April 2018 – Nick Higham

By Len Freeman and Nick Higham

Joan Eileen Walsh was born on 7 October 1932 and passed away on 30 December 2017 at the age of 85.

Joan obtained a First Class B.A. honours degree in Mathematics from the University of Oxford in 1954. She then spent three years working as an Assistant Mistress at Howell’s School in Denbigh, North Wales. In 1957 Joan left teaching and enrolled at the University of Cambridge to study for a Diploma in Numerical Analysis. This qualification was awarded, with Distinction, in 1958. At this point, Joan returned to the University of Oxford Computing Laboratory to study for a D.Phil. under the supervision of Professor Leslie Fox. She was Fox’s first doctoral student. Her D.Phil. was awarded in 1961.

Between October 1960 and March 1963, Joan worked as a Mathematical Programmer for the CEGB (Central Electricity Generating Board) Computing Department in London. In April 1963, she was appointed to a Lectureship in the Department of Mathematics at the University of Manchester. She progressed through the positions of Senior Lecturer (1966) and Reader (1971) before being appointed as Professor of Numerical Analysis at the University of Manchester in October 1974. For the academic year 1967-1968 Joan had leave of absence at the SRC Atlas Computer Laboratory—a joint appointment with St Hilda’s College, Oxford.

Joan led the Numerical Analysis group at the University of Manchester until 1985, after which Christopher Baker took over. This was a period of expansion both for the Numerical Analysis group at Manchester and, more generally, for numerical analysis in Britain. This expansion of British numerical analysis was supported by special grants from the SRC (Science Research Council) to provide additional funding for the subject at the Universities of Dundee, Manchester and Oxford, from 1973 until 1976. This funding supported one Senior Research Fellow and two Research Fellows at each Institution. Joan helped establish the Manchester group as one of the leading Numerical Analysis research centres in the United Kingdom (with eight permanent staff by 1987)—a position that is maintained to the present day.

Joan was Head of the Department of Mathematics between 1986 and 1989, and subsequently became Pro-Vice Chancellor of the University of Manchester in 1990. She held the latter role for four years, and was responsible for undergraduate affairs across the University. Joan’s tenure as Pro-Vice Chancellor coincided with substantial, and sometimes controversial, changes in undergraduate teaching—for example, the introduction of semesterisation and of credit-based degree programmes; Joan managed these major changes across the University with her customary tact, energy and determination. Joan was an efficient and effective administrator at a time when relatively few women occupied senior management roles in universities.

After 35 years’ service, Joan retired from the University in 1998 and was appointed Professor Emeritus.

In retirement, Joan returned to her studies; between 2000 and 2003 she studied for an MA in “Contemporary Theology in the Catholic Tradition” at Heythrop College of the University of London.

Over the years, and particularly during her tenure as Pro-Vice Chancellor, Joan sat on, and chaired, numerous University committees, far too many to list. She had a very long relationship with Allen Hall (a University Hall of Residence) where she was on the Hall Advisory Committee from 1975 until her retirement in 1998.

Joan served leadership roles nationally, as well as in the University. She was Vice President of the IMA (1992-1993) and a member of the Council of the IMA (1990-1991 and 1994-1995). She was elected Fellow of the Institute of Mathematics and its Applications (IMA) in 1984. She was a member of the Computer Board for Universities and Research Councils for several years from the late 1970s. She encouraged the creation of its Software Provision Committee, formally constituted in 1980 with Joan as its first Chairman, which she led until 1985. She was also President of the National Conference of University Professors (1993–1994). Further, she was a member of the Board of Governors at Withington Girls’ School, a leading independent school, for six years between 1993 and 1999.

Nowadays, all computational scientists take for granted the existence of software libraries such as the NAG Library. It is unimaginable to undertake major computational tasks without them. In 1970, Joan was one of a group of four academics who founded the Nottingham Algorithms Group with the aim of developing a comprehensive mathematical software library for use by the group of universities that were running ICL 1906A mainframe computers. Subsequently, the Nottingham Algorithms Group moved from the University of Nottingham to the University of Oxford and the project was incorporated as the Numerical Algorithms Group (NAG) Ltd. Joan became the Founding Chairman of NAG Ltd. in 1976, a position she held for the next ten years. She was subsequently a member of the Council of NAG Ltd. from 1992 until 1996. In recognition of her contribution to the NAG project Joan was elected as a Founding Member of the NAG Life Service Recognition Award in 2011.

Joan’s research interests focused on the numerical solution of ordinary differential equation boundary value problems and the numerical solution of partial differential equations. She conducted much of her research in collaboration with PhD students, supervising the following PhD students at the University of Manchester, who obtained their degrees in the years shown:

Thomas Sag, 1966;
Les Graney, 1973;
David Sayers, 1973;
Geoffrey McKeown, 1977;
Roderick Cook, 1978;
Patricia Hanson, 1979;
Guy Lonsdale, 1985;
Chan Basaruddin, 1990;
Fathalla Rihan (supervised jointly with C. T. H. Baker), 2000.

