b'Environmental geophysics The early days of computing the screen, after which you made a guessWeidelts visit he, Kurt Srensen, and of an initial model and typed it in. AfterHans Kurt Johansen worked together I attended my first programming coursea while, both data and response wereon a theory for the calculation of at the Department of Computer Science,shown on the screen, and you repeatedHankel transforms which come which in the late 1970s was a part of thethe process modifying the model untilup in almost all the mathematical Mathematics Department. Exercises werea satisfying data fit was obtained. Thatexpressions for the responses of EM done by carrying our small programs tomight take a while. It would take a fewmethods. About half a year after I a special terminal that produced punchminutes to calculate a geoelectricalstarted my Masters studies at the cards, one for each program line. Whenmodel response and you might have todepartment in 1975, I got to share finished with the punching you took thego through several trials. Today, a fullyan office with Kurt Srensen, an stack of cards out of the machine beingautomatic inversion of a geoelectricalarrangement that continued for the very careful not to mix them up. Woe besounding data set takes less than 1 ms! next eleven years. Fairly soon, I was to the one who dropped the stack onpresented with the FHT theory, and I the floor and had to put them in orderstarted coding a program that would afterwards. We took the stash of cardsTheoretical background calculate the filter coefficients used in to the counter and handed them overthe FHT. That was a bit of a mouthful. I to the machine operator who then putOne of the theoretical prerequisites to code subroutines for EM responses washad to re-study the theory of complex them in the queue and ran the programfunctions, but eventually a functioning on the university mainframe. The queuethe advanced course in EM methods that was given by Peter Weidelt, oneprogram came out of this work which - could be quite long, and half an hourwith a few additions and optimisations easily passed before your job was done,of the great theoretical geophysicists in this speciality. In 1974 he visited our- is still used. For most reasonably so often you just left to come back later,well-behaved kernel functions, our maybe the next day, to get your cardsdepartment for a period of six months, and his course lifted the theoreticalmethod is still the most robust, flexible back and a printout from the line printerand in most cases accurate method which often laconically informed you thatfoundation several levels. After Peter left the department, his course wasfor calculating FHTs (and Fourier there was a syntax error in line 52. Then ittransforms for forward responses), was all over again: find the error, punch arepeated every year by Laust Brsting Pedersen using Peters handwrittenand for any desired accuracy, the filter new card, put it in the correct sequencecoefficients can be chosen to minimise in the stack, and hand it to the operator. notes that he created concurrent with his lectures in 1974; I took the course incalculation time. The program was Everyone who has ever written computer1975. Peter Weidelts efforts and Laustsmade open source, a word that did code knows that this can happen afollow-up laid down some of the solidnot exist at the time, and it has been considerable number of times, so itfoundation stones for the internationallysent to hundreds of people all over the could take days to correct the errors inrenowned level of hydrogeophysics inworld. We became world champions in a 100-line program and many bike ridesAarhus today. Hankel transforms! With Peter Weidelts back and forth between the computeradvanced EM course and the best desk and the office. Eventually cards wereAnother of the cornerstones waspractice for doing FHTs, we had the replaced with punch tapes, and it wasthe theory and practice around Fastbest possible starting point for making a great relief that you did not have toHankel Transforms (FHT). During Petergood EM response routines.fumble with the cards.In the late 1970s, our department acquired a graphical terminal with the fantastic property that it could draw lines and graphs on the phosphorescent green screen (Figure 2, left). Before that, only TTY terminals were available that could receive and send characters at a maximum rate of 110 characters/minute, so the ones who really had the knack learned to type with a clockwork speed of 110 characters/minute. I can still perfectly recall the sound of those electromechanical terminals: tschak-a-tschak-a-tschak-a-tschak.Along the way we got more screen terminals which, however, were only able to show green characters on the screen. With the graphical screen, it became possible to make interactive inversion programs of, for example, geoelectrical soundings like the ones Hans Kurt Johansen was producing at the time.Figure 2.The (with present day eyes) clunky devices from the middle ages of computing. Left: The Tektronix The interactive, trial-and-error process4010 graphical screen (https://w140.com/tekwiki/index.php?curid = 24556). Right: The Nord 100 main frame consisted of having the data shown on(Norsk Data Historie).JUNE 2022 PREVIEW 34'