Use Case Analysis with Narrative Semiotics

Sky Marsen, Robert Biddle, James Noble

Victoria University of Wellington
Wellington, New Zealand
e-mail: [email protected]

Abstract
Use cases are the accepted contemporary vehicle for the capture, collection and management of functional
requirements for information systems and networked e- and m- commerce environments. While employed
widely, use cases lack a critical theoretical foundation. As a result, the applications of use cases vary greatly in
practice. We present an analysis of use cases based on the narrative semiotics of Greimas and on structural
narratology. Our analysis illustrates how these techniques can expose common weaknesses and implicit
assumptions latent within use case texts, and can provide a principled basis for the systematic review and
evaluation of use cases within information systems development methodologies.

Keywords
Use cases, analysis, semiotics, narrative

INTRODUCTION
Use cases are the accepted contemporary vehicle for the capture, collection and management of functional
requirements for information systems and networked e- and m- commerce environments (Jacobson et al. 1992,
Tapscott and Stevens-Guille 2002). While widely used, use cases lack a critical foundation, and as a result,
many practical applications of use cases have difficulties because of their ad-hoc and unprincipled nature
(Cockburn 2001). In this paper, we explore ways in which narrative semiotics can facilitate understanding of
the assumptions, implicit motivations, and goals involved in the development and evaluation of use cases. We
will illustrate how semiotic techniques can expose common weaknesses and implicit assumptions latent within
use case texts, and provide a principled basis for the systematic review and evaluation of use cases within
information systems development methodologies.
The main objective of the paper is to suggest a method of writing user-oriented texts that would firmly ground
this practice on a theoretical base, thereby improving clarity and consistency.

BACKGROUND
Jacobson is usually credited as the originator of use cases. In their book describing their process, Jacobson et al.
define a use case as “a behaviorally related sequence of transactions in a dialogue with the system” (1992: 127).
A more recent definition for the Rational Unified Process shows little real change, saying a use case is “a
description of a set or sequence of actions, including variants, that a system performs that yields an observable
result of value to a particular actor” (Booch et al. 1999). In the early stages of development, use cases help to
focus on interactions as a way of eliciting desirable system behavior, and so help capture requirements and
determine specifications. In the later stages of development, use cases help again because of the focus on
interactions. The interactions can now be regarded as the embodiment of specifications that the system must
meet. Cockburn's (2001) comprehensive book provides a detailed account of how to write use cases, and
includes a good summary of the different styles of use cases. In all, the general idea of a use case is to represent
intended sequences of interaction between a system (even if not yet implemented) and the world outside that
system. The nature of use case quality has been considered, and there are a number of diverse guidelines
offered by Cockburn and by others; these have been explored in broad empirical studies by Anda et al. (2001).
Our approach is to explore existing principled techniques from another domain, as we explain below.
Significantly, the roots of narrative semiotics and narratology lie in the study of folklore and mythology. Many
semiotic models and terms were developed in an attempt to trace common patterns in myths and folktales in
order to understand the ‘grammar’ of universal narrative. Their main aim was to throw light on the processes
humans use to make sense of, and interact with, their environment. This aim clearly shows the parallels between
the design of semiotic models and the design of use cases: both are intended to facilitate an understanding of
how humans translate conceptual processes into physical action. In fact, in addition to the usually credited
founders of semiotics, the philosopher Charles Sanders Peirce (1931) and the linguist Ferdinand de Saussure
(1974) [1913], semiotics is indebted to the work of ethnographers, such as Vladimir Propp (1968) [1928] and
Marsen, Biddle, Noble (Paper #198)
14th Australasian Conference on Information Systems Page 1
26-28 November 2003, Perth, Western Australia
Claude Levi-Strauss (1962), as well as cultural and literary critics (Chatman 1978 and 1990, Greimas 1986,
Greimas and Courtes 1982, Prince 1982, Floch 2001).
Semiotics and narratology already have a presence in computer science. Recent work has shown that some kinds
of computer programs directly encompass ‘stories’, such as computer games (Darley 2000, Jenkins 2000).
Semiotics has been used to study the design of human-computer interfaces (Nadin 1988, de Souza 1993); more
recently it has been adopted to address programming and computer use more generally (Andersen 1993, 1997)
and the organisational and political aspects of information systems development (Clarke 2002, Robichaud 2002,
Underwood 2002). We now address how theories of narrative are relevant for another kind of information
systems ‘text’: use cases.
The branch of semiotics known as ‘narrative semiotics’ distinguishes two levels of narrative: narrative as
trajectory and narrative as schema. The first involves the sequential development of actions, and the
transformation from one state of affairs to another, leading from an initial problem to a resolving climax. The
second involves the positioning of agents in specific roles, as they interact in this transformational sequence.
Narrative trajectory unfolds on the syntagmatic axis of discourse, where temporal enchainment dominates (the
‘and then…and then’ of stories). Narrative schema, on the other hand, involves the paradigmatic axis, linked
with the identity of agents in relation to their actions (the ‘who’s who’ of stories). This paper overviews the two
levels and indicates their significance for use case development and evaluation. For the purpose of this paper,
we will use a simple example, ‘Placing an Order’, from a textbook on use cases, Schneider and Winters’
Applying Use Cases (1998). This represents only one way in which use cases are written, and it is not without
imperfections, but, as it is a textbook example, it is influential and, therefore, characteristic of the writing of use
cases. Besides, it should be noted that the fact itself that there are numerous ways to write use cases highlights
the need for consistency and a more principled approach. In the next section, we concentrate on narrative as
trajectory, and then in a later section we address the narrative schema. We then present our conclusions.