Joan was an important figure in the development of Numerical Analysis and Scientific Computing at the University of Manchester and in the UK more generally. Her essay Numerical Analysis at the Victoria University of Manchester, 1957-1979 gives an interesting perspective on early developments at Manchester.

Brian Ford OBE, Founder Director of NAG, writes:

Joan had a brilliant career in Mathematics (particularly areas of Numerical Mathematics, Ordinary and Partial Differential Equations), Computing, University Education and Teaching, and was an excellent researcher, teacher, administrator, doctoral supervisor and colleague. But she was so much more than that!

Joan was invariably kind and thoughtful, intellectually gifted and generous with advice and guidance. Her profound Christian faith illuminated every aspect of her life. Joan’s deep reading and wide intellectual interests coupled with her prudence and clear thinking gave her profound knowledge and command. She was excellent company –amusing, modest, never belittling nor intimidating- and enjoyed fine wine and food in good company. She held firm beliefs, gently and persuasively seeking what she saw as the right way. Many people turned to her for help, advice and references and were grateful for her readily-offered help and support.

Joan was a private, even guarded, person. A devout Catholic, on her retirement she completed an MA in “Contemporary Theology in the Catholic Tradition” at Heythorp College, University of London. Fluent in Latin and reading regularly at services, she loved the traditional Tridentine Mass of the Church. Along with her local bishop in Salford and other like-minded Catholics, she therefore worked actively for the restitution of the Tridentine Mass to the liturgy of the world-wide Church (sidelined after Vatican II in favour of local languages), an involvement which culminated her joining high-level discussions at the Vatican. This bore fruit, the Tridentine Latin Mass being officially declared the extraordinary form of the Roman Rite of Mass a few years later: Joan was thrilled. Such was Joan’s commitment to things she believed in and her endless thought and work for others.

Joan was an excellent contributor to the NAG Library, believing strongly in collaboration and sharing, with high quality standards for all aspects of our work and thorough checking and testing. She was an excellent first Chairman of NAG and invaluable colleague and advisor. We thoroughly enjoyed working together, invariably in an excellent spirit. We achieved much.

Last week I had the pleasure of attending and speaking at the Conference on Scientific Computing and Approximation (March 30-31, 2018) at Purdue University, held in honour of Walter Gautschi (Professor Emeritus of Computer Science and Mathematics at Purdue University) on the occasion of his 90th birthday.

The conference was expertly organized by Alex Pothen and Jie Shen. The attendees, numbering around 70, included many of Walter’s friends and colleagues.

The speakers made many references to Walter’s research contributions, particularly in the area of orthogonal polynomials. In my talk, Matrix Functions and their Sensitivity, I emphasized Walter’s work on conditioning of Vandermonde matrices.

A Vandermonde matrix $V_n$ is an $n\times n$ matrix depending on parameters $x_1,x_2,\ldots,x_n$ that has $j$ th column $[1, x_j, \ldots, x_j^{n-1}]^T$ . It is nonsingular when the $x_i$ are distinct. This is a notoriously ill conditioned class of matrices. Walter said that he first experienced the ill conditioning when he computed Gaussian quadrature formulas from moments of a weight function.

Walter has written numerous papers on Vandermonde matrices that give much insight into their conditioning. Here is a very a brief selection of Walter’s results. For more, see my chapter Numerical Conditioning in Walter’s collected works.

In a 1962 paper he showed that

$\displaystyle\|V_n^{-1}\|_{\infty} \le \max_i \prod_{j\ne i}\frac{ 1+|x_j| }{ |x_i-x_j| }.$

In 1978 he obtained

$\displaystyle\|V_n^{-1}\|_{\infty} \ge \max_i \prod_{j\ne i} \frac{ \max(1,|x_j|) }{ |x_i-x_j| },$

which differs from the upper bound by at most a factor $2^{n-1}$ . A 1975 result is that for $x_i$ equispaced on $[0,1]$ ,

$\displaystyle\kappa(V_n)_{\infty} \sim \frac{1}{\pi} e^{-\frac{\pi}{4}} (3.1)^n.$

A 1988 paper returns to lower bounds, showing that for $x_i \ge 0$ and $n\ge 2$ ,

$\displaystyle\kappa(V_n)_{\infty} > 2^{n-1}.$

When some of the $x_i$ coincide a confluent Vandermonde matrix can be defined, in which columns are “repeatedly differentiated”. Walter has obtained bounds for the confluent case, too.

These results quantify the extreme ill conditioning. I should note, though, that appropriate algorithms that exploit structure can nevertheless obtain accurate solutions to Vandermonde problems, as described in Chapter 22 of Accuracy and Stability of Numerical Algorithms.

Ron DeVore speaking on Optimal Data Assimilation

Photo of Nick Higham’s talk by David Gleich.

Month: April 2018

Joan E. Walsh (1932–2017)

Conference in Honour of Walter Gautschi

Share this:

Share this: