Copyright Notice

The following manuscript

EWD 577: Tripreport E.W.Dijkstra, ECI-conference 9-12 August 1976, Amsterdam

is held in copyright by Springer-Verlag New York.

The manuscript was published as pages 225–229 of

Edsger W. Dijkstra, Selected Writings on Computing: A Personal Perspective, Springer-Verlag, 1982. ISBN 0–387–90652–5.

Reproduced with permission from Springer-Verlag New York. Any further reproduction is strictly prohibited.


EWD 577

Tripreport E.W.Dijkstra, ECI-conference 9 - 12 August 1976, Amsterdam.

It was the kind of conference from which one returns with a glorious headache. It had been organized by the European Cooperation in Informatica, a joint enterprise of the Information Processing Societies of the various European countries. The response to the Call for Papers had been very low: only 36 papers had been submitted. Instead of cancelling the conference, the organizers decided that, by abolishing parallel sessions and selecting 12 of the submitted papers, the conference could still be held. (After all, they had six invited papers in addition to the submitted ones!) In spite of the meagre program, they still managed to collect about 250 participants. It is very questionable whether, with those 250 participants, they have reached the break-even point; if they haven't, I just cannot have pity with them: it would serve them right, for the organization of a bad conference is worse a crime than not organizing a conference at all. (Besides that, they had been warned.) The organizers have committed what I would like to describe as "Contempt of Audience". It saddened me to see the extent to which the degrading circumstances at the European universities have broken the spirits of my colleagues, it saddened me to observe their apathy and indifference, in which they have "learned" to accept the junk as if it were the real thing. The most frightening aspect of it all is that most of them present their abolishment of all norms and (time-honoured!) quality standards as an act of great wisdom.

The conference was held in the new building of the Free University of Amsterdam. I had never been there before. It must be about the last university building erected during the time that our government thought that it was infinitely rich. It was in many respects an ideal setting, in a most important aspect, however, it was not: at no moment one got the impression of being in a seat of learning and a centre of culture, on the contrary! I walked through the University Bookshop, but it was of a shocking vulgarity —as bad as the books department of the Bijenkorf (i.e. a chain of large department stores, oriented towards the superficial, fashionable consumer)—. Comic strips that were not funny, political pamphlets full of cliche's, science fiction books and pamphlets of obscure mystical cults. And, to top it all, when we entered the Auditorium a few minutes before the conference started, there was... Muzac! I was flabbergasted. (In my previous tripreport I have made a remark about differences in noise level at both sides of the Atlantic: from that comparison, I learned, at least Amsterdam should be excluded. While portable radios are strictly forbidden in most public buildings, such as railway stations and the Eindhoven University Campus, the bar at the lounge of the Free University in Amsterdam did not observe that rule! The whole place was of a disgusting vulgarity.)

The first afternoon was devoted to Program Development and Verification in Practice and Theory. Michael Jackson gave his talk as an invited speaker. It was well-presented and certainly falls under the heading Program Development in Practice, but it had little to contribute beyond the rather limited problem area of file processing in (or: in spite of?) COBOL. Then Antonio Salvadori from the University of Guelph, Ontario, described the "Guelph Efficiency Monitor, a preprocessor system which can analyse a COBOL program at any development or running stage". I quote:

"The statistics gathered and printed consist of

The last speaker was P.Hammersley, Cambridge UK on "Team Organization in Integrated On-Line computer Projects", a talk that was well-covered by its title. His English was a pleasure to listen to, but it was a talk with little or no technical or scientific content and I —as many others— found my thoughts wandering away.

In the evening there was a reception by the State Secretary of Interior Affairs and Burgomaster and Aldermen of Amsterdam; the reception was held in the beautiful surroundings of the new Vincent van Gogh Museum, where the air was polluted by... audible wallpaper! Most of us looked back on a wasted afternoon and were worried whether the conference would get any better.

The next morning was devoted to Concepts and techniques of Database Management. The invited speaker, C.J.Date (IBM General Products Division) was that morning's best speaker. He knew what he was talking about and gave what seemed to be a good overview. For those unfamiliar with the topic, his talk was quite instructive.

The afternoon was devoted to Computer Networks. Louis Pouzin (IRIA) was the invited speaker on "Names and Objects in Heterogeneous Computer Networks". He had originally prepared a more technical presentation than in the meantime he dared to give and in a hurry he redesigned his presentation; as a result it was a bit rambling, but it was still quite clear that he knew what he was talking about. I had heard him many times before, but unaware of his past as a telephone engineer: that past was mentioned in the introduction and was quite discernable.

On Wednesday morning —again Program Development and Verification etc.— I opened the session as the invited speaker. Tony Hoare should have been that morning's session chairman but, Tony being prevented from attending, his role was taken over by L.A.M.Verbeek of Twente University. I did not present my material well. I knew that it (EWD563) was difficult to present, as I wanted to show what I had learned from a highly formal experiment that had recently taken me more than a month to conduct, lessons that I could not explain without sketching the experiment, an activity that, indeed, took too much time and was only partly successful. (The trouble was that, while trying to do it, I noticed this!) In retrospect a very simple tutorial —nothing new— on formal program derivation would have been more appropriate for this occasion. Afterwards I.D.Ichbiah (CII) showed a similar —but less formal— case study of program development; he reached his audience better I think. Then Shmuel Katz (IBM Israel Scientific Center, Technion City, Haifa, Israel) gave a well-prepared talk on "Program Optimization Using Invariants". I think that it told more about the role of science and technology in the state Israel than about computing science as such.

