A Taxonomy of Current Approaches to Systems

Analysis

A. T. Wood-Harper
School of Computing Studies and Accountancy, University of East Anglia, Norwich NR4 7TJ, UK
G. Fitzgerald
School of Mathematics, Statistics and Computing, Thames Polytechnic, London SE18, UK

In the last few years a number of systems analysis approaches and methodologies have arisen. There is confusion in the

Downloaded from https://academic.oup.com/comjnl/article/25/1/12/527282 by guest on 29 January 2023

choice of using these approaches and in this paper the authors argue that methodologies are in fact different
perceptions in thefieldof systems analysis. Each methodology is discussed and classified in terms of paradigms,
conceptual models and objectives.

Since the identification and propounding of GST many

INTRODUCTION people have tried to apply the theory to the solving of
practical problems.3 This attempt has been notably
The discipline of systems analysis is still very young and unsuccessful.4'5 The reason for this lack of success is that
in common with most other emerging disciplines it the very generality of GST makes it difficult to use and to
occasionally enters periods of radical self examination develop a methodological solution; and where a solution
and re-thinking. The authors feel that we are in the midst is arrived at it is often one which requires a revolution to
of such a phase at present; new ideas abound, arguments implement. It is not a process which often recommends
rage, and the development of technology is a powerful small incremental changes but one which more usually
impetus to the re-examination of ideas. results in the complete reassessment of structures, roles
The reason for the current turmoil in systems analysis and behaviour. Indeed it has been called by Popper,
is the emergence over the past few years of a number of Utopian Engineering, i.e. implying redesign of the whole
new approaches or methodologies.1 These approaches fabric of society.6 This may well be overstating the case
have generally originated as academic ideas and been but it is certainly unlikely to be an approach to appeal to
taken up and modified in the practising world of systems the systems analyst who has an inbuilt leaning to the
analysis. Thus there exists a confusing array of ap- pragmatic. A systems analyst who recommends giving
proaches. It is the purpose of this paper to examine some the product away free as a revolutionary solution to the
of the more fundamental approaches and to attempt to problem of invoicing the customers would probably not
classify them. It is the authors' view that the approaches last long in his chosen profession. Although he may argue
are not simple alternatives, but that they seek to do that his is the correct solution from society's point of
different things. view.
The authors have identified six major approaches to Thus the systems analyst considers the application of
systems analysis: (i) General Systems Theory Approach; GST too impractical and wide ranging for his purposes,
(ii) Human Activity Systems Approach; (iii) Participa- after all he has terms of reference within which he must
tive (Socio technical) Approach; (iv) Traditional (NCC, work!
etc.) Approach; (v) Data Analysis Approach; (vi) The counter argument is of course that these restric-
Structured Systems (Functional) Approach. Except for tions impose conditions on the systems analyst which are
the General Systems Theory Approach they are all used the very reason why he is actually not very successful in
to some extent in the industry today. General Systems his problem solving. He can only tune the engine when
Theory is included as an approach because of its perhaps what is needed is the redesign of the transport
important influence on systems thinking in general and system.
because of the contribution it has made to almost all the
other identified approaches. A number of people have sought to come to terms with
this problem and make GST more practical for problem
solving. They have striven to convert GST into a practical
methodology byfirstly,breaking down the process into a
GENERAL SYSTEMS THEORY APPROACH number of defined steps to be followed and secondly,
seeking to limit the range of alternative solutions by
General systems theory (GST) deriving from the work of introducing notions such as the identification of certain
L. von Bertalanffy has had much influence,2 but has not value systems within which the problems must be set.
really been concerned with practical systems analysis. It Perhaps the best known of these attempts is Check-
is more an attempt to come to terms with and understand land's Methodology for real world problem solving which
the nature of systems. It is theoretical model building for here we classify as the Human Activity Systems
the interpretation of complex and diverse systems. Approach.7
CCC-O01O-4620/82/0O25-0O12 $02.50
12 THE COMPUTER JOURNAL, VOL. 25, NO. 1,1982 © Heyden & Son Ltd, 1982
 A TAXONOMY OF CURRENT APPROACHES TO SYSTEMS ANALYSIS

