Competitive Design of Web-Based Courses in Engineering Education - Web-Based Engineering Education: Critical Design and Effective Tools - page 123

challenging ones. Requirements are further ranked
to reveal their relative importance in the equation
of quality. Tools like AHP are very often used to
rank requirements (Brad, 2008; Raharjo, Xie, Goh
& Brombacher, 2007; Saaty & Vargas, 2001).
The key performance characteristics of the
system are defined and linked to the stakeholders'
requirements by means of a relationship matrix.
This matrix uses a set of coefficients to establish
the level of relationship between each requirement
and each key performance characteristic. For de-
ciding upon the key performance characteristics
from a comprehensive, initial set of performance
characteristics, an evaluation process should be
conducted. A key performance characteristic fully
satisfies the related stakeholders' requirements
only when it attains its target value (Brad, 2008).
A key performance characteristic can be easily
identified from the relationship matrix in the sense
that at least a strong or very strong relationship
level of the respective performance characteristic
with a certain requirement exists.
Setting target values is of very great impor-
tance. Traditional approaches heavily rely on
experience and intuition. This leads most probably
to feasible solutions rather than optimal ones. The
target values will drive all subsequent develop-
ment activities. In this respect, the approach for
defining the right target value of each performance
characteristic is of major concern. When estab-
lishing target values, several aspects should be
considered (e.g. correlations between performance
characteristics, relationships between stakehold-
ers' requirements and performance characteristics,
current performances of systems in the same cat-
egory with the new system, available budget, etc.).
A very critical issue in planning performance
is related to the evolution of stakeholders' re-
quirements over the life-cycle of the new system
(Brad, 2008). In a complex, non-linear evolving
environment, stakeholders could change their re-
quirements both in terms of contents, ranks, target
values and minimum acceptable values (Liang,
2009; Lindroos, Malmivuo & Nousianinen, 2007;

Salihbegovic & Tanovic, 2008; Takago, Matsuishi,
Goto & Sakamoto, 2007). What today is of lower
importance tomorrow could become of very high
importance. Issues that today are missing in the
list of stakeholders' requirements could arise in
the future. Thus, input data in the planning pro-
cess are more complex from the perspective of
system's life-cycle.
Actually, several lists of stakeholders' re-
quirements, including their related ranks, should
be elaborated even before starting the concep-
tualization of the new system. Each list should
consider an interval of time within the life-cycle
of the new system. Each list is actually related
to a certain release of the system. These aspects
come up from the business strategy associated to
the new system. When a new system is going to be
developed, people should establish some business
goals in relation to the respective system, includ-
ing the estimated time of its life-cycle and the
number of releases over the estimated life-cycle.
If for the first release of the system, stakeholders'
requirements and their ranks can be determined
with classical tools (e.g. deep market research
based on interviews, questionnaires, focus groups,
conjoint analyses, etc.), for medium and long term
perspectives non-conventional approaches must
be considered in defining the list of stakehold-
ers' requirements. They should include multiple
scenarios, probabilities of occurrence, internal
and external influence factors, etc.
However, the development of each release of
the system must consider a clear defined list of
requirements, ranks, target values and minimum
acceptable values during performance planning
process. The solution proposed for the first re-
lease meets a lot of challenges because it should
smoothly and cost-effectively translates to the
next releases, too.
Correlations between performance charac-
teristics are also analyzed. Critical for system
design are the negative correlations occurring
between performance characteristics, meaning
that the attempt to improve a certain performance