

Keywords: Problem Formulation, Requirements Analysis, Information Systems Development, Cognitive Mapping,

Coloured Cognitive Mapping.

Abstract: This position paper argues that methodical and rigourous attention to problem formulation should be an

essential part of requirements analysis and proposes a method for modelling problems and potential

solutions. Currently, most IS Development Methodologies ignore the issue of problem formulation.

Furthermore, in practice, most IS development projects ignore or pay little attention to this issue. In this

paper we argue that the resulting lack of proper problem analysis and formulation is a major cause of IS

failures. In place of this lack of attention, this paper proposes and reports on research in progress on the

development and evaluation of a new technique, Coloured Cognitive Maps (CCM), for use in problem

formulation at the early stages of Information Systems Development. The notation of coloured cognitive

maps and a procedure to use them for problem analysis and solution derivation are briefly described. Initial

anecdotal results in employing CCM by students and practitioners are briefly described.

1 INTRODUCTION

The IS/IT field is the poor track record of system

development and introduction. Information Systems

Development (ISD) projects commonly (1) are late

and over budget, (2) deliver a system that doesn’t

meet requirements, cannot be used as intended, is

hard to use, or is completely unusable, or (3) fail to

deliver a system altogether (project never

completed). High failure rates in ISD are widely

recognised.

Methodological solutions proposed for these

problems address various perceived causes. Design

and software engineering methodologies address

technical difficulties in making complex systems

work properly to insure that requirements are

properly translated into designs and

implementations. Requirements analysis

methodologies precisely state requirements and

ensure that they correct and consistent. Project

management methodologies aim to overcome poor

estimation of time and costs, scope creep, and poor

project planning and control. To varying degrees, all

of the above try to address changing needs during

development.

However, we assert that many failures are due to

poor practices at the very front end of development,

when stakeholders are (or should be) grappling with

a problematic situation in order to decide what

problems are to be solved and what sort of IS would

contribute to that solution. This lack of attention to

problems and the organisational situation leads to

poor understanding of the problem(s), often to

solving the wrong problem(s) (also known as an

error of the third kind, Mitroff and ), and commonly

to poor acceptance and adoption of system(s), a

major form of IS failure. None of the above types of

methodologies address this issue.

It is the contention of this position paper that

absent or cursory problem analysis and formulation

is a key weakness in system development. Problems

are sometimes completely unstated. Often, they are

only weakly examined. Typically, problem diagnosis

and formulation are poorly performed, if at all.

Commonly, solutions are proposed in the form of a

system request without any examination of the

problem situation whatsoever. Furthermore,

agreement about problems may not be sought.

Proposed solutions are then accepted for IS

Development without consideration of the proposed

systems’ effectiveness in solving the unstated

466

R. Venable J. (2008).

IMPROVING REQUIREMENTS ANALYSIS - Rigourous Problem Analysis and Formulation with Coloured Cognitive Maps.

In Proceedings of the Tenth International Conference on Enterprise Information Systems - ISAS, pages 466-470

DOI: 10.5220/0001727204660470

 SciTePress

problem(s) or addressing the needs of various

different stakeholders.

Instead, this paper proposes that methodical and

rigourous attention needs to be paid to developing

problem formulations that are clear, correct,

properly scoped and prioritised among other

problems, and agreed by relevant stakeholders.

Further, ISD processes should then incorporate

effective transition from problem definition to

solution generation, choice, and implementation (via

detailed IS development and installation). While

some important work has been done in this area

(e.g., Soft Systems Methodology, Checkland, 1981,

Checkland and Scholes, 1990 and Multiview, Wood-

Harper et al., 1985, Avison & Wood-Harper 1990 –

see below), it is still an area of perennial weakness

and in need of much more attention.

2 ADDRESSING PROBLEMS

WITH PROBLEM

FORMULATION IN ISD

Various methods and techniques have been proposed

that address problem formulation.

An important, relevant method is Soft Systems

Methodology (SSM) (Checkland 1981, Checkland &

Scholes 1990, Checkland and Holwell, 1998). SSM

is a general problem solving method especially for

use where there are differences among stakeholders

about their understandings and goals. It can be

applied to the formulation and agreement about any

kind of problem to be solved and the design of any

kind of solution. It is not specifically designed to be

applied to Information Systems solutions, although

Checkland and Holwell (1998) do so.