HUMAN ACTIVITY SYSTEMS APPROACH the question is to what extent this is done. Even in
Mumford's scheme the users design only within the
Checkland has tackled some of these problems in the constraint of the assumption that a computer is to be
design of his methodology. First he attempts to provide used in the solution.10 What the user may design is the
a methodology for solving problems which are not precise work situation into which the computer is to be brought
or of well denned structure. He argues that these are the and he helps to determine how the man/machine
sort of problems that need to be dealt with in the interface is organized. The authors do not criticize this
environment or organizations and management; they are approach, indeed it seems to be a beneficial and
what is known as fuzzy or soft problems. Second he seeks praiseworthy method. However, one should be clear that
a solution to the problem which he acknowledges is only it is not a systems solving methodology. It is an attempt
one of perhaps many possible alternative solutions. This to achieve smooth implementation of a system, and
makes the methodology very much more practical for it successful operation of the computer in the workplace. A
is often found in reality to be very difficult to define comparison with Checkland's system reveals that it is
objectives and usually very contentious to try. not problem solving at the same level. It is concerned
with solving implementation problems.
It is not the purpose of this paper to describe the
methodology but simply to categorize it. Checkland has

Downloaded from https://academic.oup.com/comjnl/article/25/1/12/527282 by guest on 29 January 2023

developed a methodology based on the notions of GST THE TRADITIONAL APPROACH
but modified in some significant ways to be practical in
the real world, and it has been used in many cases with Perhaps the most common approach to Systems Analysis
evident success.8 The methodology can be categorized as is that advocated by the National Computer Centre
GST approach, with emphasis on unstructured problems (NCC), often termed the traditional approach.J' In some
in complex situations. It generates understanding of the form or other, this is the approach adopted by the
environment and leads to possible structural, procedural, majority of systems analysts. It contains the well known
attitudinal or environmental change that will improve phases of investigation, analysis, design and implemen-
the situation. The change is then implemented. This of tation. It is an approach based on the idea that there
course is the crux of the situation because it is here that exists a problem which can be solved by the application
the work of the Systems Analyst normally starts and yet of a computer. Each application is considered separately
this is the point at which Checkland's methodology is and the problem resolved by the design of an optimal
least concerned. The philosophy behind Checkland's subsystem. The optimization is achieved by investigation
methodology is that analysis and understanding of the of the existing system in terms of the functions that it
situation and environment within which the problem lies performs. The analysis is the distillation of the results of
leads to a possible solution. investigation into a documented form and design is
This approach to systems analysis is therefore not in achieved by consideration of the required outputs and
competition with most other approaches, and indeed designing the inputs,filesand processes to achieve those
could be regarded as complementary to them. It tries to outputs. A tried and trusted formula, the result of which
explore, identify and structure the problem before has been the piecemeal computerization of manual
'solving' it. subsystems. What has usually not been solved is the
problem that led to the demand for the introduction of a
computer in thefirstplace. At best, time has been bought
by making the manual system more efficient. This has
THE PARTICIPATIVE APPROACH been fine for particular circumstances, for example
payroll, invoicing and billing. Where it has proved
This approach is most commonly associated with inadequate is in more complex areas, such as accounting
Mumford, Land and Hawgood and stresses the impor- systems or information systems.
tance of the user in systems analysis.9 Many other people
have of course argued the importance of people and the
human element in systems but Mumford takes this a step DATA ANALYSIS APPROACH ~
further and argues that the users ought to be involved in
the design of a system in which they participate or The developing methodology of data analysis is based on
preferably actually design it themselves. This approach the philosophy that the fundamental building blocks of
seems to have generated a lot of interest recently. Possibly a system are data.12 The hypothesis is that if we can
because it directs itself to a very practical problem: that classify and identify the set of data elements (entities and
of the rejection of the system by the people who use it, or attributes) that exists within a particular situation then
if not outright rejection then much reluctance and we have identified the true nature of a system. The use to
slowness in accepting the new system. Such factors have which that particular system can be put can change or be
often been found to handicap a system fatally. An changed, but the underlying nature of the system remains
extreme example is the rejection of new technology in unchanged because the data is static, or much less likely
Fleet Street. Clearly it is not a revolutionary approach; to change than the function or processes applied to it.
it does not seek ways offindingradically new solutions to This means that one can define the system without
business problems. It is a way of making sure the defining the individual applications that need to use it—
implementation stage of a project is acceptable to the obviously a very useful concept. In addition if relation-
users by letting them design the user/machine interface. ships between the data are defined then one has, in effect,
It is revolutionary only in the idea that it should be the developed a model for the system which can be validated
users who do the designing. Every Systems Analyst will before implementation. Data analysis can be seen,
say that he already takes account of the users' wishes, but therefore, as a 'neutral' way to come to an agreed

