Software testing has been the subject of active research for approaching 30
years. In that time, there have been developments such as the invention of
techniques and tools, the morphosis of these techniques and tools to deal
with the adoption of new development paradigms and programming languages, a
nd analytical and empirical comparisons of techniques to improve our unders
tanding of their relative merits. Impressive though these developments are,
there still remains the fundamental problem of trying to relate the result
s of testing to some objective attribute of the program (such as number of
faults). It is the search for this missing link that is attracting much of
the recent research in the area. Any progress made towards this will have a
considerable impact on our understanding of the subject and ultimately on
software itself. (C) 1999 Elsevier Science B.V. All rights reserved.