REQUIREMENTS ENGINEERING FOR THE BUSINESS PROCESS

RE-ENGINEERING: an Example in the Agro-Food Supply Chain

∗

Floriana Marin

IASMAA

via E. Mach, 1

38100 – S. Michele a.A. (Italy)

Paolo Bresciani and Fabrizio Sannicol

ITC-irst

via Sommarive, 18

38050 – Trento-Povo (Italy)

Lucia Martinelli

IASMAA

via E. Mach, 1

38100 – S. Michele a.A. (Italy)

Keywords:

Requirements Analysis, Modelling Notations, Business Processes Re-engineering

Abstract:

Recent researches in Requirements Engineering and Software Engineering suggest that a deep organizational

analysis is needed as a preliminary phase of any project that aims at introducing or modifying the use of

Information Technologies inside an organization. Accordingly to this view, the social setting is the motivating

factor that justify the technological choices. In this paper, we show with an example how balancing social and

business issues can be properly addressed and analyzed by means of the Early Requirements Analysis phase

of the Tropos methodology. In particular, we refer to an ongoing project with the goal of proposing web based

technologies to spread-off information, data and knowledge on the so called Genetically Modiﬁed Organisms.

1 INTRODUCTION

The debate on modern agrobiotechnologies and, spe-

cially, the so called Genetically Modiﬁed Organisms

(GMO) is surely one of the most controversial ques-

tions of the last years, involving many levels of in-

terests and actors. Economic, political and scientiﬁc

signiﬁcance involves the discussion, since the appli-

cation of this technology has a deep impact on the

“quality of life”. Besides, the difﬁcult communication

attitude often achieved by the scientiﬁc community

towards the Society does not help a conﬁdent debate,

and allows an uncontrolled spread of incorrect infor-

mation. The question involves a great amount of data

and information, to be allocated in an efﬁcient way

to allow all the actors involved to choose for them-

selves. Differences in interests and roles should be

regarded as driving forces in the debate. In what we

could call a deliberative democracy, each part should

be provided with the same amount of information, but

∗

Research supported by Autonomous Province of

Trento, Project OSSERVA3

at the same time should be allowed to express a per-

sonal and aware opinion, free from others’ inﬂuences.

This explains the need of rationalizing the way infor-

mation and data are made available to the different

actors participating to the scientiﬁc debate on GMOs.

To answer to these needs, we are currently working at

a project (Osserva3), the aims of which include also

to provide an Information Technology and/or Multi-

media infrastructure to support knowledge and infor-

mation sharing and broadcasting for a better informed

deliberative democracy on the topic. Thus, we adopt

the Tropos methodology for Requirements and Soft-

ware Engineering (Bresciani et al., 2004) to deﬁne the

requirements of such a system (Marin et al., 2003).

The Tropos methodology deals with all the phases

of system requirement analysis and all the phases of

system design and implementation in a uniform and

homogeneous way, based on common mentalistic no-

tions as those of actors, goals, soft-goals, plans, re-

sources, and intentional dependencies. Thus, one of

the Tropos main advantages is that it allows us to cap-

ture not only the what or the how, but also the why a

piece of software is developed. The Tropos method-

538

Marin F., Bresciani P., Sannicolò F. and Martinelli L. (2004).

REQUIREMENTS ENGINEERING FOR THE BUSINESS PROCESS RE-ENGINEERING: an Example in the Agro-Food Supply Chain.

