Abstract:
INTRODUCTION
I am very honoured to have been invited to deliver the 10th Vidya Jyothi
Professor V. K. Samaranayake memorial oration. I thank the Director and the
Academic staff of the U.C.S.C., and the University of Colombo for this. I
believe I am the first non-Sri Lankan to have been invited in this way. I
understand this is not only the 10th anniversary of this memorial series, but
is also being celebrated as the 50th anniversary of computing at the
University of Colombo.
Professor Samaranayake (Prof, from now on in this talk), influenced
developments in Sri Lanka in many areas. I am a statistician and therefore I
use this occasion mainly to discuss his influence in this area. In particular I
consider some topics that combine statistics and computing. This influence
between our organisations was partly via a link between the University of
Colombo and the University of Reading. The formal link was more than 30
years ago, but some of the activities remain important today.
In 1967,50 years ago, Prof first taught FORTRAN programming to University
of Colombo students. This was the same year that I joined the Department
of Applied Statistics at the University of Reading. Statistics has been a
problem subject to many people, in many countries, for a long time. I
describe this problem in Box 1 below.
Box 1: Common problems with statistics teaching
Service course training is often dominated by analysis, with relatively little
on data organisation, or on design. Examples are usually relatively small,
hence little time is devoted to important descriptive methods.
Sometimes a "recipe-book" approach is used which usually results in
topics covered in their order of mathematical complexity, rather than
their importance. It also often results in the overuse, and often irrelevant
use, of significance tests. This approach provides little understanding of
principles
