How and Why ETAPS Started

The first ETAPS took place in Lisbon in 1998. ETAPS is now a well-established and very successful conference series, with ETAPS 2025 in Hamilton being the 28th installment and the first outside Europe.

I was the first ETAPS steering committee chair, responsible for coordinating discussions about the idea of starting a new conference series and decisions about what form it should take, and helping with the organisation of the first few installments. I hope that some members of the ETAPS community will be interested in my recollections of how and why ETAPS got started.

The situation in 1994

The story begins in 1994, several years before the first ETAPS.

In that year I was local organiser of CAAP/ESOP/CC in Edinburgh, a biennial conference series that fed into what would become ETAPS, and PC chair of ESOP 1994. The conference was a success, with a full programme of good papers and enough participants that the conference didn’t lose money.

But as the conference drew to a close, I discovered that there were no plans for its successor in 1996. There were signs that CAAP (Colloquium on Trees in Algebra and Programming—the acronym comes from the original French name, Colleque de Lille sur les Arbres en Algèbre et en Programmation) might not continue, but nobody present really knew or was taking responsibility. I was unhappy because I saw initiating the next installment as one of the responsibilities of the local organiser and PC chairs.

At that point in the discussion, somebody drew our attention to a position statement that had been circulated by Reinhard Wilhelm in the European programming languages research community a couple of years earlier:

The situation concerning European conferences in the areas of Programming Languages, Semantics, and Programming is not satisfactory. Different groups of scientists have established conference and workshop series which compete for a too small market. The competition of too many conferences for a rather small supply of scientific results has prevented any of the series to really reach a high international standing.

How to change things

1. We should stop to split a rather small cake into too many pieces; that is we should join efforts and merge some of the series.
2. We have to find some means to provide continuity for the series.
3. We should try to find a financial underpinning.

Reinhard Wilhelm, August 1992

What Reinhard—a prominent figure in the European programming languages research community, and famous as the founding scientific director of Schloss Dagstuhl—was referring to, was the complex kaleidoscope of European conferences series and individual conferences in this area that were active at that time. Here is a diagram showing the origins of the two biennial conference series that existed in 1994—TAPSOFT and CAAP/ESOP/CC—that eventually fed into ETAPS:

In 1994 there were a further 32 individual European conferences and major workshops in the same general area.

(The history of these and other European conferences is documented in their proceedings, almost all published in the Springer LNCS series.)

This was an unsatisfactory situation from several points of view. First, authors had too many choices of where to submit their latest paper. Likewise, people who wanted to attend conferences to learn about the latest results had too many conferences to choose from. Too many choices is not necessarily a bad thing. But the result was that there was no clear primary option to compete with the top international conferences.

This meant that organising a European conference was a risky business. PC chairs worried that all the competition would lead to not enough good submissions. Conference organisers worried about not enough fee-paying attendees, leading to a financial loss that would have to be covered by their institution.

Given the situation, Reinhard’s proposal made a lot of sense. Paul Klint of CWI in Amsterdam suggested the foundation of EAPLS as the basis for a possible solution:

Call for discussion: Should we restructure the European conferences in programming research?

Among some European researchers in the areas of programming languages, semantics and programming the concern is growing that this field is not adequately represented in a major, high quality, conference in Europe. The many conferences being organized in this and related fields […] are competing with each other in a too small market and therefore they have a hard job in building up sufficient critical mass to become competitive with the big conferences in the USA.

Can we join efforts and change this situation? In this note I propose to create a European organization (tentatively called EAPLS—European Association for Programming Languages and Systems—patterned after the EATCS) that aims at creating a platform of researchers that can restructure the European conferences in the desired direction.

Paul Klint, 1993

Paul’s proposal was seen as an important step towards addressing Reinhard’s wish for continuity and for financial underpinning, since a legally-established organisation was required to apply for funding from sources like the European Commission. However some people were uncomfortable about the “top-down” nature of the proposal and the relationship to existing communities, see below. And it left open the question of how best to build a strong conference series on the basis of some of the existing events.

Email during 1994 between various people, including Reinhard, Paul and the people behind CAAP, and face-to-face meetings in Santa Margherita Ligure and following the TAPSOFT PC meetings in Aarhus, led to the following message from Peter Mosses (chair of TAPSOFT 1995) and me:

European Spring Conferences on Theory and Practice of Software Science

The situation with European spring conferences on theory and practice of software science is widely regarded as unsatisfactory. There is no single event to compare with big U.S. conferences such as POPL. The result is poor conference attendance and too much competition by the various small conferences for good papers. A consensus seems to be emerging that a single annual joint conference on Theory and Practice of Software Science should be established in the slot currently occupied by CAAP/ESOP/CC in even years and TAPSOFT in odd years. This would comprise a number of existing and new conferences, covering a spectrum from theory to practice.