In Proceedings of the Sixth International Conference on Enterprise Information Systems, pages 538-542

DOI: 10.5220/0002641705380542

 SciTePress

ology is based on four phases (Bresciani et al., 2001;

Bresciani et al., 2004): Early Requirements Analysis,

to understand the problem context by studying its or-

ganizational setting; Late Requirements Analysis,to

deﬁne the system-to-be, in the context of its organiza-

tional environment; Architectural Design, to deal with

the deﬁnition of the system global architecture and

the Detailed Design phase. In particular, during the

Early Requirements Analysis the existing organiza-

tional setting is analyzed in terms of actors, who play

some role in the organization, and of their reciprocal

intentional dependencies, in the context of the orga-

nization. The output of this phase is an organizational

model which includes relevant actors and their respec-

tive intentional dependencies. Actors are character-

ized by having goals that each single actor, in isola-

tion, would be unable —or not as well or as easily—

to achieve. Intentional dependencies are used to de-

scribe this kind of relationships among actors. Goals

are the the elements around which the intentional de-

pendencies are established.

In this paper, we will concentrate on the Early Re-

quirements Analysis phase. In particular, we will base

our discussion on the assumption that a deep organi-

zational analysis is needed as a preliminary phase of

any project that aims at introducing or modifying the

use of Information Technologies inside an organiza-

tion (Bresciani et al., 2004; Donzelli, 2003; Bresciani

et al., 2001; Donzelli and Bresciani, 2003). Speciﬁ-

cally, in the Osserva3 project we focus on carrying out

a clear analysis of the current socio-economical and

scientiﬁc scenario on biotechnologies, taking into ac-

counts diverse actors, as, e.g., Institutions, Scientists,

Industries, Farmers and Breeders, Consumers, Mass

Media, Food Retailers, their own interests and their

reciprocal intentional dependencies.

2 USING TROPOS: A SIMPLE

SCENARIO

For the sake of brevity we here consider only a sim-

pliﬁed business scenario extracted from our Osserva3

project: the terminal part of the agro-food products

delivery chain. The objective is to analyze how the

consumer attitudes toward the GMOs products may

inﬂuence the marketing strategies. Of course, this is

only a small and simpliﬁed fragment of the analysis

developed inside the Osserva3 project (that includes

more then 27 actors).

The two main actors (the consumer and the food re-

tailer) are characterized by having a set of goals, and

depend each other for achieving some of these goals.

E.g., the consumers aim at environment protection,

technical development and social equality. This at-

titude is reasonable, as well as it is reasonable the fact

Consumer

food

products

are

furnished

Food

Retailer

environment

protected

technical

development

is achieved

Legend

Goal

dependency

depender dependum

dependee

maximizes

profits

happy

consumer

quality

products

are

provided

Actor

Goal

Softgoal

repeats

purchasing

low

prices

social

equality is

guaranteed

Figure 1: Actor Diagram with goals and intentional depen-

dencies between

Consumer and Food Retailer.

that the food retailer aims at obtaining high econom-

ical proﬁts. To better understand their motivations,

it is important that the two actors are not considered

and analyzed in isolation: they depends each others to

reach some objectives. The food retailer depends on

the consumer to have her as a costumer or, better, to

have her keeping on shopping with it. The consumer

clearly depends on the food retailer ﬁrst of all to buy

food, but also for its quality level and (possibly) low

prices. As well, it is interesting to consider, e.g., the

trade-off between quality and prices, a detailed anal-

ysis on what kind of quality the consumer may be

interested in, the strategies that the food retailer can

adopt to better match consumers expectations, and so

on. Tropos provides a diagrammatic notation to deal

with this kind of analysis: by using diagrams we have

a visual tool that allow us to focus different aspects

on turn, and to breach the global cognitive effort into

smaller chunks of knowledge, that can be more easily

understood and analyzed, as shown next.

