What I Owe to Jean-Raymond Abrial

When did I discover Jean-Raymond Abrial? 

Perhaps it was when I purchased the proceedings of the Cargèse conference and read his article "Data Semantics."  

Or was it while reading the publications of EDF's DER-IMA (located in Clamart)? 

 What is certain is that I discovered Jean-Raymond Abrial in the Meyer Baudoin, (Méthodes de programmation) Programming Methods at Eyrolles, then in the Delobel Adiba, Bases de données et modèle relationnel, (Databases) at Dunod.

  When I started in "business information technology" (at the "Computer Science" IUT), I found that what I was reading was hardly "scientific." "Information," "data," "rubrique," etc. For "data," I learned the theorems of Jean-Louis Rigal, who was a professor at Dauphine University. Easy: Th. 1, "Data is not given" because it is the result of an abstraction. Th. 2. "Data is not given" because once given, it is no longer data (it no longer changes the state of your knowledge. Oh! I just introduced another term!). It's no longer a scoop. Journalists understood this well. Th. 3, "Data is not given," it costs," it's not cheap! Even with the Web. Someone pays. 

 

 In 1970, I read and reread "Computer Science" by Jacques Arsac. I at least understood the difference between "information" and "knowledge" (according to J. Arsac's definitions).

Paul Namian of the Cnam had written an article in the journal Automatisme volume XI, n°10, October 1966 "Theoretical approach to the processing of administrative information" also trying to see clearly. I was lucky to be in the same premises as the ENSM (now École Centrale)! Claude Pair "The Structures of Information" (Alès, 1971) We used to talk about "methodology" (when we weren't actually studying methods; today, methodology is more chic than "method"!) Who would dare publish a journal like "Work and Methods" (where I've published quite a bit). We used to talk about "analysis," then it was "design," conceptual analysis; the word "structure" was everywhere, model too. But it was Michel Serres, invited to Nantes by Jean-Louis Gardies, who enlightened us with the example of the fable The Wolf and the Lamb, on Structure and Model. 

 This reminded me of my time at the barracks (yes, I went to the barracks very young—we used the plural—in Saint-Yrieix-la-Perche!). A instituteur (teacher) (now called "professeur"( "primary school teacher") asked us, "What do you want to be when you grow up?" A student replied, "Chartered Accountant" (his father was an accountant). Another then replied "expert hairdresser." is father was hairdresser. In the 80s, he would have replied "designer hairdresser," and nowadays I see signs with "Coiffure Concept" and other variations.

Since I'm talking about teachers, let's also talk about professors. It was the philosophy professor who told us about truth tables and sets. Admittedly, it didn't go very far! He explained the empty set using the symbol on the no-parking sign. 

 

 I mentioned Z. But there was also the DBMS (I'm making an anachronism, we didn't use the term yet) Socrate by J.R. Abrial, Socrate also discovered in the Delobel Adiba. Socrate was used under that name and then under Clio at the Indret arsenal, not far from Nantes. 

 

 When I arrived in Nantes in 1971, I taught the Cantor method. It had nothing to do with Cantor. But it ranged from the study of existing systems to the generation of Cobol code. I had learned it while interning at ICL, which used it. 

 

In 1982, in the bibliography of "Présentation du projet Thérèse" by J.M. Pendibidu, TSI, Vol. 1, N°3, May-June 1982, pp. 253-257, we found "Yaka, Fer, Setou, Bon and Yapad Bug, A metatheory of extended alfa-romero calculus: Very private communication (1979). We will talk about TSI again later. Let us recall that Georges Perec published in 1974 Experimental Demonstration of the tomatotopic organization in the Soprano (Cantatrix sopranica L.). Everyone noted the long bibliography. Jean-Raymond only made very small ones! But he read the texts included in his bibliographies.

 

 

After reading Z, I bought the first volume of Bourbaki with its durable hardcover binding. I didn't manage to get very far (but I must have gotten further than in Ulysses by the guy from Dublin! Because this Ulysses isn't exactly Joyce!). But at the end, there's the Historical Note and the small booklet of results. And I was able to read that. It's unfortunately much more readable than some of the textbooks that were distributed when the math reform was launched in education. Jean-Raymond talks about this booklet in his lecture at the Collège de France. 

 

Perhaps it was around this time that I became interested in the SETL language, which was developed at New York University. My memory tells me it was by the French. Wikipedia (English, translated into French) doesn't tell me. I had to give a tutorial to students in SETL. 

