BLIND ALLEYS IN UNIVERSITY COMPUTING (1983)


Translated from: Datalogiens veje og vildveje, Weekendavisen, København, 1983 Dec. 9.


Professor Dines Bjørner, chairman of The Informatics Committee of the Ministry of Education, writes in Weekendavisen on 1983 November 18 about an academic field or a group of such fields--the matter is not entirely clear--called informatics. It is argued that Copenhagen University ought to establish a faculty of informatics, in which the Institute of Datalogy, DIKU, and certain other fields should be placed.

As the background of the following comments on Bjørner's article I have to start by admitting a feeling of personal insufficiency. Unlike the members of The Informatics Committee, I am not among 'the persons who for more than 20 years, daily, and profoundly, in an intellectually honest manner have pondered over the didactics, affinities, development, and such like, of their field'. As a matter of fact I have only been concerned with datalogy for 17 years, more precisely since 1966 March 8, the day on which the designation datalogy occurred to me. And I have been a professor of this subject only since 1969.

Likewise in my ability to grasp what Bjørner calls a profound understanding of the major subject informatics I have to admit my shortcomings. Thus when Bjørner writes that 'The novel aspect in informatics is by its nature an intellectual universe of quality, not a material universe of quantity', then I have to say as my old physics teacher did when he was given certain types of answer from the students: This must be wisdom, because I do not understand it.

Again it may be seen as my shortcoming that I have to take exception to some of the views on datalogy that Bjørner propounds. Thus he writes: 'The frontier of datalogy will be engaged on developing knowledge processing models capable of simulating mental processes at ever higher levels... in this manner datalogy is about to establish itself as a "meta-discipline", taking upon itself to dissect the notations and methods of each individual subject and discipline'. If this opinion is the result of 20 years' daily, profound, and intellectually honest activity, then I have to admit that I have no idea what these forms of activity amount to.

There is, however, another of Bjørner's views of datalogy that I believe I understand. But this view, in my opinion, is highly dangerous for the potential of the subject to develop into a comprehensive and socially valuable activity. The issue is Bjørner's distinction between two aspects of the subject, viz. on the one hand 'a series of angles of attack, points of view, methods' and on the other 'applications of electronic data processing', and his disparaging characterization of the latter as being 'roughly speaking, entirely uninteresting'. With a view to the following discussion I shall denote the two aspects of the subject: pure datalogy and applied datalogy, in the sense that pure datalogy is the study of the conceivable potentialities of data processing, while applied datalogy is concerned with data processing in actual application to many other fields and activities, with attention to the properties and limitations of the current technology, in particular electronic data processing.

The danger of Bjørner's disparagement of applied datalogy, as I see it, is that the corresponding attitude is practically certain to make datalogy into a remote, self-sufficient, academic activity. This danger is obvious by a comparison with the state in which mathematics finds itself. Mathematics over the last century and a half has split into pure mathematics and applied mathematics. Pure mathematics nowadays is the dominant branch at most Western universities, also in Denmark. It is governed by self-sufficient ideals of mathematical beauty, generality, and elegance, and is the prime example of an academic activity in an ivory tower. It is virtually unassailable, partly because the university system allows it to persist by continued inbreeding, partly because pure mathematics is strongly attractive to a certain type of persons. In pure mathematics perfect clarity reigns, pronouncements are made in the form of theorems that may be proved true with absolute certainty. There is no room for doubt or personal evaluation that might create a feeling of uncertainty. The relation of pure mathematics to applied mathematics is like that of a game of chess to conflicts in the real world between kings, bishops, knights, and other such holders of power. In the game of chess everything proceeds according to unbreakable rules, there can be no doubt who has won a game. In the real world, if anything shall be decided one has to adopt a point of view, choose a basis of evaluation, and this basis can always be challenged. In applied mathematics one tries to understand some circumstance of the world using mathematical notions. Here again one has to choose what is to count as important, and it will rarely be clear whether the understanding achieved can be taken to be definitive.

The danger that pure datalogy will become dominant at the universities is entirely real. As a matter of fact, there is already a clear tendency in that direction at the Western universities. The Institute of Datalogy at Copenhagen University, DIKU, is unusual in so far as from its start in 1969 it has been strongly oriented towards applications and contact with other academic fields. This has not made matters easy. One will continually face difficult concerns whether one's work is scientifically defensible, whether it is solid enough. Such nagging doubts are unknown when pursuing pure studies. Here one decides on one's own evaluation basis, independently of the demands from the complicated, dirty reality. So far at DIKU we have been able to maintain the application oriented line and have been able to attract many students, who have been able to use what we try to teach them. When Bjørner claims that 'the young researchers... are frustrated time and time again by existing restrictive and illogical faculty boundaries' we at DIKU are puzzled. Both staff members and students have professional contacts in many directions, quite uninfluenced by faculty boundaries. At DIKU this is no problem.

But DIKU does have a problem, and one that has become unsolvable, to wit the imbalance between the teaching we have to provide for our many students and the resources of teachers, administrative assistants, premises, and equipment, we can draw on. Compared with those academic departments whose form of teaching is similar to ours, we have ten times as many students per teacher. At the present we have sufficient numbers of half finished students for the production of finished candidates for the next seventy years. We are just being strangled by students, with destructive consequences for our research and professional development, including the contact with other fields. This imbalance is neither unknown nor incomprehensible. It is displayed clearly in the reports on the state of the Institute of Datalogy that have been made over the last several years.

The problem is to some extent related to the circumstance that DIKU since 1976 has been forced into an organizational, and thus a resource, community with a series of departments in an entirely different situation, characterized by stagnation and cut-backs. But ultimately the problem is posed by the Ministry of Education. It can be solved only by, first, limiting the inflow of students. This is not done, one of the reasons being that Bjørner's Informatics Committee says that it shouldn't. Second, by increasing the resources. This will now result in a doubling of the teaching staff of DIKU over the coming years. This is scant help when a tenfold increase is needed. As compensation the production of candidates is expected to grow by a factor of three, and the research activity to increase. The stranglehold around our neck is being tightened still further.

With DIKU in this state of impotence it is to be expected that many other departments of Copenhagen University will feel urged to develop the applications of electronic data processing on their own. I agree with Bjørner that such an arrangement is in strong danger of making applications methodically shallow, lacking in deeper perspective. Unfortunately there is but slight prospect that this can be avoided. Rather there will be a strong pressure from many departments of the University to compensate for cuts in other activities by grants in electronic data processing, which appear to flow less stingily for the time being. Or such departments may jump on Bjørner's informatics bandwagon so as to attract the attention of the Minister of Education.

DIKU's problem cannot be solved within a foreseeable future, the opportunities were missed several years ago. We will have to continue to toil along with the hundreds of students, to cut away useful teaching activities, and to argue against colleagues in neighbouring departments who watch vigilantly for weaknesses in our endeavors, so as to justify once more that our wishes for a larger part of the insufficient resources are thwarted. Our situation will not be improved by talk in committees of the ministry of education. Such committees serve only to waste our time and effort, thus for example in producing the present reply to Bjørner's profundities.>