23rd Mach 1997, Göteborg, Scandic Crown Hotel.

Today a one day seminar titled "Development of automotive software - correctness and performance" was jointly organized by ASTEC, ARTES and Mecel AB. The purpose was to present and discuss state-of-the-art methods and techniques for development of correct and efficient automotive software. More than 85 persons, mainly from Swedish car-industry and academia, attended the seminar that was sold-out several days in advance.

After a short introduction given by Hans Hansson (director of ARTES and ASTEC-RT project), Stefan Poledna (general director of TTTech) gave a 70 minutes talk about time triggered technology and how it may be used in automotive software. He focused on Time Triggered Protocol (TTP) and how it differs from CAN that seemed to be more familiar to the audience.

In the next talk Ken Tindell (director of Northern Real-Time Group) presented the Volcano concept that has been developed within his company in collaboration with Volvo. Naturally, Tindell devoted most of his 40 minutes to present the concept and some of the developed tool support. He also related Volcano, which is based on CAN, to the TDMA based techniques presented in the previous talk. After a short lunch break he continued with a short demo in the auditorium where he showed how a reconfiguration of a Volcano communication system is done.

The last invited talk was given by Göran Anger (managing director of Industrilogik) who presented some of his experiences from using formal specification and verification techniques in automotive systems and software. Anger also talked about formal methods in general and about to what extent they have been applied in industrial projects. In particular he refereed to the Web-site Formal Methods Europe Information Resource - The Application Database (http://www.comlab.ox.ac.uk/archive/formal-methods.html) as a source of knowledge about application of formal methods and tools to industrial applications.

The second session after lunch was an overview presentation of the ASTEC-RT project. Hans Hansson introduced this hour by presenting some of the main goals. The project is focused on design of automotive software and execution time analysis. The introduction was followed by five very short presentations from co-workers within the project. Jan Gustafsson (MdH) presented some results from the research on execution time analysis, and Mikael Strömberg (Mecel AB) presented ideas about development of automotive systems. Next, Magnus Lindahl (Mecel AB) and Paul Pettersson (UU) presented the results of a case-study where the tool UPPAAL is applied to design a gear controller. Finally, Peter Altenbernd concluded the session with an overview presentation of the tools CHaRy and SEA.

During the last coffee-break of the day, demos of the two tools UPPAAL and Chary/SEA were given outside the auditorium. So many people showed interest in the two tools that the demos lasted for a longer time than originally planed (even if the following session started on schedule).

The last session of the day was a panel discussion on the topic "Provably correct software - future or utopia". In the panel was the three invited speakers of the day together with Bengt Jonsson (director of ASTEC), Mikael Strömberg (Mecel AB) and Mats Larsson (Volvo Technical development). The main topic during the panel discussion was how to bridge the gap between academia and industry and how results from projects such as ASTEC-RT may be used in the industry. The panel agreed on that companies like NRTG (Ken Tindell) and TTTech (Stefan Poledna) are good examples on how academia results can be transfered to the industry.

Author: Paul Pettersson (paupet@docs.uu.se), DoCS, UU.