 

 I don't remember if I taught Z (the Oxford Z) before or after discovering NIAM (Nijssen Information Analysis Method) of Control Data in Europe. NIAM is the use of the binary relational model. I published the first book in the world (!) on NIAM (but also dealt with the n-ary relational model and others) in 1988. I had spent a year on delegation at CMILACO (Crédit Mutuel Informatique Loire Atlantique Centre Ouest) where NIAM was used. In 1985, Jean-Pierre Giraudin and Monique Chabre-Peccoud published in Bigre+Globule "For a rehabilitation of the binary relational model". It was Jean-Pierre who saved the 2025 conference by preparing a presentation on Z, Socrate in Grenoble.

 

1984 J. Raymond published in the journal TSI, "Specifier ou comment maître l'abstrait", Vol. 3, No. 3, pp. 201-219. The article begins with a boxed quote from Proust. Normal for someone who called a DBMS Socrate. But that's not done! You have to read the introductory box of the article by an academic! The article is impossible to find these days. Calling all! But I kept the introductory box on a slide. 

 

 1985 Jean-Raymond gave a lecture at CEPIA in Rocquencourt. Guy Laffitte from INSEE listened to him. And then he frequented Rue des Plantes in Paris, where Jean-Raymond lived. We could have crossed paths at one of these two places. It was probably in 1985 that I visited Jean-Raymond on Rue des Plantes. I remember that before dinner he showed me his proof booster. I also remember that he was spending long periods in Oxford at the time, long enough, he told me, to absorb the English accent. In the B-Book, Jean-Raymond wrote: "G. Laffitte influenced this work by his careful reviews, his accurate criticisms, and the sometimes very serious rearrangements he proposed for some of the mathematical developments in this book." 1986 I read in the bibliography of my Introduction to Specification (Masson), 1993, A Small Case Study in Program Design, Nov., unpublished. This must therefore have been the first year of my exchanges with Jean-Raymond.

 

In 1987, I read "The Case of Proportional Elections in LCP, Deductive Programming, J.R. Abrial's Method," IUT Nantes. Claude Pair had addressed the case in deductive programming. For LCP, it was Alain Coulon and M. Koutchouk de Bull. I also read "A Formal Introduction to Abstract Machines," published by the author on May 4, 1987. If I'm not mistaken, it was in 1987 that a CNAM engineering diploma was awarded with the subject "Four Specification Methods," SADT, Merise, F1 Formalism, Abrial's Mathematical Specification, Application to an Example, the Electronic Alarm Clock." Jean-Raymond provided his writing. One day, perhaps the day of the defense, he came with his tools to give us a demonstration. He must have been working at the CNAM in Paris at the time. In 1988, B User manual, Tech. Report, Programming Research Group, Oxford University 

 

In September 1989, I seem to remember that Guy Laffitte came from nearby during my third MOACSI conference to present Jean-Raymond's text "A Formal Approach to Software Construction." 

 

1996 (or '95?) The book I'm talking about later was published in '96. I went to give a course on NIAM at the École Polytechnique Fédérale de Lausanne. And I remember that J.R. Abrial had received a request to teach there. A colleague had gone in his place. The course was published in a book, Génie logiciel: principes, méthodes et techniques, Presses polytechniques et universitaires romandes. A colleague from EPFL had sent me my course rewritten with mathematical notation. This reminded me of a course I had given to math teachers on the Merise method. During the break, someone came to me and rewrote the CDM (Conceptual Data Models) in mathematical notation. I confirmed that he understood everything!

 

In NIAM, a graphical notation with two symbols is used to express the 16 cases of binary relations. The V, which intersects a line, provides the universal quantifier ("for all") for the totality constraint, and the line for the uniqueness constraint (as in Codd's n-ary relations, the key is underlined), and a few others for the constraints of equality, inclusion between relations, domains, and codomains. The important thing is that we must write at best two subject-verb-complement sentences for the relation and its inverse. It is then easy to move to an "n-ary relational model," which in NIAM is not called a "logical model" (why "logical"?!), but a "grouped model." I see that in my 1988 book, I cite Data Semantics 1974 and "Proof in Specifications" in "Prototyping and Software Specifications," Journées Afcet, Insa Lyon, January 1987. I note that Nijssen was one of the organizers of the Cargèse conference. In 1993, a philatelist colleague and I were specifying the Yvert and Tellier Catalogue in Merise, Niam, and Z. 

 

One day, a colleague from Teesside University in Middlesbrough (the city that kept its transporter bridge while Nantes lost it) toured the universities where Z was taught. He went to Grenoble and then came back to Nantes. Z was the lingua franca for several teachers. Through Erasmus, there were several meetings of teachers and students from different countries where Niam diagrams and then Z were created. And in the UK, Robin Milner's CCS was developed. Then colleagues from Teesside (Bill Stoddart, Steve Dunne, Andy Galloway) taught B and published on B. And went to see Jean-Raymond in Marseille.