2.1 Building Tropos Actor Diagrams

First we identify the most relevant actors in the social

environment. Initially, we concentrate only on few

actors, together with their goals and reciprocal inten-

tional dependencies: the

Consumer and the Food Re-

tailer

, as shown in the Actor Diagram of Figure 1. Ac-

tor Diagrams are used to represent goal dependencies

among actors. Actors are represented by means of cir-

cles, labeled by the actor names. In Figure 1, some of

the possible goals of the two actors

Consumer and

Food Retailer are represented. The main Food Re-

tailer

goal is maximizes proﬁts. Goals are represented

by means of labeled ovals. In the case of

maximizes

proﬁts

the goal appears in the diagram attached to the

actor that aims at fulﬁlling it (

Food Retailer). Simi-

larly, the actor

Consumer wants to attain the goals so-

cial equality is guaranteed

, technical development is

achieved

, and environment is protected, as well as the

goal

happy consumer. Some other goals in Figure 1

are not directly attached to any actor, but in the middle

REQUIREMENTS ENGINEERING FOR THE BUSINESS PROCESS RE-ENGINEERING: AN EXAMPLE IN THE

AGRO-FOOD SUPPLY CHAIN

539

of a path of the kind actor1→goal→actor2 (e.g., as

Food Retailer→repeats purchasing→Consumer).

The meaning of such a pattern is that actor1 (referred

as the depender) depends on actor2 (referred as the

dependee) to achieve the goal (referred as the depen-

dum), either because she is not able to satisfy it by

herself, or not as easily or not as efﬁciently. In Fig-

ure 1, the

Consumer depends on Food Retailer for ful-

ﬁlling the soft-goals

low prices, quality products are

provided

, and the goal food products are furnished.

Finally

Food Retailer delegates the goal repeats pur-

chasing

to the Consumer.

2.2 Building Tropos Goal Diagrams

After identifying the relevant stakeholders, their

goals, and intentional dependencies, the Tropos Early

Analysis phase proceeds by decomposing each goal

in subgoals by means of techniques of AND-OR de-

composition and contributions analysis (Chung et al.,

2000; Mylopoulos et al., 2001; Dardenne et al.,

1993). The resulting diagrams are called Goal Dia-

grams (Bresciani et al., 2001). AND-OR decomposi-

tion allows for a combination of AND and OR decom-

positions of a root goal into subgoals, thereby reﬁning

a goal structure. In particular, AND-decomposition

implies that all subgoals have to be fulﬁlled for

achieving the root goal, while for OR-decomposition

the fulﬁllment of one of them is enough. Contribution

analysis allows us to point out goals and soft-goals

that can contribute positively or negatively at reaching

the goal under analysis. Using these techniques, we

can build some diagrams in which we analyze each

goal from the actor point of view, to acquire a deeper

understanding on how and why to achieve that goal.

The idea is that a goal may be decomposed in one or

more subgoals, that could be delegated to other actors

or fulﬁlled in isolation.

Figure 2 shows a Goal Diagram where the goals

of the actor

Food Retailer are decomposed by means

of AND-OR decomposition. In particular, we OR-

decompose the goal

maximizes proﬁts into the sub-

goals

reduces costs, increases quantity,orincreases

prices

: the satisfaction of (at least) one of the sub-

goals guarantees also the satisfaction of

maximizes

proﬁts

. Again, the goal reduces costs is further re-

ﬁned into the two subgoals

wholesale product prices

are reduced

and optimizes management. The former

is delegated to the actor

Agro-food Industry, establish-

ing a new goal dependency between the actor

Food

Retailer

and Agro-food Industry. It is worth notic-

ing that, here, the dependency arrow points directly

A distinction is made in Tropos between goals —aka

hard-goals— and soft-goals. While hard-goals deﬁne objec-

tives the satisfaction of which can be deﬁned with clear-cut

criteria, soft-goal achievement cannot be sharply deﬁned.

increases

prices

reduces

costs

increases

quantity

Food

Retailer

maximizes

profits

gives

trust

reduces

prices

Consumer

a good

image

is built

Public

Relations

Agency

clear

and detailed

label

Legend

Contribution

+/-

OR decomposition

optimizes

management

Agro-food

Industry

wholesale

product

prices

are reduced

organic

products

are

provided

GMO free

products

are

provided

happy

consumer

quality

products

are

provided

food

quality

AND decomposition

repeats

purchasing

Figure 2: Goal Diagram from the Consumer and Food Re-

tailer

point of view.

