Date of Graduation


Document Type


Degree Type



Statler College of Engineering and Mineral Resources


Lane Department of Computer Science and Electrical Engineering

Committee Chair

Bojan Cukic.


This work describes the production of an application Error Propagation Metrics from XMI which can extract process and display software design metrics from XMI files. The tool archives these design metrics in a standard XML format defined by a metric document type definition.;XMI is a flavour of XML allowing the description of UML models. As such, the XMI representation of a software design will include information from which a variety of software design metrics can be extracted. These metrics are potentially useful in improving the software design process, either throughout the early stages of design if a suitable XMI-enabled modelling tool is deployed, or to enable the comparison of completed software projects, by extracting design metrics from UML models reverse engineered from the implemented source code.;The tool is able to derive the error propagation of metrics from test XMI files created from UML sequence and state diagrams and from reverse engineered Java source code. However, variation was observed between the XMI representations generated by different software design tools, limiting the ability of the tool to process XMI from all sources. Furthermore, it was noted that subtle differences between UML design representations might have a marked effect on the quality of metrics derived.;In conclusion in order to validate the usefulness of these metrics that can be extracted from XMI files it would be useful to follow well-documented design projects throughout the total design and implementation process. Alternatively, the tool might be used to compare metrics from well-matched design implementations. In either case design metrics will only be of true value to software engineers if they can be associated empirically with a validated measure of system quality.