 

In 1994, we were two years away from the publication of the B-Book. Jean-Raymond Abrial came to Nantes to teach APPC students (am I making an anachronism here again? Année Post Premier Cycle followed "Année Spéciale"), a course that was videotaped at the IUFM (University Institute for Teacher Training) in Nantes, which had a studio, and distributed by the IUT of Nantes and the publisher Teknea in Toulouse. There are 6 VHS cassettes introducing B and 8 case studies, accompanied by copies of the slides. A very educational course. Jean-Raymond used the papygrams. On this subject, I remember that when I provided the students with the Merise notation, I told them to ask their contacts in companies to provide them with the papygrams. And that they would realize that... When I met them on the train, they reminded me of my advice. And that I was right. They had tested it. J. Raymond also explained the refinement theorem very well. I was then able to teach a B course. 

 

 I remember telling him about Niam's simple notation. But Z notation was then widely used and had even been standardized. I presented it at the 2025 Nantes conference. It uses only the stroke --------, the V, and the points > or <. The less constrained a relation is, the fewer symbols we have. Any relation is represented by ------. The most constrained is represented by <-V---V->. We called it Oc, in reference to Occam and his razor... and to Occitan.

In 1995, Jean-Raymond and I had planned the first international conference on B in Nantes. But the publication of the B-Book was delayed. So I changed the name of the conference. It was shortened to Z2B (it was fashionable to use the two to make "to"), and then to "Z Twenty Years On - What is its Future?" ISBN 2-906082-19-8 On the occasion of this conference, I created the APCB association, the Association for the Management of International B Conferences. In November, 1996, the first Conference on the B Method was held in Nantes. Proceedings: ISBN 2-906082-25-2 

 

1995-1996 was also the year of the BUG, the B User Group attended by Atelier B users. I remember Ranan Fraer, who helped out a lot of people. I met Ranan at the EDF, CEA, INRIA Summer School at the Château de Bréau-Sans-Nappe from June 12 to 23, 1995, where Fernando Mejia distributed a document entitled "The Method and the B Language." Ranan was at Atelier B day and night. And I learned from him that it was the first time he had ever had access to Atelier B! On his page (still available: it's rare!) https://www-sop.inria.fr/croap/personnel/Ranan.Fraer.html I read: "Ranan Fraer Formal development in B of a Minimum Spanning Tree Algorithm distinguished with the "Best Paper Award" at the First B Conference, Nantes (France), November 1996."

 

So he had to leave with the B-Book! 

 

 And "I'm also an enthusiastic supporter of Jean-Raymond Abrial's B method. There's a growing interest in B, as it has already been successfully applied on several large-scale industrial projects in France and the UK."

 

1996 was the year of the Steam Boiler Control Specification, at the Château de Dagsthul in the north of the Saarland.  

 

And of the first international B conference in Nantes. Jean-Raymond announced what would become the B event with his presentation entitled: Extending B without changing it (for developing distributed systems). 

 

In 1997, at the AFADL Toulouse, J.L. Lanet spoke following Jean-Raymond's presentation. Gem+ would use B for smart cards. And later, I would spend a weekend in Marseille with Jean-Raymond after an IUT student's internship at Gem+ was defended. I haven't forgotten. During the defense, there were PhD students who discovered that the other student... was at a university technology institute (IUT)! I also discovered the calanques by following Jean-Raymond. We went to the Luminy university and took the bus back. When we arrived at the Vieux Port, I had a lot of difficulty walking. The next day, I spent it lying down. And Jean-Raymond left for the calanques again in the morning. He had just completed a desert expedition. Was it the particularly grueling Tamanrasset-Djanet camel trek where, on the first day, one of the group members died and his body had to be carried throughout the expedition?

 

In 1997, the Computer Science Department of the IUT of Nantes published the book "The B Method, Concepts, Approach, Examples of Applications" by Karl Emeriau, ISBN 2-906082-26-0, written following his CNAM engineering thesis. Karl was self-taught in B thanks to the B-Book. All the applications are so-called "management" applications developed with Atelier B, right down to the automatically generated C code. 

 

1997 was also the year of the 1st IEEE International Conference on Formal Engineering Methods, which took place in Hiroshima (M.G. Hinchey, Shaoying Liu, Eds). I had submitted a paper written with a colleague, "Formal Specification of Dynamic Constraints with the B Method." I knew we had a problem (a theorem to prove, I think). One day, I received a call from Jean-Raymond. I was ashamed. I knew he was a member of the program committee. I immediately told him I shouldn't have sent the paper. He had a problem. But Jean-Raymond told me, "I'm calling to tell you that all you need to do is..."—a tiny correction, but an essential one! I anonymously thanked a reviewer in the article. I had to cancel my trip at the last minute. In Japan, I would have met up with several English friends I made through Z and then B. 

