00 Metrics Data Gathering Environment

Author(s): Marjan Hericko, Ivan Rozman, Romana Horvat, Tomaz Domajnko, and Jozsef Gyorkos
Venue: Technology of Object-Oriented Languages, TOOLS 24. Proceedings
Date: 1997

Type of Experiement: Case Study



This paper introduces the OOMetDaGa Environment, a new environment that provides automatic support for metrics data gathering and provides some empirical data to support this new environment. OOMetDaGa, or OO Metrics Data Gathering Environment, is built based on successes and failures of previous metrics tools including OOMetTool, PROCESSUSTool, and PRO+. By combining these tools into one, the goal of OOMetDaGa is to provide both static analysis and dynamic analysis of a piece of software. In order to perform the dynamic aspects this tool must gather meta-data including results from audits, inspections and all other activities and measurements that take place in software companies.

The case study performed in this paper took 10 C++ projects, 3 Delphi projects, Smalltalk projects, 9 software process maturity assessments, and some other external data in order to get results on OO metrics. The conclusion the authors came to based on their empirical data was that lines of code and number of classes are very poor metrics deciding the complexity of the code. Depth of Inheritance tree was the other aspect looked at closely in the empirical study, showing that the biggest difference in the DIT was due to the experience of the programmers in OO technology and their knowledge of the program domain. The DIT study confirmed that reuse wasn't happening as often as it should, which is what the tool proposed would help show early on in the life cycle.