© Heyden & Son Ltd, 1982 THE COMPUTER JOURNAL, VOL 25, NO. 1,1982 13
 A. T. WOOD-HARPER AND G. FITZGERALD

understanding about the nature of an organization. It fits proaches to systems analysis. This diversity is a reflection
in very nicely with (indeed it has developed from) the of the differing viewpoints embodied in the approaches.
concept of the data base. We seek to map the system as The differing viewpoints arise because it is so difficult to
it exists and then later solve our problems easily by observe objectively a system that exists 'out there' in the
functional applications on the data base.' 3 Data analysis real world. Our perceptions of that reality are different
however is not orientated to problem solving in itself. and subjective and it is these different perceptions that
What it seeks to do is clearly to define the basis, in terms lead to the differing approaches. It is the authors'
of data and relations of the system in which the problem contention that the differing approaches can be better
exists. Ideally it seeks to do this on a company or understood by an examination of the paradigms, concep-
organization-wide basis. A clear mapping and under- tual models and their objectives, that the approaches are
standing of the way an organization works is indeed very based upon.
useful. If the mapping or logical data model forms the
basis of all the subsequent applications the worry is that
this base will in time prove inadequate because its source PARADIGMS
was the existing system. Data analysis concentrates on
the investigation and analysis phases of Systems Analysis First the authors feel that a major contribution to the
and says very little about system design or problem

Downloaded from https://academic.oup.com/comjnl/article/25/1/12/527282 by guest on 29 January 2023