Food

Retailer

Agro-food

Industry

GMO free

food

producers

Competitive

producers

wholesale

product

prices

are reduced

ISA

clear

and detailed

label

organic

products

are

provided

GMO free

products

are

provided

Organic

food

producers

Figure 3: Actor Diagram with the new actors introduced by

effect of the goal analysis.

from inside the context of the depender toward the

dependee. Thanks to this kind of analysis, we can

complete the Actor Diagram depicted in Figure 1 with

new details like hard-/soft-goal dependencies, new ac-

tors, and so on. The Tropos methodology foresees

to incrementally increase the detail level of each dia-

gram, by means of a cyclic process, until a sufﬁciently

detailed model is produced.

The so called contribution analysis allows us to

highlight hard- and soft-goals which contribute pos-

itively or negatively at reaching the goal under anal-

ysis. For example, the arrow labeled with a “+” and

pointing from the goal

reduces costs to the goal re-

duces prices

means that the ﬁrst goal contributes posi-

tively at satisfying the second one. Again, the goal

re-

duces prices

contributes partially positively at fulﬁll-

ing the goal

happy consumer and negatively at maxi-

mizes proﬁts

(note the label “−”).

ICEIS 2004 - INFORMATION SYSTEMS ANALYSIS AND SPECIFICATION

540

Agro-food

Industry

GMO free

food

producers

Competitive

producers

ISA

Organic

food

producers

repeats

purchasing

low

prices

Consumer

happy

consumer

quality

products

are

provided

organic

products

are

provided

maximizes

profits

Food

Retailer

clear and

deteiled

label

food

products

are

furnished

GMO free

products

are

provided

wholesale

product

prices are

reduced

Figure 4: Global Actor Diagram with goal contributions.

3 REVISING ACTOR DIAGRAMS

Accordingly to the Tropos Early Requirements Anal-

ysis process, each diagram may evolve incrementally,

following an iterative and incremental development.

For example, it maybe here interesting considering

how the dependencies introduced during Goal Anal-

ysis can be further developed in the Actor Diagram

again. Let us consider the dependencies among

Food

Retailer

and Agro-food Industry, introduced with Fig-

ure 2. A better picture of these dependencies can be

given in the Actor Diagram of Figure 3. Here, an ex-

tra level of detail is provided, which states that three

of the dependencies posed on the

Agro-food Industry

are indeed to be delegated to more speciﬁc kinds of

it. Three new actors,

Organics food producers, GMO-

free food producers

, and Competitive producers (see

Figure 3) are introduced as specializations (ISA)of

the actor

Agro-food Industry. The three new actors

inherit the dependencies and goals of the actor

Agro-

good Industry

, like, e.g., the delegations of the soft-

goal

clear and detailed label. As well, they are the

dependees in the speciﬁc dependencies for achieving

the goals

organic products are provided, GMO free are

provided

, and wholesale product prices are reduced

respectively. These last observations let us refocus

on the original task: to analyze and understand which

possible impact may have the introduction of GMO

products in a agro-food products delivery chain. In

fact, from the contribution analysis of 2, we can evi-

dence that price competitiveness is in general a factor

contrasting with quality. Since quality is often related

by many consumers to the absence of artiﬁcial fac-

tors —like chemicals but also GMO— in the prod-

ucts, organics products and GMO-free products are

frequently perceived as having a higher quality stan-

dard, and this ﬁts the expectations of the average con-

sumer, despite the fact that she has to pay a higher

price. Thus, after balancing all the elements in favour

and against (that of course do not have all the same

weight), we can propose to adopt our analysis to jus-

tify market and business choices.

Finally, we brieﬂy introduce an interesting exten-

sion for the Analysis process, where contributions

with a relevance for different actors are represented

in a more organic and compact way. In Figure 4, all

the contributions with a relevance for more than one

actor, i.e., spanning outside the local contexts of each

single actor, are replicated in a global Actor Diagram.