SSM incorporates a number of useful techniques,

including Rich Pictures, CATWOE Criteria, Root

Definitions, and Conceptual Models. Rich Pictures

are especially used to model and explore a

problematic situation.

The Multiview methodology (Wood-Harper et al.

1985, Avison & Wood-Harper 1990) incorporates

Rich Pictures. Multiview can be described as an

eclectic method, which draws on a number of other

methods and their techniques and builds those into a

coherent, overall approach to systems development.

Multiview incorporates techniques from SSM,

Human-Computer Interface (HCI) design methods,

and Structured Analysis and Design methods, among

others (Wood-Harper et al. 1985, Avison & Wood-

Harper 1990).

Mathiassen et al. (2000) incorporate Rich

Pictures into their Object-Oriented Analysis and

Design methodology. They also incorporate revised

versions of CATWOE and Root Definitions, which

they call FACTOR and System Definitions. These

revised versions are specifically tuned for modelling

the concerns of scoping and defining information

systems to be designed and built.

However, none of the above methods have a

technique for formally and rigourously modelling

problems and their causes. A form or modelling that

includes causal analysis is needed. Fortunately, other

work addresses this need.

Problematiques (Roberts 1994) are a diagram for

causal analysis that can be used to explore a problem

or group of problems, their causes, and their

consequences. The technique consists of nodes and

links, where the nodes are succinct statements of

parts of problems and the links being arrows from a

cause to a consequence of that cause. However,

problematiques are somewhat limited in their

semantics compared to Cognitive Maps, as described

in the next paragraph.

Cognitive Maps (Eden 1988, Eden & Ackermann

2001, Ackermann & Eden 2001) were developed

primarily for strategy development, not for problem

analysis. A key element of cognitive maps is that the

text in a node may have two ‘poles’, a primary pole

which is the content, and a secondary pole, which

provides more meaning through contrast (e.g.

“increased sales … (as opposed to) continuing poor

sales”). While this technique was not developed for

problem analysis, the notation can be used to explore

conceptualisations of problems and solutions at the

front end of ISD.

Venable (2005) proposed a new form of

cognitive maps, called Coloured Cognitive Maps

(CCMs) to address the above issue and added

various enhancements, including: (1) a

conceptualisation of two forms of problem

statements and corresponding forms of coloured

cognitive maps: problems as difficulties, which

focus on the current undesirable or problematic

situation and problems as solutions, which focus on

statements of some different, desirable future

situation, (2) a procedure for straightforward

conversion between these two forms of cognitive

maps, (3) colouring of nodes to indicate desirability

or undesirability, and (4) an overall process for

problem analysis with cognitive maps.

CCM supports rigourous problem diagnosis and

formulation through drawing a model of the problem

as difficulties. CCM also supports derivation of

alternative problem solutions, which could include

IMPROVING REQUIREMENTS ANALYSIS - Rigourous Problem Analysis and Formulation with Coloured Cognitive

Maps

467

IS solutions and their features. Through the use of

the CCM method, problem solutions and their

features are rigourously related back to the original

problem. The method supports contrasting of

different solutions as to how they apply to solving

the different aspects of the problem(s) to be solved.

Therefore, it also supports contrasting of alternatives

for decision making.

3 COLOURED COGNITIVE

MAPPING FOR

REQUIREMENTS ANALYSIS

Based on the above advantages, we propose that

Coloured Cognitive Mapping (Venable, 2005) could

be used effectively to support methodical and

rigourous problem formulation at the front end of

Requirements Analysis in ISD. We will now briefly

summarise the notation and mapping process for

using Coloured Cognitive Maps (Venable 2005).

3.1 Notation

Two symbols are used in coloured cognitive maps:

nodes and arrows (see figure 1). Nodes are drawn

with rounded rectangles, ovals or some other

convenient symbol and represent some aspect of a

problem. Text is placed within each node, which

captures the meaning of the node. The text in the

node can also be split into two parts or poles, which

are separated by an ellipsis symbol (“…”).