There will be a discussion on this matter at TAPSOFT’95 in Aarhus.

Peter Mosses and Don Sannella, 1994

At this point, the decision had already been made to claim the relatively empty period in the spring, the slot occupied by CAAP/ESOP/CC and TAPSOFT, in order to build on the success of these conferences and to avoid direct competition with the busy summer conference season. The confusing interleaving of biennial conference series in even/odd years would be replaced by a single annual event. And the area to be covered was referred to as “theory and practice of software science”—almost the same as “theory and practice of software development”, the meaning of the TAPSOFT acronym, while signalling a broadening of scope.

Deciding what ETAPS should be

Plans for ETAPS took shape during the discussion at TAPSOFT 1995 and by email after that. Since the hope was to bring together European research communities and to build on the success of existing European conferences, the main people involved were representatives of these communities/conferences:

• The theory community: CAAP (1976-1997), EATCS (1972-), FoSSaCS (1998-); Maurice Nivat, Wolfgang Thomas, André Arnold, Max Dauchet, Sophie Tison
• The formal methods community: TAPSOFT (1985-1997), FASE (1993-), nascent EASST (2000-); Hartmut Ehrig, Marie-Claude Gaudel, Fernando Orejas, Pierpaulo Degano, Tom Maibaum, Egidio Astesiano, Ugo Montanari, Peter Mosses
• The programming languages community: ESOP (1986-), CC (1986-), EAPLS (1993/1996-); Hanne Riis Nielson, Peter Fritzson, Tibor Gyimothy, Reinhard Wilhelm, Paul Klint, Alan Mycroft, me
• The tools community: TACAS (1995-); Bernhard Steffen, Kim Guldstrand Larsen, Ed Brinksma

Discussions at that point, 30 years ago, were by email or telephone or else face-to-face. The web was just starting to be used but at that point it was just for publishing, not for interaction. So we used it for recording the state of the discussion and publication of proposals. It was so new that Peter Mosses and I included instructions in our first email on how to access it in order to read the initial proposals.

One of the first decisions was that CAAP should be replaced by another theory conference along similar lines but with a broader focus (“computation structures” was Maurice’s proposal, and it stuck) in place of trees. This became FoSSaCS.

As already indicated, there was agreement on a “joint” or federated conference in which the conferences that joined would retail their identity but be organised as a single large event. It would be structured so as to encourage partipants to attend talks from any combination of member conferences they wished, and to maximise opportunities for cross-fertilisation. There would be five days for the main conferences, covering the themes

• Computation Structures and Foundations of Software Science, initially FoSSaCS
• Formal Methods, initially FASE
• Programming Languages and Systems, initially ESOP and CC

with talks in synchronised parallel sessions and invited speakers in plenary sessions. Satellite events would take place in the weekends before and after the main conferences.

The status of what is now TACAS was a point of disagreement. It started as a satellite workshop at TAPSOFT 1995, not having been part of TAPSOFT or CAAP/ESOP/CC before that, and so it was not included at the very beginning of the ETAPS discussions. Although arguments for inclusion were convincing, there were concerns about overlaps with the other conferences.

See this page for a copy of a web page from around the time of TAPSOFT 1995 that summaries the state of discussion at that point.

But there were some tensions

After agreement on the idea of a combined event, there were serious disagreements on how best to bring it into existence. Each conference that had agreed to participate represented a community, and there were worries from some of these communities about being “taken over” by the others. An aspect of this was concern about the role of the newly-formed EAPLS in the new event. Proposals for the name of the event played a role, with the desire of some to continue with the TAPSOFT label being viewed by others as confirmation of the “taking over” agenda.

The strong feelings are demonstrated by the following contributions to the email discussion:

I STRONGLY disagree with the following propositions of Reinhard.

The steering committee mentioned should be the EAPLS steering committee after securing appropriate representation of the TAPSOFT community.

XYZ is the flagship conference of the European Association of Programming Languages and Systems (EAPLS).

The steering committee of the European Association of Programming Languages and Systems (EAPLS) will decide whether or not to admit proposed new events.

A lot of people interested in TAPSOFT (attendees or organizers) have nothing to do with EAPLS. […] The aim of the new TAPSOFT is to gather several conferences about software science whatever community a conference was originated in. Organizing it under the banner of EAPLS goes in the opposite direction.

The steering committee must be constituted of people involved in the creation of this new event (and of its predecessors). I suggest that the steering committee requests organizations like EATCS, EAPLS and others to appoint a representant in the committee. If EAPLS needs a flagship conference, it could be a part of the new TAPSOFT.

1. I am absolutely against keeping the name TAPSOFT as a label on this conference week. it suggests too much of a take-over from TAPSOFT. Also, neither myself nor most of my colleagues go to the current TAPSOFT conferences, so TAPSOFT is neither well-known nor successful for us. Why not just call it ESOFT—European Software conference week. […]

2. As for version 2 of the draft proposal, I think it is reasonable, with one exception: the introduction gives too much of an impression of a new great conference that will swallow and replace the existing conferences. This is however not the case, which is made clear a bit further down in the text. Some modification in the introduction could change this impression. There should be an emphasis on a conference week rather than a new confederated conference.

