Authors:
Nelufar Ulfat-Bunyadi
;
Rene Meis
and
Maritta Heisel
Affiliation:
University of Duisburg-Essen, Germany
Keyword(s):
Four-Variable Model, Context, Context Modelling, Contextual Decision, Satisfaction Argument, Domain Knowledge, Requirement, Specification.
Related
Ontology
Subjects/Areas/Topics:
Context
;
Context Analysis
;
Context Representation
;
Paradigm Trends
;
Software Engineering
Abstract:
A control system usually consists of some control software as well as sensors and actuators to monitor and
control certain quantities in the environment. The context of the control software thus consists of the sensors
and actuators it uses and the environment. When starting development of the control software, its context
is often not predefined or given. There are contextual decisions the developers can make (e.g. which sensors/
actuators/other systems to use). By means of these decisions, the context is defined step by step. Existing
approaches (like the Four-Variable Model) call for documenting the environmental quantities (monitored,
controlled, input, and output variables) that are relevant after making these contextual decisions. The environmental
quantities that have originally been relevant (i.e. before deciding which sensors/actuators/other
systems to use) are not documented. This results in problems when the software shall later on be reused in
another, slightly different
setting (e.g. with additional sensors). Then, it is hard for developers to decide which
environmental quantities are still relevant for the software. In this paper, we suggest an extended version of the
Four-Variable Model, the Six-Variable Model, and, based on that, a context modelling method, that combines
existing approaches. The benefit of our method is that the environmental quantities that are relevant before
and after decision making are documented as well as the contextual decisions themselves and the options that
were selectable. In this way, later reuse of the software is facilitated.
(More)