Information Technology Reference
In-Depth Information
Though the necessity to associate end-users of the future system with its spec-
ification and development steps has long been advocated [1], several approaches
rather propose to deal with the discovery of such constraints by analysing the
content of a related database (or at least, a set of relevant data samples). How-
ever, they rely on the preexistence of large sets of data, which can obviously be
problematic in the process of designing a new information system: there might
be no usable legacy database, or gathering and reencoding a significant amount
of data would be unrealistic.
In this context, the RAINBOW approach [2] provides an alternative and in-
teractive process based on the analysis of a limited set of user-provided data
samples in order to elicit and suggest database constraints and dependencies for
a given schema. RAINBOW is a collaborative and interactive user-oriented ap-
proach to develop the static data model of an information system based on the
reverse engineering of user-drawn form-based interfaces. It relies on the adapta-
tion and integration of principles and techniques coming from various fields of
study, ranging from Database Forward and Reverse Engineering to Prototyping
and Participatory Design.
In this paper, we present how to use the RAINBOW approach to discover
constraints and dependencies. In particular, we focus on the elicitation of func-
tional dependencies, which are a fundamental and critical aspect of conceptual
modelling that has proved dicult to apprehend. The remainder of the paper is
structured as follows. Section 2 delineates the research context, while Section 3
describes the related works. The main principles of the proposal are detailed
in Section 4. Section 5 discusses the evaluation of this process, while Section 6
discusses the proposal and concludes this paper.
2 Research Context
2.1 The RAINBOW Approach
The RAINBOW approach is a collaborative and interactive user-oriented ap-
proach to design database conceptual schemas in the context of Information
System engineering [2]. It exploits the expressiveness of user-drawn form-based
interfaces and prototypes, and specialises and integrates standard techniques
to help acquire and validate data specifications from existing artefacts in or-
der to use such interfaces as a two-way channel to communicate static data
requirements between end-users and analysts. The approach is formalised by a
semi-automatic seven-step process dealing with the progressive modelling of the
application domain:
1.
Represent
: the end-users are invited to draw and specify a set of form-based
interfaces to perform usual tasks of their application domain;
2.
Adapt
: the forms are “translated” into data models, which basically consists
in extracting a data model from each interface using mapping rules;
3.
Investigate
: the data models are cross-analysed to highlight and arbitrate
semantic and structural similarities and produce a pre-integrated schema;
