Department of Computer Science, Lund University

The early history of the journal BIT is closely related to the development of the first electronic computers. The need for large scale computation arose during the war, particularly in connection with the atomic bomb. An early construction is rather well known, namely the ENIAC (Electronic Numerical Integrator and Computer), completed in 1945. It could store just 20 numbers, but its operational speed was quite impressive. A newspaper suggested the following characterization: ``It has the size of a destroyer and at work it roars like an admiral!'' However, more modern computers were waiting just around the corner, and there soon arose a need to describe and implement numerical processes in a better and more compact way than before.

The basic ideas behind such constructions were certainly important, and several more sophisticated projects were started along these lines. In the late forties and early fifties excellent machinery became available in several places. The Stockholm computer BESK, designed by Erik Stemme, was completed in November 1953 and became operative almost immediately. A simplified copy SMIL was built in Lund, first following a much simpler design with a slow magnetic drum as memory, but after a few years a full-fledged member of the family. Other computers appeared and it soon turned out that programming now had become the black sheep. With this fact as a back-ground it is not surprising that there was a need for discussions of mathematical and technical problems, programming features, applications of different kinds, and so forth. Now it may seem somewhat surprising that the Swedish armed forces, particularly the navy were quite interested in all kinds of applications. Anyhow, The Swedish Board for Computing Machinery together with the Swedish Navy passed an invitation to all interested parties in the Nordic countries to attend a two days meeting May 14-15, 1959, in Karlskrona, a naval base on the Baltic in the southern part of Sweden. As a result 207 people from Sweden, 28 from Denmark, 15 from Finland and 16 from Norway participated, in all some 270 persons. The weather conditions were excellent and the guests soon understood that Karlskrona was a city with a most interesting history.

The lectures treated six different areas: general problems, data processing, programming, numerical analysis, operations analysis, and technical features. The problems in these groups in some cases were of a rather special nature, and they would nowadays be either reformulated or neglected. On the other hand, problems within some areas, e.g., numerical analysis and programming, are still recognized in this form, particularly since they have a well-defined structure, and it soon turned out that realistic methods for their solution can be constructed and analyzed.

The lectures at the conference were met with great interest and it was decided to print a report for each subject. All contributions were collected and published in a book with title ``Nordic symposium on the use of computers'', (mimeographed, Karlskrona and Lund; 445 pages). In the following discussions there were suggestions to find suitable procedures to publish similar papers also in the future, but that is as far we got by this time. However, one important decision was taken, probably a few month later, when specialists in these subjects were invited to a second meeting. It was held August 26-31, 1960 in Copenhagen with Niels Ivar Bech as host. An important reason for his engagement was the fact that better than any one else he understood the importance of cooperation between the Nordic countries in this respect. The common Nordic symposia (NordSAM, later NordDATA) became an important forum for such an exchange of ideas and information. However, it was generally felt that there was something missing in this picture, and alternative methods to treat problems belonging to different classes appeared as highly desirable. Many scholars could clearly see the need for a common Nordic journal within this area.

In connection with the symposium in Copenhagen 1960 I had an opportunity to discuss these matters with Niels Ivar Bech (who was in charge of the Danish computer institute), particularly on the question whether there were practical possibilities to start such a scientific journal. Almost immediately Bech agreed with the general idea and very quickly he was able to arrange an informal Nordic meeting where the plans were discussed in more detail. Probably, also Peter Naur from Denmark, Jan Garwick from Norway, and Olli Lokki from Finland participated. It was not too difficult to decide what subjects we wanted to cover in the journal: computer technology, numerical analysis, operations analysis, programming techniques, and data handling. During the fall of 1960 we worked hard on a number of practical problems: selection of editorial staff, economic planning, referee problems, checking and improving the manuscripts, printing, proof-reading, reprint handling, and subscriptions, to mention just the most important ones.

By this time the name of the journal became crucial, and first there was quite some sympathy for NORDSAM. However, criticism arose from Norway due to the possibility of mixing the names of our symposia and their documentation on one hand and the journal on the other. ``Nordisk Tidskrift for InformationsBehandling'' would be an acceptable but unwieldy long name. However, after a suitable permutation of the letters T, I, and B we obtained the result BIT which was finally accepted.

An essential problem for the practical implementation of course concerned economy; starting a new journal without economical back-ground is indeed impossible. A first step consisted in establishing contacts with suitable firms which were able to produce four issues per year of 48 pages each. However, it soon turned out that a yearly output of 200 pages would not by far be sufficient to accommodate all papers already waiting to be printed. On the other hand we would need at least 1000 subscriptions to cover this fairly reasonable volume. However, additional space would be expensive and more money must be supplied from external sources.