NARRATIVE AS TRAJECTORY
Narrative trajectory favours the verbal (referring to actions) as opposed to nominal (referring to objects) aspect
of narrative, focusing on what agents do, and how these actions fit in a pattern of change from one state to
another. The basic model has the narrative beginning with a problem, and leading to a series of transformational
stages in a developmental sequence aimed at solving the problem. The final stage solves or fails to solve the
initial problem. From a top-down perspective, the narrative is constructed around a Subject’s quest to solve a
problem and/or attain a desired goal, and is designated by the term ‘program of action’ (PA). Zooming in, this
main narrative is divided into a series of sequences, which constitute mini-narratives and reflect the structure of
the whole. Figure 1 represents this graphically:
initial problem actions and their performers resolution
(designation of a lack or question) (liquidation of lack, or answer)
Figure 1: model of basic narrative transformation
In this model, each set of actions transforms the previous set, like the final stage transforms the initial stage. It
should be noted that this is not generated by the principle of causality, as it might at first seem. A
transformation logically presupposes the previous one, but the first should not be seen as causing the second.
Each situation or stage has alternative developments, which, even though they evolve from the previous stage,
are not determined by it. For example, although an action is preceded by the perception of a problem, the
perception of the problem does not entail this particular action. Correlatively, each choice of action dissolves
other possibilities.
Although seemingly simple, this model can be traced at the base of many human activities. In fact, it can be
applied to the ‘quest for knowledge’ itself, as this is exemplified by the process of hypothesising:
Uncertain situation Action Resolution

Hypothesis Experiment / Test Verification / Falsification

Thesis Antithesis Synthesis

Figure 2: Underlying narrative structure of the quest for new knowledge

Marsen, Biddle, Noble (Paper #198)

14th Australasian Conference on Information Systems Page 2
26-28 November 2003, Perth, Western Australia
The model of the narrative trajectory has been reformulated over time. A significant version in the history of
literary and text analysis is the one designed by Algirdas Greimas (first presented in his Structural Semantics
1966), which borrows concepts from the work of Saussure, and ethnographers Propp and Levi-Strauss to
articulate the narrative trajectory in terms of a heroic quest. According to this, a Subject/Hero undertakes to
solve the initial problem through a series of tests that lead to the acquisition (or not) of a desired goal. Some of
the terminology may seem odd in the context of information systems, but the ideas should be familiar. Greimas
distinguishes three categories of tests following the initial problem:
Initial Problem ― Qualifying Test ― Decisive Test ― Glorifying Test
Figure 3: Model of the narrative quest
From a semantic perspective, these stages correspond to three sets of actions, or situations, namely:
• the qualifying stage corresponds to competence: the possession by the Hero of a set of resources that
enable him to perform the action needed to progress to another stage
• the decisive test corresponds to performance: the carrying out of actions based on the problem in
combination with the resources available to the Subject
• the glorifying test corresponds to recognition: the final solution (or lack of solution, for failed projects)
that provides closure to the project.

