I am devastated to be told today that my colleague Nick Higham has died. Nick was an outstanding research leader in numerical analysis, but he was also a deeply kind man and feel I need to say something about the limited time I spent with him.

I have always worked on the boundary of electronic engineering and computer science, often working on efficient hardware implementations of numerical algorithms. While as a research student I was aware of the early work of Jim Wilkinson, it was only after my PhD was complete in 2001/2 that I became aware of Nick’s work via his outstanding book Accuracy and Stability of Numerical Algorithms. There are only a few books I have purposely bought multiple copies of over the years, and this is one. I have at least three – one for home, one for work, one for my PhD students. It happens to be sitting on my desk at work next to my laptop as a write this. As does an early Muldivo mechanical calculator – another shared joy we bonded over: Nick’s office featured a Brunsviga model last time I was there.

I can’t remember how or when I first met Nick in person, but we hit it off very quickly. Despite being as eminent as they come, he took a deep interest in the work of a junior member of academic staff in a different discipline. He invited me to present in Manchester on a number of occasions, included me in SIAM conference mini-symposia he was organising, and involved me in a Royal Society event he co-organised. We had deep technical discussions, yes, but we also spoke about careers, about my role as a school governor (and about the school his son was then attending), about Jim Wilkinson’s work, and he encouraged me to think about fellowship of SIAM. One of my favourite conference experiences was attending SIAM CSE in Spokane, Washington, at Nick’s invitation, primarily because it was so different to the usual Electronics and Computing conferences I tend to attend.

Nick was a profoundly practical person as well as being deeply involved in theoretical analysis. After I told him about some experimental observations a PhD student and I had made some years earlier regarding finite precision effects in certain Krylov methods, he was surprised and a few days later had coded up an example, verified this behaviour, and written a SIAM News column about it! We connected over the interrelation of theory and practice, and he told me that he very much liked the suggestion I made that the rise of custom hardware and FPGA-based compute meant that time had come for computer science, numerical analysis, and electronic engineering to come back together, after a long period of not talking as much as disciplines as I would like. He was also amused by my interest in “very pure” (his words) real algebraic geometry.

Nick taught me – whether he knew it or not – how to communicate to mathematicians, as well as how to use Latex for slides… even down to emailing me suggestions for Latex code hacks to make my slides better. His presentations were always a model of clarity and depth.

Many colleagues in the Numerical Analysis and Linear Algebra world will know Nick much better than I did, but he touched my research life more deeply than most. Somewhere in the back of my mind, I had the idea that I’d like to take a sabbatical leave with Nick in the future. I thought there would be plenty of time.