 

In 1998, Claude Boksenbaum organized the second B conference in Montpellier in April. Didier Bert edited the proceedings, the first in Springer-Verlag's LNCS. In 2000, the first B and Z user conference was held in York in September. The editors were Jonathan Bowen, Andy Galloway, and Steve King.

 

In 2001, I published "Formal Specification with B," Hermes Lavoisier, ISBN 2-7463-0302-2, with the collaboration of Jean-Yves Lafaye and Marie-Laure Potet, with a Foreword by Fernando Mejia (Manager of Software Development Methods and Tools, Alstom Transport Signalisation) which begins as follows: "I had been experimenting with algebraic specification methods for some time, at what was then called Bull Systèmes, when I discovered, in 1985 or 1986, during a training course given by Jean-Raymond Abrial, what would become the B language and method a few years later. And it was love at first sight!" He tells us his story of B. 

 

In 2003, the 3rd Z and B User Conference was held in Turku, Finland. Turku, where Ralph-Johan Back, who is often cited in Jean-Raymond's articles, practiced at Abo University. It was in June, a time when the nights are not dark. Note that it is a Swedish-speaking university. Of course, we met Jean-Raymond and many others there.

I have preserved the exchanges on the B Forum managed by INRETS forum B from 2004 to 2008. Here are the names of the participants whose messages I have preserved: Michael Leuschel (HH Univ. Dusseldorf), Samuel Colin (Loria), Simão Melo de Sousa (Universidade da Beira Interior), Guy Vidal-Naquet (Supelec), Georges Mariano (INRETS), Marc Guyomard (Enssatt, Lannion), Steve Dunne (Teesside University), Dominique Cansell (Loria), Daniel Zingaro (McMaster University), Ken Robinson (UNSW Australia), Ib Holm Sorensen (BB-Core), Elisabeth Ball (CSE Sotton), Jeremy L. Jacob (York University), Yann Zimmermann (Keesda)  

 

2004: Start of the Rodin project with Laurent Voisin. Jean-Raymond was a professor at ETH Zurich until 2009. where he will work on Rodin with Laurent Voisin and others. He talks about it in his lecture at the Collège de France. 

 

In 2008, I will retire in October. I will organize the conference "The B Method, from Research to Teaching," where Jean-Raymond will be attending. In 2009, the second conference "The B Method, from Research to Teaching" will take place. 

 

In 2008, it will be the first ABZ conference at the British Computing Society in London. 

 

In,2014 I will attend the ABZ conference in Toulouse and meet Jean-Raymond there, who was returning from China. Also present, of course, were Egon Börger (he must not have missed one)—who was a visiting professor in Nantes for a few months—Leslie Lamport, and Tony Hoare. When I went to visit the Toulouse Aviation Museum with the grandchildren, I recognized the restaurant where the gala dinner was held.

 

2015 Jean-Raymond's lecture at the Collège de France. I remember him telling me that his laptop had been stolen before the lecture. 

 

2022 These were my last email exchanges with Jean-Raymond. He had just told me that he was in a wheelchair and in a suitable facility. He was reading Montaigne's Essays and was interested in Japanese civilization. I suggested we go see him, but he replied that he wasn't well. In his presentation at ABZ 2025, Dominique Cansell ends with two photos taken during his last two visits to Jean-Raymond. That was in 2022. But Dominique communicated with Jean-Raymond in subsequent years. 

 

 On May 25, 2025, the J.R. Abrial lecture, a pioneer in the scientific development of computer languages and formal methods and their large-scale application in industry, will be held as part of the Scientific Days of the University of Nantes. We learned of Jean-Raymond's passing on June 3, the day before the conference. 

 

And from June 10 to 13, 2025, the ABZ conference took place, where Dominique Cansell presented work done with Jean-Raymond a few years ago.  

 

This was 30 years after the first conference on Z and B in Nantes! 

 

And I never imagined a year ago that I would be publishing an article on B at a conference at almost 79 years old! (Formal Modeling of Information System Evolution using B) An article I found while preparing the conference on Jean-Raymond. My colleague had to go back to Atelier B because only bits of B specifications remained. And the tools have evolved since 2008! The article deals with applications where the rules change according to phases. This is the case, for example, with the academic year. It's also the case with changes in the law. 

 

 

Thank you, Jean-Raymond, for everything you taught me, from Socrates to Z, to B, to Event-B, to Rodin, for your teaching skills, for sharing your work with so many people, for your welcome, whether in Paris or Marseille. Thanks to you, I've connected with many people who have also said how much you've taught them. Thanks to you, I didn't work; I enjoyed myself. 

 

Henri Habrias