Applicability of the narrative trajectory to Use Cases

To examine how using the narrative trajectory is relevant for use cases, consider this example of ‘Placing an
Order’, presented as an instance of how to apply use cases in systems development by Schneider and Winters
(1998: 26). We substitute numbers for bullet points to facilitate analysis.

Precondition: a valid user has logged into the system

• Flow of events:
o Basic path:
1. The use case starts when the customer selects Place Order.
2. The customer enters his or her name and address.
3. If the customer enters only the zip code, the system will supply the city and state.
4. The customer will enter product codes for the desired products.
5. The system will supply a product description and price for each item.
6. The system will keep a running total of items ordered as they are entered.
7. The customer will enter credit card payment information.
8. The customer will select Submit.
9. The system will verify the information, save the order as pending, and forward payment
information to the accounting system.
10. When payment is confirmed, the order is marked Confirmed, an order ID is returned to
the customer and the use case ends.
o Alternative paths
• In step 9, if any information is incorrect, the system will prompt the customer to correct
the information.
• Postcondition: The order has been saved in the system and marked confirmed
This use case is clearly based on a developmental series of actions transforming an initial problem: the user has
not placed an order and wants to do so. The competence is that the user has all the resources necessary to be
‘valid’ so as to log into the system. The qualifying test of ‘recognition’ is that the system saves the data and
produces a confirmation sign. As is stands, however, the use case has some weaknesses.
Firstly, not having taken into account the full implications of narrative, the case is confused as to what agent
occupies the position of Subject (the positioning of agents is the topic of the next section; however, enough has
been said already to make the present discussion comprehensible). Tasks 1, 2, 3, 4, 7 and 8 place the customer
as Subject, while 5, 6, 9, 10 place the system in this position, and task 3 has a conditional Subject. In other
words, this case is made up of two different PAs misleadingly intertwined. Actually, taking the customer as
Subject, tasks 5, 6, 9 and 10 are not tasks, but responses to the Subject’s actions. Therefore, it would be more
productive to divide this use case into two PAs, one with customer and the other with system as Subject. This
would show clearly what is expected to happen in each step and would highlight what could go wrong.

Marsen, Biddle, Noble (Paper #198)

14th Australasian Conference on Information Systems Page 3
26-28 November 2003, Perth, Western Australia
Secondly, focusing on the customer as Subject would show that this is actually one sequence in a larger
narrative. It can be assumed that the customer’s main goal is not to place an order but to receive the desired
products. This may not be directly relevant to the system’s story but it is contextually pertinent and should be
acknowledged if more realism and effective customer service are sought. For example, recognising this may
encourage the designer of the system to include information about other avenues of purchase or access to goods
and so make the system more competitive.
Thirdly, by not recognising that there are stages, and not just steps, involved in this PA, the case makes some
unwarranted assumptions about the customer’s competence. In order for the PA to progress from qualifying
through to glorifying stage, the Subject must make use of the skills and resources that form his competence. In
the above use case, there are two steps that assume a competence that may not be there: the request for a zip
code, which defeats localization principles and limits the internationalism of the system by assuming a specific
type of geographical location, and the request for a credit card number, which limits usability by closing off
alternative paths of payment. By placing the customer as Subject and by dividing the steps into stages that list
the assumed or ideal competence for each stage, important details would be highlighted, facilitating the design
of a more comprehensive and user-friendly system. Here is how a narrative-semiotic approach to this use case
would look like:
• Main Program of Action: a user wishes to obtain X goods and uses the software as a helper to achieve
this goal.
• The narrative sequence in which the user requests this help takes this form:
o Problem: the user has not placed an order to receive the desired goods
o Qualifying Test: the user needs
1. access to the system (specify requirements for this), and knowledge how to act in it
2. contact details (name and address)
3. a credit card
o Decisive Test: the user will carry out these actions consecutively:
1. Log in
2. Select Place Order
3. Enter name and address (assumed competence zip code – is this necessary?)
4. Enter product codes for desired products
5. Enter credit card information (assumed competence credit card – any other ways to pay?)
6. Select Submit
o Glorifying Test: the user will receive a Confirmed sign at the end of the actions.
Placing the system as Subject we would get this:
• Main Program of Action: is this task part of the broader functions of the software?
• Problem: the system must assist the user in placing an order
• Qualifying Test: the system needs:
o clear indications of what the user requires to place the order
o a prompt function to help direct the user in cases of error
• Decisive Test: the system will carry out these actions in response to user input:
1. Give prompt to log in
2. Give directions on how to place an order and on user competence (i.e. what the user needs
to successfully complete the task)
3. Give prompt to enter name and address (+ zip code?)
4. Give prompt to enter codes for desired products
5. Give descriptions of products and their prices
6. Maintain record of products as they are entered
7. Describe payment options
8. Accept user’s mode of payment
9. Verify information
10. Save order as pending
11. Forward payment information to accounting system
12. Mark order as confirmed and issue an order ID
• Glorifying Test: the order has gone to the responsible parties for filling it.
Finally, the questions that need to be asked when analysing a project according to the narrative trajectory are:
Marsen, Biddle, Noble (Paper #198)
14th Australasian Conference on Information Systems Page 4
26-28 November 2003, Perth, Western Australia
 • is the initial problem made explicitly clear (without which the project may remain ambiguous)?
• how many sequences are there between initial situation and resolution (making sure that the Subject
remains the same throughout the sequence and point of view does not change inexplicably)?
• what competence is assumed for each task or action?
• what possible outcomes does each task or action have (which could anticipate troubleshooting)?