In coloured cognitive maps, the nodes are

coloured to indicate whether the node represents

something that is desirable or something that is

undesirable. Green nodes represent desirable

circumstances and red nodes indicate undesirable

circumstances. Generally, one of the poles in a node

should be desirable and the other one undesirable,

with the colour corresponding to the primary pole

(the text that comes first). Where colour cannot be

used, another indication is needed, such as bold

print, darker lines, darker shading, or a different

node shape for undesirable nodes, as shown in figure

1 below.

Nodes are connected to each other with arrows.

Arrows represent causality between the nodes, i.e.

the node at the tail of the arrow causes the node at

the head of the arrow. Table 1 shows some further

synonyms for the meaning of causality. Note that the

arrows do not mean flow of information or goods

and should never be used as such.

Node:

- Goal, activity, problem,

cause, implication, etc.

- Poles separated by ellipsis,

- Red/bold = undesirable, Green = desirable

Arrow:

- Causal or contributory

- Plus sign or minus sign (plus assumed)

Figure 1: Coloured Cognitive Mapping Notation.

Table 1: Synonyms for the meaning of the arrow.

n arrow with

a plus (or no)

sign means

n arrow with a minus sign

means

Causes Causes the opposite pole

Implies Implies the opposite pole

Enhances Reduces

Contributes to Detracts from

Increases Decreases

llows Disallows

Enables Prevents

Arrows may optionally have plus or minus signs

attached to them. If there is no sign, a plus sign is

assumed. If a minus sign is attached, the node at the

tail prevents (rather than causes) the node at the head

or causes its opposite pole. Table 1 also shows

alternative meanings for the arrow when it has a

minus sign attached.

3.2 A Procedure for Analysing

Problems with Cognitive Maps

The coloured cognitive mapping procedure is

divided into three stages (see figure 2). First is

Problem Diagnosis, in which a cognitive map is

developed of the problem as difficulties. The second

stage is CCM Conversion, which converts the

cognitive map of the problem as difficulties into a

cognitive map of the problem as solutions. The

resulting cognitive map is incomplete, but a basis for

progressing in the third stage. The third and final

stage is Solution Derivation, in which the cognitive

Give

Poor …

Good

+ or -

Provide

Good … Poor

Service

ICEIS 2008 - International Conference on Enterprise Information Systems

468

Figure 2: Procedure for Problem Analysis with Cognitive Maps.

Figure 3: Example Conversion to an Initial Cognitive Map of a Problem as Solutions.

map of the problem as solutions is expanded with

various candidate or potential solutions.

The goal of problem diagnosis is to obtain a clear

(and hopefully agreed) understanding of the causes

and consequences of the problematic situation.

Solving a problem effectively requires that the

problem solver(s) develop a rich understanding of

the problematic situation before proceeding. The

problem solvers need to understand what is

undesirable about a problematic situation, why it is

problematic to the stakeholders, and what the causes

of the problem are – i.e. what things allow the

undesirable circumstances to exist. Note that

cognitive maps of problems as difficulties will

primarily have nodes that are undesirable (coloured

red, bolded, and/or oval shaped), but some nodes

will likely be desirable ones. As they say, “Every

cloud has a silver lining.”

Cognitive Map Conversion is the process of

converting a CCM of the problem as difficulties into

an initial CCM of the problem as solutions. Figure 3

below shows an example. This step is a (nearly

completely) mechanical process of changing every

node in the CCM of the problem as difficulties from

either undesirable to desirable or desirable to

undesirable. The colour of every node is changed

and the text is changed by switching the poles and

rewording so that it makes sense. The example is of

Problem

Diagnosis:

Cognitive Mapping

of a Problem as

Difficulties

CCM Conversion:

Convert from CCM of

Problem as Difficulties

into CCM of Problem

as Solutions

Solution

Derivation: Cognitive

Mapping of a Problem

as Solutions

Poor … good

customer

service

Work

done

poorly …

well

Insufficient

… enough

time

Too

much …

right amount

of work

Lower ….

normal repeat

business

Provide

enough …

insufficient time

Reduce

workload … too

much work

Do work well

… poorly

Increase …

lower repeat

business

Improve …

poor customer

service

Reverse

poles of

problems,

symptoms,

implications, or

causes

IMPROVING REQUIREMENTS ANALYSIS - Rigourous Problem Analysis and Formulation with Coloured Cognitive

Maps

469

course extremely simplified compared to a normal

problematic situation.

