DR. K. WINKELMANN
HOHENBRUNNER STR. 27B
Comments on the POP-ART project proposal
This report is written as part of the INRIA evaluation process for the proposed project
POP-ART . It was requested by
François Sillion, président du comité des projets
General scope and direction of POP-ART
The project proposal presents a comprehensive view of the field of designing safe
real-time control systems, from the perspective of applying scientifically well-
grounded engineering principles.
More specifically, it comprises the areas of
• discrete as well as continuous control,
• mapping algorithms to implementations using real-time operation systems,
including multi-tasking and scheduling,
• fault tolerance,
I agree with the proposal authors on the high relevance of these topics for a wide
variety of applications in modern information and automation technology, including
systems in aerospace, robotics, communication, automotive and others. In particular,
the topics are well aligned with the INRIA research themes as quoted in the proposal
– namely production of high-quality “safe” software, and automation of complex
systems in several domains.
The proposal is strongly based on the existing experience of the authors, in particular
the results from various R&D groups in synchronous programming languages, state-
of-the-art compilation techniques, and the control architecture and domain knowledge
from the BIP robotics project.
While the classical theory of feed-back control for continuous systems is well
established and understood, there is in modern industrial applications a demand for
very complex discrete controllers, as well as controllers combining discrete and
continuous components. The synchronous languages provide a very powerful
framework to describe such phenomena. Based on this approach, the project’s
background spans theoretical advancements, such as synthesis algorithms, as well as
practical aspects, such as the various programming languages that are used in today's
programmable logic controllers.
The general presentation of the existing solutions and the open problems in these
fields demonstrates that the researchers are on the leading edge of these technologies.
Rather than basing their proposed work on a single narrow method and corresponding
tool development, they aim at building on the rich body of existing techniques and
software, especially in, but not restricted to the French community of synchronous
language research. This seems to me a wise decision. It represents a good balance
between the extremes of a from-the-scratch approach based on one single paradigm
on the one hand, and a universal openness risking incoherence on the other.
Planned research activities
The proposal describes a long-term research area, for which it is naturally not
adequate to present a detailed project plan for the full duration. Rather, initial research
activities are presented, each of which can be expected to produce tangible results
within one or two years.
These activities are well positioned along the “axes de recherche” as presented,
• Common frame-work, including languages and modelling,
• Implementing synchronous programmes,
• Scheduling and control,
• Controller synthesis.
These axes, together with the initial activities provide a reasonable basis for
monitoring the progress of the project in broad terms.
However, some of the initial activities should clarify their goals to become a bit more
verifiable. It is acceptable that some activities have as objective “d’etudier...” this and
that, but – being influenced by my industrial background – I would prefer somewhat
more precise goals in the sense “by month so-and-so we want to achieve a new
solution to such-and-such”. Even if some of these will fail, such measurable goals
could be expected to drive and focus the actual work.
Scientific and industrial Impact
An important factor is the given industrial relations. It seems the group has a number
of interesting connections, and it can be hoped that these are indeed exploited well,
both to work on practically relevant problems and to transfer R&D results into real
Even stronger seem their scientific contacts. The group members are well linked with
all the important French work-groups in their field, and some abroad. They present
together 37 recent scientific publications, and have contributed to organising several
conferences and workshops.
The scientific record of the proposers, and the responsibilities that they have mastered
already, including their involvement in the BIP project, demonstrates the good
leadership, vision and dynamism required to direct an ambitious long-term research
project such as POP-ART.
Klaus Winkelmann received his Ph.D. in computer science and a degree in
Mathematics at Erlangen University. Currently he is a project manager in the Infineon
CVE formal verification team.
Before joining Infineon Technologies in April 2002, he worked for Siemens
Corporate Research on formal techniques for the design and verification of embedded
software, in particular verification and synthesis of finite automata, synchronous
languages, discrete event systems and their applications to industrial control. He holds
a patent on controller synthesis.
Prior to that, he worked on the application of AI techniques to problems of automation
and control, diagnosis expert systems, configuration of automation equipment and
software configuration management.
He acted as reviewer of several European projects, as project manager for the ESPRIT
project SACRES, and as technical director for 5th framework project SafeAir.
As French is not my native language, I assure that I understand it well enough to
evaluate a scientific document.
Source of information - Reviewer’s relation to the proposers
The present report is based on the reference . I had the opportunity to discuss its
content briefly with Eric personally in January 2002. The other proposal authors I
have met only briefly or not at all.
I have known Eric Rutten since 1995 from several scientific contacts, notably the
SACRES project (1996-1999) funded under the ESPRIT program. On behalf of
Siemens AG, Corporate Technology, I acted as project leader for the consortium,
while Eric, at that time with INRIA Rennes, contributed to the technical success of the
project in the area of modelling embedded systems, based on his knowledge of
synchronous languages. The industrial relevance of that work lies in the capability to
unify several different design styles incorporated in state-of-the-art industrial design
tools, and in opening the door to formal analysis of such designs. The practical
application domains range from automotive to avionic systems as well as robot
control and other complex embedded systems.
The proposal defines an important and active research area, well in line with INRIA’s
objectives. The team is well qualified for the kind of work proposed, and has good
industrial and scientific contacts.
The proposal could, in my view, further benefit from spelling out some short-to-
medium term goals more explicitly.
I believe Eric Rutten will – based on his firm commitment and open-minded
personality – easily motivate students and co-workers for the POP-ART research
objectives, and make the project a valuable contribution to the area of real-time
 Alain Girault, Gregor Gößler, Eric Rutten, Daniel Simon: “Proposition de
projet INRIA, thème C, POP ART, Programming and Operating systems for
Applications in Real Time, Contrôl-commande temps-réel sûr”, http://www-
http://www.inrialpes.fr/bip/people/rutten/pop-art.ps.gz (Version Janvier 7,