Quantitative Security Characteristics of Perl Programs

Abstract

The program quality is used to be characterized with error number per 1000 code lines. This parameter is calculated by a statistical regressive analysis of error numbers for successive code versions, with a subsequent extrapolation for the future. This procedure is very tedious even for large companies. It is very hard to verify this estimate for common users, as they have no initial data. There are a lot of methods to estimate code error number, e.g., models Shooman, Musa, Bell-LaPadula, Jelinski-Moranda, Schick-Wolverton, Mills, Lipov, Corcoran, Bernoulli simple intuitive software reliability model, Nelson's software reliability. But often we deal with programs that formally have no errors, at the same time their quality is not evident. The method is proposed to estimate quantitatively a code quality for Perl-routines. This method can identify weaknesses in certain program components, where errors are possible. The proposed method is based on programming style analysis. The method is applicable for any programs with open sources (Python, Perl, PHP, etc). The method can be used for quality comparison and choice of the programs solving similar tasks.

Keywords

• security
• testing
• structured programming principles