Automatic extraction of abstract-object-state machines based on branch coverage

Download: PDF.

“Automatic extraction of abstract-object-state machines based on branch coverage” by Hai Yuan and Tao Xie. In 1st International Workshop on Reverse Engineering To Requirements (RETR 2005), (Pittsburgh, PA, USA), Nov. 2005, pp. 5-11.

Abstract

Some requirement modelling languages such as UML's statechart diagrams allow developers to specify requirements of state-transition behavior in a visual way. These requirement specifications are useful in many ways, including helping program understanding and specification-based testing. However, there are a large number of legacy systems that are not equipped with these requirement specifications. This paper proposes a new approach, called Brastra, for extracting object state machines (OSM) from unit-test executions. An OSM describes how a method call transits an object from one state to another. When the state of an object is represented with concrete-state information (the values of fields transitively reachable from the object), the extracted OSMs are simply too complex to be useful. Our Brastra approach abstracts an object's concrete state to an abstract state based on the branch coverage information exercised by methods invoked on the object. We have prototyped our Brastra approach and shown the utility of the approach with an illustrating example. Our initial experience shows that Brastra can extract compact OSMs that provide useful information for understanding state-transition behavior.

Download: PDF.

BibTeX entry:

@inproceedings{YuanX2005,
   author = {Hai Yuan and Tao Xie},
   title = {Automatic extraction of abstract-object-state machines based
	on branch coverage},
   booktitle = {1st International Workshop on Reverse Engineering To
	Requirements (RETR 2005)},
   pages = {5--11},
   address = {Pittsburgh, PA, USA},
   month = nov,
   year = {2005}
}

Back to Publications whose methodology uses invariant detection.