Wednesday afternoon was again devoted to Database Management. Rudolf Bayer (Technische Universitaet Muenchen) was invited speaker on "Integrity, Concurrency and Recovery in Databases." His talk was well-prepared and well-presented (but for the fact that he tried to show too much). I was not convinced by his proposed solution: it was complicated and his "conclusion" that deadlock prevention was impractical did not seem to me to be sufficiently justified. I have more the impression that he had failed to discover how to do it. Later, in private, I had a long discussion with him about the current database folklore. It was very instructive for me; he, at least, is willing to challenge the common tacit assumptions, even if they have already found their way in standard proposals! (So does Michael Jackson, who repeatedly expressed his strong fear that via the standardization bodies already identified mistakes will be "cast in concrete".)

The afternoon ended with a Panel Discussion on Database Management with P.J.H.King (United Kingdom), G.M.Nijssen (Belgium) and A.A.Verrijn Stuart (The Netherlands) as panelists and T.W.Olle as chairman, who found it necessary to address in public the panelists by their Christian names. It was just terrible! A few weeks earlier Olle had sent six questions to the three panelists and each question was answered by each panelist. For more than an hour we had an eighteenfold demonstration of the well-known saying "If you ask a silly question, you'll get a silly answer." In our innocence we thought that this panel discussion would be the absolute low of the conference: little did we suspect what the future still held in store for us.

From 20.00 to 24.00 a Dutch evening with plenty to eat and to drink had been announced. What had not been announced was that conversation would be absolutely impossible thanks to constant "music" produced by two alternating groups. Thanks to uncontrolled electronic amplification they produced a deafening noise that was physically painful. I had been stupid enough to try to talk: the next morning I had a sore throat; Horst Huenke had the same experience. In short: if you want to torture and insult your guests, delegate the organization of your parties to "Martinair Party Service". I guarantee absolute success.

The invited speaker for the last morning —devoted to Architecture— was W.G.Spruth from IBM, Boeblingen, Germany, on "Trends in Computer System Structure and Architecture". But for a scathing remark about Honeywell not marketing MULTICS and another scathing remark about Burroughs and the ILLIAC IV the speaker only mentioned IBM products, mentioning their catalogue numbers at a higher rate than I could factorize them. It was a bloody shame. Van Wijngaarden, who was chairman of that session, took the precaution of not allowing any discussion and announced the coffee break. But was this a wise decision? During the coffee break several youngsters came to me, seriously worried by the fact that that shameful show had been allowed. They must have left the ECI-conference with in their mouths the bitter taste of dishonesty. Is this the way to educate our next generation?

It ended with a closing speech by van Wijngaarden in his capacity of Conference Chairman. Apart from thanking all the people who had contributed —he did so very nicely— he somewhat repaired the situation by commenting on Spruth's performance, be it in veiled terms. A large portion of his closing speech was devoted —in less veiled terms, but admirably chosen— to a public rebuke to me for my lack of tolerance. I had not only been annoyed by the music —and had shown so— but also by the fact that at the occasion of this conference a well-publicized ECI Computer Chess Tournament had been arranged. I did not like that at all —and had shown so— because I am of the considered opinion that, contrary to public superstition, the game of chess is of no relevance for computing science. By organizing the tournament, the ECI had only added fuel to that public misconception. It is perhaps easier to be tolerant, as soon as one doesn't care anymore.....

*         *         *

I learned a few things about Databases. I learned —or: had my tentative impression confirmed— that the term "Database Technology", although sometimes used, is immature, for there is hardly any underlying "science" that could justify the use of the term "technology". I have even my doubts when I am asked to believe that "database technology is still in its infancy", for that strikes me as being asked to regard the quacks at the fairs as the infancy of medical science. The point is that the way in which the database management experts tackle the problems seems to be so grossly inadequate. They seem to form an inbred crowd with very little knowledge of computing science in general, who tackle their problems primarily politically instead of scientifically. (In this respect the panel discussion was very revealing: at least half of the time was devoted to problems related to standardization! From the history of programming language development they should have learned to what disaster that premature concern about standardization may lead.) Often they seemed to be mentally trapped by the intricacies of early, rather ad hoc solutions to rather accidental problems; as soon as such a technique has received a name, it becomes "a database concept". And a totally inadequate use of language, sharpening their pencils with a blunt axe.

Lousy use of language —and therefore confusing— was a fairly general phenomenon. Allow me to end with the following anthology of crazy expressions. (Most of them are meaningless; if they mean something, it is something nonsensical.)

"virtual systems"
"virtual terminals"
"logical names"
"physical names"
"logical abstractions"
"mapping one level of abstraction onto the layer below"
"data structures are mapped into several layers of abstraction"
"a programmer efficiency index"
"an effective implementation view of the corporate data model"
"different levels of abstraction of view of data"
"dynamic change"

and, to crown the confusion,

"the computer playing this game."

No, gentlemen, three times No: computers don't play.


Plataanstraat 5
The Netherlands
prof.dr.Edsger W.Dijkstra
Burroughs Research Fellow

P.S. To give you some impression of how "international" this conference was: more than 150 participants were Dutch.

P.P.S. A final quotation from our IBM-spokesman:

"Interfaces decrease performance and increase manufacturing cost."