In Solution Derivation, the initial CCM of a

problem as solutions is enhanced to explore different

potential solutions and the consequences if someone

were to implement one or more of the potential

solutions. Solutions cause the reduction or

elimination of causes and therefore indirectly solve

or alleviate problems.

In the case of using CCM to support

Requirements Analysis for IS Development,

Solution Derivation would add nodes pertaining to

IS solutions to the problems at hand. In the example

given in figure 3, nodes would be added showing

means for reducing workload or providing enough

time. One can imagine various different IS-based

means for reducing workload. Placing them on the

diagram allows one to visualise what problems will

be solved or reduced by an IS solution, and what

consequences there will be of introducing an IS

solution. Of course, one must also ask what other

consequences there might be, some of which might

be undesirable.

One can also consider alternative candidate

solutions and analyse tradeoffs or one alternative vs.

others through the use of the CCM diagram of the

problem as solutions.

Thus, using CCM can be helpful for rigourously

and systematically analysing a problematic situation

and considering potential IS (or other) solutions.

4 RESEARCH PROGRESS

The author has taught the CCM technique to one of

the principal owners/operators of a small company

engaged in IT infrastructure and systems

development, who have introduced CCM into their

organisational practice. The owners report that use

of the technique is straightforward and easily learned

and adopted by their staff. They also report

significant (ca. 20%) drops in project completion

time, due to fewer problems being encountered

during projects that use the technique.

While the above initial result is encouraging,

much more rigourous and detailed evaluation of the

technique is needed. We plan to undertake action

research so that we develop a richer understanding

of CCM in use.

5 CONCLUSION

This paper asserts the importance of methodically

and rigourously analysing and formulating the

problem(s) to be solved when analysing

requirements for solution via IS Development –

before deciding on a solution and trying to

implement it. In particular we have proposed that

Coloured Cognitive Mapping could be usefully

employed to improve the efficiency, effectiveness

and efficacy of ISD.

REFERENCES

Ackermann, F. and C. Eden, 2001. SODA – Journey

Making and Mapping in Practice, Chapter 3 in

Rational Analysis for a Problematic World Revisited.

J. Rosenhead & J. Mingers (eds.), Wiley, Chichester,

UK.

Avison, D. E. and Wood-Harper, A. T., 1990. Multiview:

An Exploration in Information Systems Development.

McGraw-Hill, New York City, USA.

Checkland, P. B., 1981. Systems Thinking, Systems

Practice. Wiley, Chichester, UK.

Checkland, P. B. and S. Holwell, 1998. Information,

Systems, and Information Systems. Wiley, Chichester,

UK.

Checkland, P. B. and J. Scholes, 1990. Soft Systems

Methodology in Action. Wiley, Chichester, UK.

Eden, C., 1988. Cognitive Mapping, European Journal of

Operational Research, Vol. 36, pp. 1 – 13.

Eden, C. and F. Ackermann, 2001. SODA – The

Principles, Chapter 2 in Rational Analysis for a

Problematic World Revisited. J. Rosenhead & J.

Mingers (eds.), Wiley, Chichester, UK.

Mathiassen, L., Munk-Madsen, A., Nielsen, P. A., and

Stage, J., 2000. Object Oriented Analysis & Design.

Marko Publishing, Aalborg, Denmark, http://marko-

publishers.com.

Roberts, P., 1994. Systems and the Problematique.

Futures, Vol. 26, No. 7, September, pp. 730-740.

Venable, J. R., 2005. Coloured Cognitive Maps for

Modelling Decision Contexts. Proceedings of the First

International Workshop on Context Modeling and

Decision Support, Paris, France, 5 July 2005, CEUR

Workshop Proceedings, ISSN 1613-0073, online

http://CEUR-WS.org/Vol-144/03_venable.pdf

(accessed 30 October 2005).

Wood-Harper, A. T., L. Antill, and D. E. Avison, 1985.

Information Systems Definition: The Multiview

Approach. Blackwell Scientific Publications, Ltd.,

Oxford, UK.

ICEIS 2008 - International Conference on Enterprise Information Systems

470