understanding of the differing approaches can be
solving. Yet the inference is that successful data analysis achieved by examining the underlying paradigms on
leads to successful design. We need to question this. Data which the approaches are based. Paradigm is used here
analysis may not solve underlying problems that the in the sense identified by Kuhn as a specific way of
organization might have. Indeed it may have 'captured' thinking about problems encompassing a set of achieve-
the existing problems into the new data model, and made ments which are acknowledged as the foundation of
them even more difficult to solve in the future. further practice.16 A paradigm is usually regarded as
This is not to say that data analysis is not a powerful subject free, in that it may apply to a number of problems
tool to help in clarification, understanding and commu- regardless of their specific content.
nication, but that it is not necessarily going to provide a The authors identified two basic paradigms to be of
solution to a business problem. interest in this context, firstly the science paradigm and
secondly the systems paradigm.
The science paradigm embodies the scientific method
STRUCTURED SYSTEMS ANALYSIS which has led to the development of the 'hard' sciences
APPROACH and without becoming ensnared in the controversies that
surround discussions of the philosophy of science we opt
Structured Systems Analysis is described perhaps most for Checkland's description of the science paradigm as
eloquently by de Marco14 and Gane and Sarsons.15 It being a learning system characterized by reductionism,
appears that this approach or method is one that is repeatability and refutation.17
beginning to be used in an increasing number of 'We reduce the complexity of the variety of the real
organizations. It is an attempt to solve some of the world in experiments whose results are validated by their
problems of the traditional approach, such as the repeatability and we build knowledge by the refutation
departmental/subsystem viewpoint, the problems of co- of hypotheses.'
ordinating a large team of analysts, and the problem of The systems paradigm is not so easily condensed
complexity of a large organization or system. The because there is no unique acceptance of what constitutes
approach provides new tools for analysis and documen- the systems paradigm. What is clear is that the origin of
tation, such as data flow diagrams, the concepts of data the systems paradigm derives from the concern of some
dictionaries and structured English. The use of these that the science paradigm was proving inadequate when
enables the clear documentation of existing systems and faced with living systems and particularly human activity
proposed new systems. It suggests methods of analysis systems. These systems exhibit openness, low separability
but none for design. Indeed when de Marco comes to the and high interdependence. Such systems, if reduced in
design of the new system he states clearly that this is not the method of the science paradigm, lose their meaning
one of his aims. He says 'It is at this time that the Systems and the ability to be explained. This is so because such
Analyst exercises his experience and imagination to systems are more than the addition of their individual
come up with new systems concepts. This is where he components because, as Checkland explains, the parts,
'invents' the new system. I won't tell you how to go about when aggregated, display emergent properties. Thus the
this—I have restricted myself to teaching new tools for systems paradigm is a holistic one.
analysis, and no tool that I could think of would aid the Let us now examine our six identified approaches in
invention process, when you have come up with your terms of the paradigms which they adopt. First, General
invention, however, the tools of Structured Analysis are Systems Theory is by its nature wholly based on the
exceedingly useful for documenting it and trying it out.' systems paradigm. Its interdisciplinary nature; its con-
It is this aspect that is often forgotten and it is often cern with value systems and objectives; its analysis of the
thought to be a design methodology. interaction of subsystems; and its resulting ad vocation of
major structural and social change are clearly in the
mainstream of the systems paradigm.
CLASSIFICATION SCHEME Checkland's approach also incorporates the systems
paradigm, by his own admission, and by the concentra-
The preceding discussion perhaps serves to illustrate the tion on open systems, fuzzy problems and purposeful
complexity and diversity of the various identified ap- activity. This, despite the fact that Checkland reduces

14 THE COMPUTER JOURNAL, VOL. 25, NO. 1,1982 © Heyden & Son Ltd, 1982
 A TAXONOMY OF CURRENT APPROACHES TO SYSTEMS ANALYSIS

his area of concern from the global level to that of the the geology and yet another the climate. Each map is a
organizational level, in order to make the methodology subjective representation of reality—the view, just as the
more practical in the real world. conceptual model is implicit in each methodology and is
The third approach incorporating the systems para- the subjective representation of reality to the researcher
digm is the participative approach. Here it is the belief or practitioner in constructing or using a methodology.
in the interaction of the social and the technical General systems theory is based on a model of the
subsystems that leads to an advocation of the participa- world which exhibits pure systems traits. Systems have
tive design philosophy. The work system is analysed for objectives, they are composed of interacting subsystems
variances or weaknesses which prevent the system which exhibit particular behavioural characteristics. The
objectives being realized. These variances are often conceptual models help to achieve a better understanding
discovered at subsystem boundaries, particularly where of the system by identifying the interactions. These
the social and technical subsystems meet. The ideas of conceptual models are constructed from the following
job enrichment and participative design are particular criteria: Abstraction—the mental resolution of the salient
solutions to the more common variances which are features of a systems structure; Congruence—identifi-
encountered. Thus the underlying paradigm for this cation of the problem situation and its solution; Eclec-
approach is argued to be the systems one. tism—the interdisciplinary nature of a model; Syncre-
tism—the admission of different value systems. The latter

Downloaded from https://academic.oup.com/comjnl/article/25/1/12/527282 by guest on 29 January 2023

