Abstract
Software testing is the process that controls the quality of software (Myers, 1979). Software testing is comprised in any development process and every method of development applies practices for software testing (Burnstein, 2003; Kaner et al., 2002). Traditional methods of development, like the waterfall approach, allocate software testing in a given phase of the overall development process – e.g., toward the end of the software lifecycle. In modern methods, practices of software testing rather permeate the whole development process in an iterative and increasing way (Black, 2002; Spillner et al., 2007) – e.g., in XP.
Top7.2 Testing In The Open Source Development
The majority of OS projects combines the feedback from the community and an internal strategy of testing to release competitive and stable software products.
The most common practices of internal testing in the open source projects are the use of the nightly builds and the frequent releases. The frequent releases guarantees a fast and iterative reporting from all the community. The majority of the projects use Bugzilla or a modification of it to collect - from the internal teams or from the volunteers - reports on failures, request of modification, or enhancements. The same tool is used to divulgate solutions, patches, or occurrences of defects. Many of OS projects declare to use automated tools or agile practices for testing, like Test Driven Development (TDD). As we shall see, the reality is different: in many cases test classes are totally missing or appear in a very low percentage.In what follows we discuss the existence of test classes in projects stored in the following on-line repositories:
In Table 1 we report the number of classes (files), the programming language dominant in the project, the number of test classes (tests) and the percentage of test with respect to the files.