The objective of the session is to agree an agenda, for both the industrial and research communities, for improving the use of measurement to control software production.
Tom DeMarco has suggested that "you can't control what you
can't measure". In object oriented systems it isn't always
that easy to know what to measure but object oriented projects
still have to be controlled. The purpose of this working group
then is to examine the relationship between the object oriented
development process and software measurement. When we look at
this relationship there seems to be a gap between research and
current practice, a gap we would like to try to close.
It is all too common for research in this area to consider metrics
in isolation from the process within which measurement is made.
From an industrial view point measurement often appears as an
academic topic unrelated to the problems and requirements of software
production in the real world. Metrics such as those from Chidamber
& Kemerer have been proposed in theory and used in practice,
but never in a really convincing fashion.
If software production is to truly achieve the engineering status
for which it strives software production must be put under quantitative
process control. This requires that academia and industry collaborate
and agree a common agenda for producing sound measurement research
and effectively transferring this research into industrial practice.
The scope of the session includes: the mechanics of measuring software products and processes; the integration of the measurement process with the development process; and the use of measurement results in controlling projects.
Back to conference programme
Back to OT98 main page