The traditional approach to systems analysis on the
other hand we believe, embodies the science paradigm. is the most important criterion in structuring the
This comes as a surprise in that it is usually thought to be conceptual models in the General Systems Theory
based on the systems paradigm. The theory of the approach.'8
approach clearly adopts the systems aspect, but we argue The Checkland Model is similar but does not include
that the practice clearly embodies more of the science the notion of definitive objectives. The objectives of a
paradigm. The approach is reductionist in that systems system are by no means clearly or easily defined in
are broken down into their component subsystems and Checkland's model. Differing participants may have
that these subsystems are then optimized and imple- conflicting objectives, thus Checkland avoids objectives
mented. The approach rarely examines the overall system in his model. He introduces the concept of soft or fuzzy
and its interactions, but simply takes a subsystem as it systems in which the conceptual model is generated from
exists and converts it to a computerized form. the encapsulation of a system, based on the actors in the
The data analysis approach would also seem to embody problem situation, the client of the system, the owner of
the science paradigm. Here the observations about the the problem, the transformation process and the system
complex real world of organisations are reduced to the environment.19
study of data. It is postulated that the data are the basic The participative approach has a totally different
building blocks of the organization. conceptual model, somewhat at variance with those
Key elements of data, described as entities and already examined. These models are based primarily on
attributes with the relationships between entities, are people and their needs, and seek to find a 'fit' between
identified. This data model serves to define the organi- the people and their needs and the work environment:
zation or subsystem of the organization, in a time and the organization structure, technology (computers) and
function independent way. The hypothesis is validated the work tasks.
by the participants in the organization being able to The traditional approach to conceptual models is
define the functions they perform in terms of the data functional. The breakdown of the functions and the
model. The systems paradigm is clearly rejected by the optimization of these subfunctions is the basis of the
concentration on the single area of data and the ignoring model. The goal is the optimal functioning of each
of any notions of subsystems interacting with or on the individual subfunction which will produce an efficient
data. and workable computer subsystem.
The structured approach also incorporates the science The manifestation of the model should be achieved in
paradigm. It is concerned with breaking systems down the anatomy of a computerized information system on
into lower and lower levels until each component can be the application processes, integrated files, file mainte-
easily understood. It is also concerned with providing nance, recovery, control, monitoring and information
tools and techniques for this purpose. Thus it can be said retrieval.20
to rely heavily on the scientific method. In the data analysis approach, the conceptual model is
clearly one of data: entities, attributes and relationships.
Whereas in the structured systems approach, the empha-
CONCEPTUAL MODELS sis is more on the functions rather than the data, although
the functions are observed from the viewpoint of the
Having examined the paradigms underlying the various data rather than the viewpoint of any person or
approaches we now turn to the question of their organization. The use of these models concentrates on
conceptual models. using tools and techniques on the real problem in
The conceptual model is defined as a subjective communication within systems analysis.
representation of reality which is implied in each
methodology. To illustrate this definition, it is as if a
number of different people have looked at a piece of land OBJECTIVES
and set about drawing a map of it. The results all might
be very different because of the differing choices made as An examination of the objectives of each of our chosen
to what is of importance. Thus one person might draw a approaches to systems analysis is illuminating. General
map indicating the roads, another the contours, another Systems Theory, Data Analysis, and the structured

© Heyden & Son Ltd, 1982 THE COMPUTER JOURNAL, VOL. 25, NO. 1,1982 15
 A. T. WOOD-HARPER AND G. FITZGERALD

PARADIGM SYSTEMS SCIENCE

MODEL VALUE SYSTEMS FUZZY/SOFT PEOPLE PROCESSES DATA TOOLS/TECHNIQUES

OBJECTIVE ANALYSIS PROBLEM PROBLEM PROBLEM ANALYSIS ANALYSIS

SOLVING SOLVING SOLVING

APPROACH GENERAL HUMAN

I
PARTICIPATIVE
I
TRADITIONAL DATA STRUCTURED
SYSTEMS THEORY ACTIVITY ANALYSIS

Figure 1 . Taxonomy of systems analysis approaches.