THE NARRATIVE SCHEMA

Stories are not just about change, actions and events. In the tradition of myth and folklore, they are also about
heroes, villains and desired objects. The guiding questions of the narrative schema include ‘how are the
constituent agents and elements of a project defined?’ and ‘How are they positioned in relation to one another?’
In other words, ‘who’s in the story and what do they do?’ The main operational tool devised by narrative
semioticians to deal with such identities and relationships is the actantial model (Greimas 1982, Floch 2001).
An actant is the formal set of properties that positions a sign in a category relative to others within a semiotic
system. The actant designates the possible performers of actions. An actor, or agent, on the other hand, is the
specific object that gives concrete form to the abstract properties of the actant. Combining the logic of sentence
structure (which consists of a subject (active agent), verb (action) and object (passive agent)) with the logic of
myths and fairy tales, the actantial model recognizes six actants and places them in these positions:
The actants represent categories, which means that each of the six actants may include one or several agents (or
actors). For instance, the position of Sender, which includes the motivations or reasons of an act, belief or event
and, thus, underlies the causal principles with which a project is created, could be occupied by an abstract sign,
such as Love, Gravity or Mind, as it could by a figural sign, such as a King, a User or a Computer Virus.
Similarly, if the one who performs the actions leading to the final result is the same as the one who motivates
them, the positions of Subject and of Sender are occupied by the same agent – and the same is true of the other
positions. Finally, as regards long or complex narratives, the performers may be ‘re-shuffled’, so that, say, the
Helper in one sequence may become the Opponent in another.
The actantial model does not only show the relational positioning of agents in a project; it also helps to highlight
the emphasis that a project may give to one set of relations over other alternatives. This, in turn, can be used to
classify the types of documents produced for an IT project. The three sets of relations identified by narrative
semiotics are Struggle, Contract or Communication. Applying these to IT projects, we would get something like
this. User instructions focus on the relations between Subject and Object and the actions necessary to conjoin
them; so, they correspond to narratives of Struggle. Project sections that profile users and aim to ascertain their
needs, skills and objectives concentrate on the relations between Sender and Subject, and correspond to
narratives of Contract. Persuasive material, such as proposals for funding and promotional copy for marketing of
products focus on relations between Sender and Receiver – the needs or desires that the product will satisfy and
the benefits it will provide. These are relations of Communication. When a document is constructed primarily
around the question ‘why is this happening?’ it tends to involve contractual relations between Sender and
Subject. When a document is constructed primarily around the question ‘how can I do this?’ it tends to involve
relations of struggle between Subject and Object. And when a document is primarily constructed around the
question ‘what are the benefits of this?’ it tends to involve relations of communication between Sender and
Receiver.

