Authors:
Benny Akesson
1
;
Jozef Hooman
2
;
Roy Dekker
3
;
Willemien Ekkelkamp
3
and
Bas Stottelaar
4
Affiliations:
1
Embedded Systems Innovation by TNO, Netherlands
;
2
Embedded Systems Innovation by TNO and Radboud University, Netherlands
;
3
Embedded Systems Innovation by TNO and Thales Nederland, Netherlands
;
4
Altran, Netherlands
Keyword(s):
Model-based Engineering, Domain-specific Languages, Simulation, Validation, Threat Ranking.
Abstract:
Changing an established way of working can be a real headache. This is particularly true if there are high
stakes involved, e.g., when changing the development process for complex systems. New design methods,
such as model-based engineering (MBE) using domain-specific languages (DSLs) promise significant gains,
such as cost reductions and improvements in productivity and product quality. However, transitioning between
design methods comes with a great deal of uncertainty, as any approach has associated pains. While the gains
may be intuitively appreciated, it may be less clear what the new pains will be and whether or not they will
cancel out the gains. For this reason, it may sometimes feel safer to stick with the devil you know than to meet
the one you do not, preventing the full design potential of the company from being reached.
This paper is an experience report from an investigation into how to mitigate the pains associated with a
transition to a model-based design flo
w using DSLs. The main contributions of the paper are: 1) a list of
14 pains related to MBE as a technology that is representative of our industrial partners designing high-tech
systems in different domains, 2) a selected subset of six pains is positioned with respect to the state-of-the-practice,
3) practical experiences and pain-mitigation techniques from applying a model-based design process
using DSLs to an industrial case study, and 4) a list of three open issues that require further research.
(More)