Through discussion and compromises, with the obvious advantages of the proposal over the status quo as a strong source of motivation, agreement was eventually reached. The final issue was the name, which was decided by a very close vote in summer 1995 (each vote indicating preference on a scale 1–5):

The deadline for voting on the name/acronym has passed. The result, obtained by simply adding up all the numbers, is as follows:

87 European Conference on Software Science (ECOSS)
79 Theory and Practice of Software Science (TAPSOFT)
87 Theory and Practice of Software Science (TAPOSS)
110 THEORETICAL AND PRACTICAL ASPECTS OF SOFTWARE SCIENCE (TAPAS)
80 European Software Conference Week (ESOFT)
111 EUROPEAN JOINT CONFERENCES ON THEORY AND PRACTICE OF SOFTWARE (ETAPS)

Planning for ETAPS 1998

At this point, focus switched to concrete planning for the first ETAPS in 1998, which José Fiadeiro had already agreed to organise in Lisbon, an excellent location for an event in late March. José managed to secure the superb facilities of the Gulbenkian Foundation at no cost. I had been elected to chair the steering committee, and the PC chairs were:

• FoSSaCS: André Arnold and Wolfgang Thomas
• FASE: Egidio Astesiano
• ESOP: Chris Hankin
• CC: Kai Koskimies
• TACAS: Bernhard Steffen

Some critical early decisions were that deadlines for submission and acceptance would be synchronised and that publication of proceedings - at that time physical Springer LNCS volumes—would be coordinated. Participants would register for ETAPS, not for individual conferences, with registration including a choice of two conference proceedings volumes. There would be an extra fee for attending satellite events, but the fee structure would encourage attendees to stay for both the main conferences and their selected satellite events.

In the period before the first ETAPS, everybody involved in the organisation was a little worried about the number and quality of submissions it would attract, and the number of participants. But it was a huge success, exceeding all of our expectations. The facilities and the local organisation were perfect and publicity was excellent. There was a small list of good satellite events, enough to establish the viability of the concept. The number and quality of submissions and the number of participants confirmed the agreement of the scientific community on the ideas behind ETAPS.

Retrospective

The details of the arrangements for ETAPS have changed over the years, but the overall concept has worked well. Participants regularly move between parallel conference sessions, showing the value of co-locating conferences with synchronised timing of talks. It is a very challenging event to organise and all of the local organisers deserve our heartfelt thanks.

The differing size and strength of the participant conferences has been managed well. There has been some change in the set of conferences, but the original plan was for more organic change over time. I think it would be good to broaden the spectrum of ETAPS by adding at least one less theoretical conference to the mix. ETAPS has remained financially healthy, even when disaster (COVID-19) struck. That is partly due to the establishment of ETAPS eV, which I naively thought would not be necessary. Some travel funding for ETAPS participants came from the European Commission via EAPLS in the early days, but despite early expectations this has never been a significant source of support.

In 1995, as we were starting to organise the first ETAPS, I made the following prediction for how ETAPS would be described in the future:

Date: Thu, 14 Sep 1995 00:56:40
From: Don Sannella <dts@dcs.ed.ac.uk>

ETAPS is a loose and open confederation of conferences and other events that takes place in Europe each spring and covers a wide range of topics in Software Science. It receives sponsorship from various bodies and is organized by a steering committee composed of representatives from all of its constituent events and sponsoring bodies. Since its establishment in Lisbon back in 1998, ETAPS has consistently attracted the best papers in the area and attendance has become effectively mandatory for all active researchers in the field.

I am pleased and proud to say that most of my prediction from 30 years ago has become reality!

Don Sannella is Professor of Computer Science in the Laboratory for Foundations of Computer Science at the University of Edinburgh where he has spent his entire career, with the exception of a year as Professor at Universität Bremen. His research interests include functional languages, semantics of programming languages, foundations for algebraic specification and formal software development, and static analysis. He was founder and CEO of Contemplate Ltd, a startup company that produced static analysis tools for finding concurrency bugs in large Java codebases. He started programming when he was 12—not unusual nowadays, but in 1968 hardly any children had even seen a computer. His activity in conference organisation includes serving on many PCs and his involvement in starting ETAPS and organising the first annual ETAPS snowball fight (Warsaw, 2003), see the leftmost photo at the top of this article. He served 2001–2024 as Editor-in-Chief of Theoretical Computer Science, responsible for Part B (Logic, Semantics and Theory of Programming). He is a Fellow of the Royal Society of Edinburgh. In his spare time he enjoys skiing, gardening and DIY.