Resolving uncertainties during trace analysis. (bibtex)
by Alexander Egyed
Abstract:
Software models provide independent perspectives onto software systems. Ideally, all models should use the same model element to describe the same part of a system. Practically, models elements are not shared because of syntactic and semantic differences among modeling notations. Trace dependencies explicitly maintain the commonalities among the distinct model elements. Generating and maintaining trace dependencies is difficult, costly, and highly error-prone. Automated trace analysis techniques are scarce. This paper extends an existing, testing-based technique for generating and maintaining trace dependencies. It is based on the commonality principle: if two model elements of different perspectives are the same then they must have the same source code. The existing approach associates test scenarios with model elements, tests them, and observes what lines of code are being executed. Model elements are considered the same/similar if their testing uses the same/overlapping lines of code. This paper extends the existing approach (and tool) by giving the user a richer, more powerful, yet precise language on how to relate model elements, test scenarios, and source code (the input). This allows some forms of uncertainties to exist in input data without sacrificing reliability. The extended approach also identifies "shared code". Shared code works against the commonality principle in that model elements do not relate if they overlap solely on their use of generic source code (e.g., queue). As a prerequisite, our approach requires an executable and observable software system and test scenarios.
Reference:
Resolving uncertainties during trace analysis. (Alexander Egyed), In Proceedings of the12th Symposium on Foundations of Software Engineering (FSE2004), Irvine, CA (Richard N. Taylor, Matthew B. Dwyer, eds.), ACM, 2004.
Bibtex Entry:
@Conference{DBLP:conf/sigsoft/Egyed04,
  author    = {Alexander Egyed},
  title     = {Resolving uncertainties during trace analysis.},
  booktitle = {Proceedings of the12th Symposium on Foundations of Software Engineering (FSE2004), Irvine, CA},
  year      = {2004},
  editor    = {Richard N. Taylor and Matthew B. Dwyer},
  pages     = {3-12},
  publisher = {ACM},
  abstract  = {Software models provide independent perspectives onto software systems.
	Ideally, all models should use the same model element to describe
	the same part of a system. Practically, models elements are not shared
	because of syntactic and semantic differences among modeling notations.
	Trace dependencies explicitly maintain the commonalities among the
	distinct model elements. Generating and maintaining trace dependencies
	is difficult, costly, and highly error-prone. Automated trace analysis
	techniques are scarce. This paper extends an existing, testing-based
	technique for generating and maintaining trace dependencies. It is
	based on the commonality principle: if two model elements of different
	perspectives are the same then they must have the same source code.
	The existing approach associates test scenarios with model elements,
	tests them, and observes what lines of code are being executed. Model
	elements are considered the same/similar if their testing uses the
	same/overlapping lines of code. This paper extends the existing approach
	(and tool) by giving the user a richer, more powerful, yet precise
	language on how to relate model elements, test scenarios, and source
	code (the input). This allows some forms of uncertainties to exist
	in input data without sacrificing reliability. The extended approach
	also identifies "shared code". Shared code works against the commonality
	principle in that model elements do not relate if they overlap solely
	on their use of generic source code (e.g., queue). As a prerequisite,
	our approach requires an executable and observable software system
	and test scenarios.},
  doi       = {10.1145/1029894.1029899},
  file      = {:Conferences\\FSE 2004 - Resolving Uncertainties during Trace Analysis\\Resolving Uncertainties during Trace Analysis-preprint.pdf:PDF},
  keywords  = {},
}
Powered by bibtexbrowser