SENDER RECEIVER
the one that motivates the one that profits
the action from the action
SUBJECT
the one that carries out
the dictates of the Sender

OBJECT
the desired goal or state
of affairs

Marsen, Biddle, Noble (Paper #198)

14th Australasian Conference on Information Systems Page 5
26-28 November 2003, Perth, Western Australia
 HELPER OPPONENT
the one that assists the one that hinders
the Subject in its quest the Subject in its quest
Figure 4: Actantial model

Applicability of the narrative schema to Use Cases

To examine the relevance of the actantial model to use cases, consider this actantial model representation of our
running example, the use case presented earlier from Schneider and Winters (1998: 26).

Commercial requirements
how the user can buy the products User
Desire for products
Ease of e-commerce

User

Place Order
Receive desired products

System System
Eligibility Intransigence
what makes the user ‘valid’ inability to negotiate or ask for clarification

Figure 5: Actantial model of a sample use case

From the information given, and placing the user as Subject, the System can be either a Helper or an Opponent,
depending on smooth usability. What motivates the Subject in his/her quest is the desire to obtain the goods, and
the purchasing requirements: e-commerce is practical and time saving. As far as we know, the user is purchasing
own products (retail), so s/he is the beneficiary of the task. Finally, the fact that the user is eligible to purchase
goods in this manner helps his/her quest, while the technical formality of the process (i.e. the lack of personal
contact) is often a hindrance. Being aware of this danger enables the system designers to take measures to
minimize the risk of this opposition as much as possible.
The advantage of representing this information in this form is that it renders explicit such elusive but very
important aspects of a use case, such as what exactly motivates the Subject, what their fears and aspirations are,
what precisely they can use to their advantage to accomplish their goal, and what the short and long term
benefits would be. In fact, the main purpose of the actantial model was to pinpoint and articulate the abstract,
and thereby make it more conducive to analysis and understanding. The model is also likely to bring to light any
‘paradoxes’ of use cases – the cases where the goals are seemingly contradictory, and the cases where the same
agent occupies opposing positions. For use cases, putting information in actantial positions would most
advantageously be done in the planning and review stages of requirements analysis, because that is when it is
necessary to become aware of anticipated helpers, opponents and benefits, define participants and objects, and
establish realistic goals.
Finally, the questions that need to be asked when analysing a project according to the narrative schema are:
• whose perspective or point of view is taken in the arrangement of information?
• what is the rationale for placing agents in their positions?
• who is motivating the Subject in his or her quest for a desired Object?
• whose point of view has created the desired Object?
• who will benefit from the acquisition of the Object?

Marsen, Biddle, Noble (Paper #198)

14th Australasian Conference on Information Systems Page 6
26-28 November 2003, Perth, Western Australia
CONCLUSION
In this paper, we have introduced a novel approach to the theoretical analysis of use cases: narrative semiotics.
Our main objective was to suggest a method of writing user-oriented texts that would firmly ground this practice
on a theoretical base, thereby improving clarity and consistency. Considering use cases as narratives can
address many of the common questions about them, to assist both the analysts who develop use cases and the
stakeholders who must review them. We hope that these semiotic analyses can provide a platform for future
progress in the research and application of use cases.
As a final note, it should be added that the benefits of interdisciplinary research in software engineering and
narrative semiotics are reciprocal. Software projects are not the only side to gain from the application of
narratological theory; the converse is also true. The interactivity of software will also have an effect on theories
of narrative. Art (Holtzman 1997), education (Brown 2000) and even perceptions of self and other (Heim 1993)
have taken different forms under the influence of the digital media. It remains to be seen whether these new
forms can still be theorized using the traditional concepts and methods of narrative semiotics, or whether
random access reality, projected selves, Internet identities and holographic images have created new collective
mental representations that require a new theoretical language. But this is beyond the scope of this paper.

