Publikationen aus dem Jahr 1997
Jürgen Ebert, Claus Lewerentz (Hrsg.)
Proceedings of the 8th Conference on Software Engineering Environments (SEE'97)
Cottbus, April 1997, 174 pp, IEEE Computer Society Press, Los Alamitos, California, 1997, ISBN 0-8186-8019-9
HA 97-016 / 1997[01]
Claus Lewerentz, Heinrich Rust
Zur Bedeutung von Spezifikationen in verschiedenen Teilaufgaben der Entwicklung kundenspezifischer Software
Technischer Bericht Nr. I-21/1997, Institut für Informatik, Brandenburgische Technische Universität Cottbus, 1997
HA 97-017 / 1997[02]
Claus Lewerentz, Frank Simon
Integrating of an object-oriented metrics tool into SNiFF+
Technical Report No. I-22/1997, Computer Science Department, Brandenburg Technical University at Cottbus, 1997
HA 97-018 / 1997[03]
Integration of an object-oriented metrics tool into SNiFF+
Claus Lewerentz, Frank Simon
Software and Systems Engineering Group
BTU Cottbus
In the project Crocodile tools for supporting software quality measurement are developed. In particular, object oriented metrics are used to define product quality models for object oriented programs and frameworks. Our current tool implementation is based on and integrated into the SNiFF+ programming environment.
Crocodile’s approach is to use the parsed data from SNiFF’s symbol table for extracting a basic set of structural program properties. This symbol table always contains the up-to-date data like classes, inheritances and references of the currently open software project. Because SNiFF+ does the mapping from the programming language into the symbol table, our tool itself is language independent. The whole measuring process is done within a SNiFF+ session: It’s started by a new SNiFF+ -command, it’s configured through it and the feedback of the measured values is given back to SNiFF’s user interface. Crocodile offers its own query-language for defining specific measures and quality models. A special mechanism to filter the amount of calculated measurement values to the critical ones helps the engineer to effectively do and understand the measuring. This data reduction is achieved by defining subsets of classes to be measured - considering usage and inheritance context - and by defining critical ranges for the self-defined object oriented measures.
Thus, Crocodile provides an easy way to create a specialized measurement process considering the user’s specific goals in engineering or re-engineering. Furthermore, measurement activities are well integrated into the particular software development model and tool support through the usage of SNiFF+.
Abstract 1997[03]