Downloaded from https://academic.oup.com/comjnl/article/25/1/12/527282 by guest on 29 January 2023

approaches all seek to achieve a better analysis of a of various approaches to systems analysis, as classified in
situation or an organization as their objective. This better Fig. 1. It has been argued that the approaches are diverse
understanding may lead to the solving of various and perhaps should be understood in terms of the
problems that the organisation or the system may have, paradigms that they incorporate and the conceptual
but these approaches do not actually seek to provide a models and objectives they hold. Perhaps by identifying
method for solving them. The other approaches, namely the differences in these areas the reader may view the
Checkland, the participative and the traditional ap- approaches in a new light, and perhaps utilize the
proach, all actually claim to be methods of problem approach that meets his requirements. It may also serve
solving and are attempting to achieve more than just the to establish why the various approaches exist and perhaps
provision of a better understanding. The authors argue enable them to function side by side.
that this question of objectives is often ignored and leads
to misunderstandings concerning the various approaches.
Acknowledgments
CONCLUSIONS The authors wish to thank Professor K. N. Bhaskar, Dr P. A.
Dearnley, P. H. Hinings, B. C. Williams and L. D. Fitzgerald for their
The authors have sought to describe the characteristics constructive criticism and comments.

REFERENCES

1. P. Hammersley et al. New approaches to systems analysis and 12. H. C. Ellis, Analysing business information needs, in Data
design. The Computer Journal'23 (No. 1), 1-33 (1980). Analysis for Information Systems Design, ed. by R. N. Maddison,
2. L. von Bertalanffy, General Systems Theory, George Braziller, BCS Conference Papers, British Computer Society, London
New York (1968). (1978).
3. G. J. Klir, An Approach to General Systems Theory, Van 13. M. J. R. Shave, Entities, functions and binary relations: steps to
Nostrand, Princeton, New Jersey (1970). a conceptual schema. The Computer Journal 24 (No. 1), 4 2 -
4. D. Berlinski, On Systems Analysis, M.I.T. Press, Cambridge, 46(1981).
Massachusetts (1976). 14. Tom De Marco, Structured Analysis: Systems Specifications,
5. D. Lilienfeld, The Rise of Systems Theory, Wiley, New York Yourdon, Prentice-Hall, New York (1980).
(1978). 15. C. Gane and T. Sarsons, Structured Systems Analysis: Tools
6. K. Popper, The Open Society and its Enemies, Routledge, and Techniques, Prentice-Hall, New York (1979).
London (1945). 16. T. S. Kuhn, The Structure of Scientific Revolutions, Chicago
7. P. B. Checkland, Towards a systems-based methodology of University Press, Chicago (1962).
real world problem solving. Journal of Systems Engineering 3 17. P. B. Checkland, Science and the systems paradigm, Interna-
(No. 2), 87-116 (1972). tional Journal of General Systems 8, 127-134 (1976).
8. P. B. Checkland, The Development of Systems Thinking by 18. J. W. Sutherland, Systems Analysis. Administration and Archi-
Systems Practice—A Methodology from Action Research tecture, van Nostrand Reinhold, New York (1975).
Programme, Progress in Cybernetics and System Research, Vol. 19. D. S. Smyth and P. B. Checkland, Using a systems approach:
II, Ed. by R. Trappl and F. de P. Hawika, pp. 278-283, the structure of root definitions. Journal of Applied Systems
Hemisphere, Washington DC (1975). Analysis 5 (No. 1), 65-73 (1976).
9. Mumford, Land and Hawgood. A participative approach to 20. S. J. Waters, Towards comprehensive specifications, The
computer systems. Impact of Science on Society 28 (No. 3), Computer Journal 22 (No. 3), 195-199 (1979).
235-253(1978).
10. E. Mumford and D. Henshall, A Participative Approach to
Computer Systems Design; A Case Study of the Introduction
of a New Computer System, Halsted Press, New York (1979). Received March 1980
11. Barry Lee, Introducing Systems Analysis and Design, Vols I and
II, National Computer Centre, Manchester (1978 and 1979). © Heyden & Son Ltd, 1982

16 THE COMPUTER JOURNAL, VOL. 25, NO. 1,1982 © Heyden & Son Ltd, 1982