Projects of the Experimental
Software Engineering Group at the University of Maryland
|
OPT: Organization and Process Together
|
Approach
|
This project began with the development of a framework, based on the
Quality Improvement Paradigm, for systematically analyzing organizational
and process issues. This framework was designed and tested in an
exploratory case study at the IBM Software Solutions Toronto Laboratory in
1993. Then the project's focus narrowed to the exploration of the specific
characteristics of organization and process mentioned above. An empirical
study was designed to investigate the relationships between these factors,
a pilot study was conducted at IBM in Toronto in 1994 and 1995, and a
larger study was conducted at the
NASA SEL
in early 1996. Both were
studies of code inspections, and the amounts of effort expended in various
communication activities that are part of the inspection process. Both
qualitative and quantitative data collection and analysis techniques were
used, including participant observation, structured interviewing, simple
graphical display of data, and statistical tests of significance. |
Project Status
|
Inactive - Studies completed, but a followup study of
error data is being contemplated. |
Results
|
The results of both the pilot and larger study point to interesting and
non-intuitive relationships between organization and process factors. For
example, the participation of inspectors who are unfamiliar with each other
tends to increase communication effort, but also tends to uncover more and
varied issues pertaining to the code being inspected. Another result shows
that a group of inspectors, most of whom are from the same organizational
unit but which includes a few outsiders, will take longer to communicate
than a group of inspectors all of whom are from the organizational unit, or
all of whom are from different units. In short, the studies show that the
relationships between organization and process are complex, and require
further study of many kinds. The major contributions of these studies are,
in fact, a set of methods designed to study this issue, as well as a large
set of hypotheses with which to start the investigation.
|