In 1960, Carl attended a summer school in Manchester on modern methods of X-ray crystallography organized by D.W.J. Cruickshank. At the time, determining the structure of even a simple small molecule required intense computing power. Structure factors were calculated using mechanical calculators to add precomputed sine and cosine functions listed on so-called Beevers-Lipson strips. Most people then used a homemade analog computer for Fourier summations, the Hägg-Laurent machine, where three-dimensional Fourier sums were broken down into three successive one-dimensional summations. In a series of lectures Cruickshank described how the method of least squares could be applied to the refinement of X-ray structures and also outlined his program for one of the first digital computers in the United Kingdom.
Inspired by these lectures, Carl developed a collaboration with Stig Åsbrink of Stockholm University to write a similar program for the first Swedish electronic computer, BESK. This computer had a memory of 1,000 words, equipped with a larger magnetic drum memory, and used paper tape for input and output. There was no compiler for a higher language so the whole program was written in machine code. It took six months for Brändén to write a detailed flow chart, four months for Åsbrink to write the machine code and one year for them to debug the program. Over the next ten years this program was used by the entire Scandinavian crystallographic community and was top-rated in the use of computer time for both BESK and its much improved successor FACIT.