A first step towards the solution of this problem was taken in a letter to Bech in November 1960. At a following meeting in Copenhagen we discussed the difficulties and a solution was found. Bech could offer a yearly support of 3,000 crowns which by that time was an impressive sum. With this back-ground it became possible to secure a similar support from a Swedish official source, namely Matematikmaskinnämnden (Swedish Board for Computing Machinery). To this we could add the income from subscriptions which, however, was of a somewhat smaller order. (The subscription price for the first volume of four issues was set to 12 Swedish Crowns or US $ 3.) The following is a quote from the first issue:

BIT is a new quarterly journal in the computer field supported by the Scandinavian countries including Finland and Iceland. Formal editors are the societies for information processing in these four countries.

BIT will contain scientific articles on numerical analysis, computer-technical problems, operational research, programming and data-
processing, including office-automation. As a rule, these articles will be written in English, German or French. To some extent BIT will also publish informative articles on important new subjects; such articles will in general be written in one of the Scandinavian languages with an English summary appended. It is also intended to publish algorithms written in ALGOL.

Scientific papers of the kind mentioned above are welcomed also from foreign countries. Such papers should be sent to the editor for further submission to other members of the editorial board. As a rule the author will get 50 reprints free of charge.

The first issue of BIT was distributed in February 1961, and it could be regarded as a test product. The new journal was noticed internationally rather quickly; it was mentioned already in the April-May issue 1961 of Scientific Information Notes.

In connection with the third and fourth NordSAM meetings in Oslo, August 18-22, 1961, and in Helsinki, August 16-20, 1963, meetings were arranged for the Editors of BIT. The economic problems dominated, but in addition we discussed editorial work, e.g. problems with reviews and advertisements. The following persons participated in Oslo: Nils Ivar Bech, Bergsson (Island observer), Tage Carlsson, Stig Comét (Sweden, observer), Olle Dopping, Carl-Erik Fröberg, Jan Garwick, C. L. Godske, Olli Lokki, and Bent Scharøe-Petersen.

Concerning the contents of the journal we were eager to publish contributions also from authors outside the Nordic countries. Already the first issue in 1961 contained a contribution from a non-Nordic author, and this trend towards a more universal state prevailed. As an example we mention that the 1963 issues contained contributions from Gene Golub, Edgar Karst, Mok-Kong Shen, and Peter Wynn. We think that a list of the papers published in the first issue of 1961 could be of interest; some of them are purely scientific while others have the character of review papers.

           S. Bragnum, Varför tunneldioder
G. Ehrling, On the numerical computation of incomplete elliptic integrals.

C.-E. Fröberg, On the sum of inverses of primes and of twin primes.

J. V. Garwick, The programming of large logical systems.

B. Jansson, Study of anti-aircraft systems by simulation with a Monte-Carlo   model.

J. Jensen and P. Naur, An implementation of Algol 60 procedures.

R. Kivivuori, A method for checking numerical codes using the 1401.

B. Langefors, Information retrieval in file processing.

The years 1965-1985, were characterized by a consolidation process, particularly associated with improved contacts with foreign countries. This process also included search for improved facilities concerning the treatment of manuscripts, refereeing, printing, and distribution. Then the role of a splendid administrative support from Erik Bruhn, Copenhagen, must be mentioned.

For many years BIT published scientific results within computer science and numerical mathematics side by side. From 1980 different sections corresponding to this division were established. A decision to split BIT into two different journals was taken in October 1992. The name BIT was kept for the numerical journal, while the other, called Nordic Journal of Computing, started with a mostly new Editorial Board headed by Esko Ukkonen, with the administration located to Helsinki, Finland. To emphasize the split the name BIT was changed to BIT Numerical Mathematics from 1995.

During the years 1963-1968 BIT received economical support from the Swedish and Norwegian Natural Science Research Councils and during 1969-1983 from the joint Nordic Publication Board for Natural Sciences. From 1984 it became possible to make BIT independent of external economic support, although BIT has remained being published on a non-profit basis.

Two persons have served on the Editorial Board (almost) from the beginning and during 30 years or more, namely Peter Naur, 1960-1993, and Germund Dahlquist 1962-1991. Peter Naur was Area Editor for Computer Science 1980-1993. Åke Björck, who joined the Editorial Board as Area Editor for Numerical Mathematics in 1980, took over as Editor for BIT from 1993.

At this point we feel that it is reasonable to ask the question whether there were some papers which opened new areas with respect to theory and solution techniques. It would carry to far to give a thorough description, but we feel that one example ought to be given. Germund Dahlquist contributed several ground-breaking papers on the numerical solution of differential equations. This area is still one of primary importance in the latest volumes of BIT.