Spatial Complexity Metrics: An Investigation of Utility

Gold, N E, Mohan, A M and Layzell, P J (2005) Spatial Complexity Metrics: An Investigation of Utility. IEEE Transactions on Software Engineering, 31 (3). pp. 203-212. ISSN 0098-5589

Full text not available from this repository.
Item Type: Article
Additional Information: Originality: Spatial complexity metrics (including ones developed by Douce/Layzell) claim benefits over conventional lines of code methods. This work evaluates this claim and adds substantially to the body of empirical evidence on the effectiveness of complexity metrics. Originality derives from the application of theoretical measures to a live, mature software system. Rigour: The work is based upon an empirical study employing a highly evolved, large commercial software system. Two hypotheses are tested using statistical techniques to establish the plausibility of spatial metrics and advantages of complexity metrics. Significance: Analysis shows spatial complexity metrics are plausible and provide behaviour consistent with established laws of software evolution, thus establishing their credibility. The work also shows that spatial complexity largely does not provide metrics which are superior to lines of code, with the exception of the Douce/Layzell metric which provides some greater benefit. The work is also significant in adding a credible case study to the body of empirical software research knowledge. Impact: Further work is being undertaken at King's College London on metrics for program comprehension.
Schools and Departments: School of Engineering and Informatics > Informatics
Depositing User: Paul Layzell
Date Deposited: 07 Feb 2012 15:52
Last Modified: 28 Mar 2012 14:03
URI: http://sro.sussex.ac.uk/id/eprint/36120
📧 Request an update