Open source software is often developed with a different management style from the industrial ones, therefore the quality and reliability of the code need to be studied. The characteristics of the source code of these projects need to be measured to obtain more information about it. Various kinds of code measurements can be quite helpful in obtaining information about the quality and fault-proneness of the code.

Within the project, we analyzed the structural architecture and source code quality of the complex open source program package called OpenOffice.org (which consists of more than 5 million lines of source code). The applied methods involved the analysis of the source code, extraction of various facts about it, and the determination of the positions in the source which are problematic regarding its quality.

The methods developed by experts were refined by utilizing machine learning algorithms. According to the results, we made suggestions for improvements. After the professional revision of these suggestions the developers of EuroOffice/MagyarOffice (based on OpenOffice.org) modified the critical parts of the source code.

The project resulted in a source code with improved quality that can be extended and maintained more safely. Furthermore, the developed software tools offer the possibility to continuously monitor the quality of the program package.

Web pages:

• OOoMo, the OpenOffice.org Monitor
• Official home page

Partners:

• Multiráció Kft.
• FrontEndART Kft.

• Printer-friendly version