The idea is to highlight the impact that the different

goal (hard and soft) may have on the dependencies

among the actors. In this way we abstract from con-

textual details —the single, personal point of views—

and can present a more objective view of the global

contributions network. The aim is to ease the task

of capturing the dependums that play a crucial global

role, so to facilitate the identiﬁcation and understand-

ing of the most essential goals on which to focus pos-

sible choices and decisions.

4 CONCLUSION

Starting from a simpliﬁed business scenario in the

context of the current debate on the adoption of

GMOs for food products, dealing, in particular, with

the consequences that possible alternative choices

about the option of selling GMO-food may have on

the business strategies of the terminal part of the prod-

ucts delivery chain, we introduced in this paper a tech-

nique for analyzing business scenarios, also aimed

at implementing the Information Systems Require-

ments Engineering processes. The technique —called

Early Requirements Analysis— is part of the Tropos

methodology for Requirements and Software Engi-

neering and it is presented here by means of our case-

study. As well, we showed how consequences on the

business process can be directly derived by means of

the analysis proposed by using Early Requirements.

In particular, in the last part of the paper 4, we

brieﬂy sketched a compact Actor Diagram view,

aimed at summarizing the propagation of contribu-

tion analyses—which normally reside only inside the

scope of each single actor Goal Diagram across an

extended view including all the relevant actors.

REFERENCES

Bresciani, P., Giorgini, P., Giunchiglia, F., Mylopoulos, J.,

and Perini, A. (2004). TROPOS: An Agent-Oriented

Of course, this is a very simplistic view of the scenario

—we are far from proposing, here, any real marketing stat-

egy. Our aim, is just to give an idea of the feasibility of the

application of this approach.

REQUIREMENTS ENGINEERING FOR THE BUSINESS PROCESS RE-ENGINEERING: AN EXAMPLE IN THE

AGRO-FOOD SUPPLY CHAIN

541

Software Development Methodology. In press. Jour-

nal of Autonomous Agents and Multi-Agent Systems.

Kluwer Academic Publishers.

Bresciani, P., Perini, A., Giunchiglia, F., Giorgini, P., and

Mylopoulos, J. (2001). A Knowledge Level Soft-

ware Engineering Methodology for Agent Oriented

Programming. In M

uller, J. P., Andre, E., Sen, S., and

Frasson, C., editors, Proceedings of the Fifth Interna-

tional Conference on Autonomous Agents, Montreal.

Chung, L. K., Nixon, B. A., Yu, E., and Mylopoulos, J.

(2000). Non-Functional Requirements in Software

Engineering. Kluwer Publishing.

Dardenne, A., van Lamsweerde, A., and Fickas, S. (1993).

Goal-directed requirements acquisition. Science of

Computer Programming, 20(1–2):3–50.

Donzelli, P. (2003). Agoal-driven and agent-based Require-

ments Engineering Framework. Requirements Engi-

neering, Springer-Verlag, (ISSN 0947-3602).

Donzelli, P. and Bresciani, P. (2003). Goal oriented require-

ments engineering: a case study in e-government.

In Eder, J. and Missikoff, M., editors, Advanced In-

formation Systems Engineering (CAiSE’03), number

2681 in LNCS, pages 605–620, Klagenfurt/Velden,

Austria. Springer-Verlag.

Marin, F., Sannicol

o, F., Bresciani, P., and Martinelli, L.

(2003). Tropos methodologies applied to an analya-

sis of actors and goals dependencies in the agro-

biotech debate. In Proceedings of XLVII SIGA Annual

Congress, Polo Didattico ”G.Zanotto”, 24/27.

Mylopoulos, J., Chung, L., Liao, S., Wang, H., and Yu, E.

(2001). Exploring Alternatives during Requirements

Analysis. IEEE Software, 18(1):92–96.

ICEIS 2004 - INFORMATION SYSTEMS ANALYSIS AND SPECIFICATION

542