REFERENCES
Anda, B., Sjøberg, D., and Jørgensen, M. (2001) Quality and Understandability of Use Case Models. In Object-
Oriented Programming: 15th conference proceedings (ECOOP 2001), Berlin, Springer-Verlag, 2001.
Andersen, P.B. (1993) A semiotic approach to programming. In The Computer as Medium, Cambridge:
Cambridge University Press, 16-67.
Andersen, P.B. (1997) A theory of computer semiotics. Cambridge Series on Human-Computer Interaction.
Cambridge: Cambridge University Press.
Brown, J. S. (2000) Growing up digital: How the web changes work, education, and the ways people learn.
Change March/April.
Chatman, S. (1978) Story and discourse: narrative structure in fiction and film. Ithaca: Cornell University Press.
Chatman, S. (1990) Coming to terms: The rhetoric of narrative in fiction and film. Ithaca: Cornell University
Press.
Cockburn, A. (2001) Writing effective use cases, Boston: Addison-Wesley.
Clarke, R.J. (2002) Intertextuality at Work: Large Scale Organisation of Workpractices. In Lui, K. et al.,
(Editors) Coordination and Communication Using Signs: Studies in Organisational Semiotics. London:
Kluwer Academic Publishers.
Darley, A. (2000). Visual digital culture: Surface play and spectacle in new media. London: Routledge.
De Souza, C.S. (1993) The semiotic engineering of user interface languages. International Journal of Man-
Machine Studies, 39(5):753-773.
Floch, J-M. (2001) Semiotics, marketing and communication. Basingstoke: Palgrave.
Greimas, A. J. (1966) Semantique structurale: Recherche de methode. Paris: Larousse.
Greimas, A. J. (1987) [1982] On meaning: Selected writing in semiotic theory. Minneapolis: University of
Minnesota Press.
Heim, M. (1993) The metaphysics of virtual reality. New York: Oxford University Press.
Holtzman, S. (1997) Digital mosaics. New York: Simon and Schuster.
Jenkins, H. (2000). Art form for the digital age. Technology Review. 103: 5, 17-120.
Jacobson, I., Christerson, M., Jonsson , P. and Overgaard, G. (1992) Object-Oriented Software Engineering.
Boston: Addison-Wesley.
Levi-Strauss, C. (1966) [1962] The savage mind. London: Widenfeld and Nicolson.
Nadin, M. (1988) Interface design: A semiotic paradigm. Semiotica, 69(3):269-302.
Peirce, C. S. (1931). Collected papers of Charles Sanders Peirce. Cambridge, MA: Harvard University Press.
Prince, G. (1982) Narratology: The form and functioning of narrative. Amsterdam: Mouton.
Marsen, Biddle, Noble (Paper #198)
14th Australasian Conference on Information Systems Page 7
26-28 November 2003, Perth, Western Australia
Propp, V. (1968 ) [1928] Morphology of the folktale. Austin, TX: University of Texas Press.
Robichaud D. (2002) Greimas' Semiotics and the Analysis of Organisational Action. In Lui, K. et al. (Editors)
Coordination and Communication Using Signs: Studies in Organisational Semiotics. London: Kluwer
Academic Publishers.
Saussure, F. (1974) [1913]. Course in general linguistics. London: Fontana.
Schneider, G. and Winters, J. P. (1998): Applying use cases, Reading, MA: Addison-Wesley.
Tapscott, D., Stevens-Guille (2002) M. Evolution of Location-based Services – Truth Versus Myth, Wireless
Technology, London: Business Briefings.
Underwood J. (2002) Translation, Betrayal, and Ambiguity in IS Development. In Lui, K. et al. (Eds).
Organisational Semiotics: Evolving a Science of Information Systems. London: Kluwer Academic
Publishers.

COPYRIGHT
Sky Marsen, Robert Biddle, James Noble © 2003. The authors assign to ACIS and educational and non-profit
institutions a non-exclusive licence to use this document for personal use and in courses of instruction provided
that the article is used in full and this copyright statement is reproduced. The authors also grant a non-exclusive
licence to ACIS to publish this document in full in the Conference Papers and Proceedings. Those documents
may be published on the World Wide Web, CD-ROM, in printed form, and on mirror sites on the World Wide
Web. Any other usage is prohibited without the express permission of the authors.

Marsen, Biddle, Noble (Paper #198)

14th Australasian Conference on Information Systems Page 8
26-28 November 2003, Perth, Western Australia