December 1999
Volume 2 • Number 1
Contents
A Worldwide Survey on Best
Practices Toward Software Engineering Process Excellence
A multipart survey was designed and conducted
to seek the practical foundation of base process activities
(BPAs) in the software industry and to support research in
modeling the software engineering processes. A superset of
BPAs compatible with current software process models, such
as ISO 15504, CMM, ISO 9000, and BOOTSTRAP, was identified
for developing the questionnaire.
This article reports the survey’s findings on a
comprehensive set of software engineering processes and
BPAs. Each BPA is benchmarked on attributes of mean importance
and ratios of significance, practice, and effectiveness.
By comparing the benchmarks with current organizational
practices, recommendations can be made as to which areas
should have processes established first and which areas
should be the highest priority for process improvement.
Key words: customer services, organizational processes,
organizational structure, software process assessment, software
process improvement, software process reference model
by Yingxu Wang and Alec Dorling Center for Software
Engineering, Graham King, Margaret Ross, Jeff Staples, and
Ian Court, Research Center for Software Engineering
INTRODUCTION
A software process system can be broken into the following:
practice, process, process category, process subsystem, and
process system. A practice is an activity or state in a software
process that carries out a specific task or set of coherent
tasks. A practice is the minimum unit that can be modeled
in a software process system. In this article, a practice
will be referred to as a base process activity (BPA). A software
process is a set of sequential BPAs that are functionally
coherent and reusable for software project organization, implementation,
and management.
A basic argument for existing process models is that their
BPA sets have been selected empirically, and the validation
of these BPAs on usefulness and effectiveness has not been
reported. In this article, for the first time, a superset
of 444 BPAs is quantitatively characterized and validated.
Based on this work, the BPAs and processes in current process
models are supported by a set of quantitative benchmarked
and characterized attributes.
This work is based on a worldwide survey of BPAs aimed at
supporting software process excellence (Wang et al. 1997a;
1997b; 1997c; 1997d). The multipart survey was conducted by
the Research Center for Systems Engineering at Southampton
Institute in the United Kingdom, in collaboration with the
IVF Center for Software Engineering, European Software Institute,
and BCS. The survey was carried out on the Internet via professional
lists, such as SPICE (SUGAR), WWSPIN, ISO 9000, and ISO 9000-3,
as well as at a number of international conferences on software
engineering and software processes, such as the ICSQ ’96
(Ottawa, Canada), SP ’96 (a congress incorporating ICSP
’96, SPICE ’96, SPI ’96, and ISCN ’96
in Brighton, U. K.), and SQM ’97 (Bath, U. K.).
Using a benchmarked software process model, an organization’s
software process practices can be evaluated quantitatively
and process improvement opportunities can be identified and
prioritized based on the significance and effectiveness of
the BPAs within the organization.
Figures A-1 through A-5
Figures A-6 through A-10
Figures A-11 through A-18
Figure A-19
|
