A Look Back at ECRTS

In July, we held another successful ECRTS, with high quality papers, great sense of community, novel concepts, and the same spirit as in previous years. Behind the scenes, however, significant changes happened. The group of persons responsible for the development and organization of the conference, the Executive Board of the Technical Committee of Real-time Systems of Euromicro, has changed, and I have stepped down as Chair of the Technical Committee.
Time for a look back.

In the late 1990s after around 10 editions, the real-time workshop of Euromicro was about to be cancelled, numbers had been going down for a while; only authors presenting papers were attending, etc. Around that time I became chairman of the Euromicro Technical Committee of Real-time Systems with the goal of saving one of the then very few European outlets for real-time systems research.

I believed quality was key; unpopular decisions needed to be taken with only the good of the conference in mind. As I had been mainly active in IEEE conferences, I was familiar with the major researchers in the area and managed to get a number of them involved in an executive board. With this support, we established quality procedures for paper selection and conference organization, and communicated these to the broader real-time community. This boot strapping worked and we started to get increased submission and participation numbers.

In 1999, we felt procedures were strong enough to change name and format to conference, then Euromicro Conference on Real-time Systems, ECRTS. There had been warnings in the beginning, that quality orientation would endanger the workshop, as rejecting papers would reduce author participation, but soon the numbers showed that participation rose, and for a while correlated to the number of submitted papers, not accepted ones.

Once the quality of the main conference track was on track, we started to include other formats, such as work-in-progress session, keynotes, and workshops. These help to increase the value of attending the conference and to serve the community better: workshops, for example, can focus on topics complementing the main track, provide a forum for quick exchange of ideas of particular communities. Thus the conference can dynamically explore and foster new topics without compromising the main track.
The workshop on worst case execution time analysis is a good example, it helped to establish and became the main meeting place for the WCET community. Overall, workshops covering over 10 different topics have been held so far.

The names of the members of the Executive Board of the Technical Committee were not made public initially, to ensure members would join for their contributions, not the formal status. Relevant in the beginning, this could have changed a while ago. I strongly believe diversity of opinions in such a group is essential: conferences, evaluation processes, and the humans involved are too complex for a single mind to understand. Differences of opinions, which can appear to delay simple things at first, allow looking at more angles of an issue and to get more experience on the practicality of different solutions and confidence in decisions.

ECRTS continued to innovate and has been at the forefront of recent innovations in the real-time community such as firm deadlines, Work-in-Progress Poster sessions, Call for Actions, Artifact Evaluation, Industrial Challenges, etc. Many can now be found at the other conferences, e.g. Outstanding Papers, recently the Industrial Challenge or Real-Time Pitches.

The workshops and other additional formats turned out to be a great way to involve junior persons, who brought in most of the new ideas we are proud of now. Providing logistics and experience of the organizers allows new ideas to be tried out with modest effort and the people involved to get to know each other. ECRTS prides itself to have familiarized some junior researchers with organizational roles, who have become leading figures in the real-time community.

Proceedings had been published with IEEE from the beginning, although the IEEE was not involved in any way. This had been good in establishing the quality perception of the conference, but also countered our goal of wide dissemination via paywall, as e.g. pointed out by the Linux community involved in the conference. We felt this wrong on a number of levels, in particular that it excluded parts of our community: both authors, who may not be able to finance high open access fees by the IEEE and readers, who may not be able to pay to read, and so we looked for a different publication model. There were warnings against this, with some pointing to “IEEE publications” being important for evaluations etc. somehow conflating “IEEE” and “high quality”. We realized the choices were: (i) giving in to the argument and consequently staying with the status quo indefinitely or (ii) taking a leap of faith and working hard to establish the quality aspect of ECRTS on its own merits. We teamed with LIPIcs–Leibniz International Proceedings in Informatics to provide genuine open access to the proceedings since 2017, with processing charges covered by conference fees, removing restrictions on author and readership.

ECRTS considers the real-time community to be constituted by everyone working in the field of real-time, not just the regulars at conferences. We tried to reach out to parts of the community not yet involved, including support of regions with travel difficulties, industrial opinions and results, researchers not from well established groups.
Such inclusiveness is not only a matter of fairness, but also strengthens the community, e.g., by having results with major real-world impact published at our conferences, which has not always been the case. We made efforts to include research topics not represented adequately in the conference. Obviously this included outreach, via persons, papers, workshops, but also fairness of topics via identification of the specific contributions and evaluation criteria for various types of research and papers.

One important measure towards such inclusiveness is transparency of the paper evaluation process, e.g., transparency and communication of the actual evaluation criteria, and having same text (Call for Papers) as the basis for authors as well as PC members, reviewers, and discussions in the PC. Then, everyone knows and uses the same criteria.

Although we learned and took best practices from other conferences, especially in the initial phase of the conference, we did not just copy from them. Rather, when looking for solutions to problems, we took a different approach, not only driven by numbers: we try to identify the actual problem first, then come up with measures to address it. Bad review quality, for example, is more often a problem of bad program committee composition, guidance, or processes, than just a low number of reviews. Adding more reviews is unlikely to change the underlying quality problem.
In general, I believe quantitative approaches to quality problems will result in higher workloads in shorter conference timelines, a burden in particular on junior PC members; which needs to be carefully balanced with expected improvements in quality.

What used to be a small single track only workshop about to be cancelled has now become a major event, one of the top tier quality conferences in the field, with impact on other conferences, and a benefit to the entire real-time community. It has been a privilege and pleasure to lead these efforts.
All this could not have happened without the support and work of so many. Sincere thanks to the previous Executive Board, Sophie Quinton, Isabelle Puaut, Alan Burns, and Sanjoy Baruah, for the support and many good thoughts, and the large number of organizers who devoted their efforts to running the conferences; working with you was fun.

I feel deep gratitude and joy that a new group of persons has taken on the responsibility as new Executive Board, so we can leave knowing things are in competent and enthusiastic hands. Now I can join you in attending ECRTS and just enjoying the conference. I hope to meet you there soon.

Author Bio: Gerhard Fohler is Professor for Real-time Systems at the Technische Universität Kaiserslautern, Germany. He was Chairman of the Technical Committee on Real-time Systems of Euromicro, which is responsible for ECRTS, the prime European conference on real-time systems.