Information Technology Reference
In-Depth Information
A Quantitative Analysis into the Economics of
Correcting Software Bugs
Craig S. Wright and Tanveer A. Zia
School of Computing and Mathematics
Charles Sturt University, NSW 2678
cwrigh20@postoffice.csu.edu.au, tzia@csu.edu.au
Abstract. Using a quantitative study of in-house coding practices, we demon-
strate the notion that programming needs to move from "Lines of Code per day"
as a productivity measure to a measure that takes debugging and documentation
into account. This could be something such as "Lines of clean, simple, correct,
well-documented code per day", but with bugs propagating into the 6th iteration
of patches, a new paradigm needs to be developed. Finding flaws in software,
whether these have a security related cost or not, is an essential component of
software development. When these bugs result in security vulnerabilities, the im-
portance of testing becomes even more critical. Many studies have been con-
ducted using the practices of large software vendors as a basis, but few studies
have looked at in-house development practices. This paper uses an empirical
study of in-house software coding practices in Australian companies to both dem-
onstrate that there is an economic limit to how far testing should proceed as well
as noting the deficiencies in the existing approaches.
Keywords: Software Development Life Cycle, Model Checking, Software
Verification, Empirical studies.
1 Introduction
The deficiency of published quantitative data on software development and systems
design has been a major ground for software engineering's failure to ascertain a
proper scientific foundation. Past studies into coding practice have focused on soft-
ware vendors. These developers have many distinctions from in-house projects that
are not incorporated into the practices and do not align well with in-house corporate
code development. In the past, building software was the only option but as the in-
dustry developed, the build vs. buy argument has swung back towards in-house de-
velopment with the uptake of Internet connected systems. In general, this has been
targeted towards specialized web databases and online systems with office systems
and mainstream commercial applications becoming a 'buy' decision.
As companies move more and more to using the web and as 'cloud applications'
become accepted, in-house development is becoming more common. This paper uses
an empirical study of in-house software coding practices in Australian companies to
both demonstrate that there is an economic limit to how far testing should proceed as
well as noting the deficiencies in the existing approaches.
 
Search WWH ::




Custom Search