Requirements Engineering for Enterprise Systems: What We Know and What We Don’t Know? - Intentional Perspectives on Information Systems Engineering - page 117

3 Progress to Date
We checked the published approaches regarding their underlying ideas and assump-
tions, and the availability of empirical evidence about their effectiveness and the
known problems about their use in practice.
3.1 Elicitation Techniques
Our review found that more than 20 requirements elicitation approaches have been
proposed and tried out in real life project settings. In our observation, all these
approaches are based on domain knowledge, however they differ regarding how
they organize domain knowledge and how they create a domain dictionary (e.g.
what knowledge sources they use for this). We clustered the elicitation methods in
five groups:
1. Process-mining based methods , which employ a kind of process-mining tech-
nology in support of requirements elicitation activities. Examples of such
approaches are presented in [ 11, 57] . These researchers came up with special-
ized tools to capture the complete business environments in which the ES will be
put in operation. The result of using such a tool is then considered a first sketch
of the ES process requirements. The idea of process-mining first surfaced in the
1990s, when Intellicorp, an SAP Development Partner, launched the LiveModel
tool capable of identifying all transactions being in use in the current SAP envi-
ronment in a company. The process models generated through this tool have
been used by SAP implementation teams to draft the first version of process
requirements in SAP upgrade, consolidation, or migration projects.
2. Reference-model-based approaches , which rely on predefined process and data
models (termed reference models) and help clarify the questions of (1) what
tasks must be performed and what package-embedded business functionality
can support these tasks, (2) which organizational units should execute these
tasks, (3) what information is needed for executing the tasks in a more effi-
cient way, and (4) what information exchange must happen among tasks and
how the package and other applications would support this. We refer to the ref-
erence models as to reusable and general descriptions of the commonalities in
organizations, business sectors or systems that can be used as a base to derive
other models from. In the ES RE field, we distinguish between two types of
reference models: (i) software-independent reference models which represent
generic descriptions of business processes and data flows in a certain enterprise
area (e.g. accounting) or in a certain business sector (e.g. telecommunications),
and (ii) ES reference models which are “conceptual descriptions of customizable
software” (as defined by [ 43] ). We must note that in RE-for-ES, there is a number
of reference-model-based elicitation approaches that proved their market value
in the past 20 years. Among the software-independent-reference-model-based