Structural comparison of source code between multiple programming languages

Software developers are often faced with the task of comparing two or more versions of software. Typical usages of software comparison utilities include: a code-review prior to checkin, tracking down a recently introduced regression, and searching for code-clones in the source code (for future ref...

Full description

Bibliographic Details
Main Author: Biehn, Rolf
Language:English
Published: University of British Columbia 2014
Online Access:http://hdl.handle.net/2429/46547
id ndltd-LACETR-oai-collectionscanada.gc.ca-BVAU.2429-46547
record_format oai_dc
spelling ndltd-LACETR-oai-collectionscanada.gc.ca-BVAU.2429-465472014-04-24T03:42:57Z Structural comparison of source code between multiple programming languages Biehn, Rolf Software developers are often faced with the task of comparing two or more versions of software. Typical usages of software comparison utilities include: a code-review prior to checkin, tracking down a recently introduced regression, and searching for code-clones in the source code (for future refactoring). However, most traditional source code comparison tools typically use simple text-to-text comparison (with some simple rule-based comparisons for comments), which has the drawback of showing superfluous differences during comparison. Many projects, for a variety of business reasons, ship products and software development kits (SDKs) using multiple programing languages. It is desirable to compare amongst languages in order to detect potential errors and understand the meaningful differences between the two codebases. In some cases, fixes may be implemented in one language, but not in the other. In this paper, we create a tool called the Software Difference Analyzer Tool (SDAT), a tool capable of comparing Java and CSharp code, to address some of the unique problems associated with cross-language comparison. Automated testing demonstrated SDAT reduces the number of reported differences by up to 40%. User testing has shown a 37% increase in speed and 28% increase in accuracy. 2014-04-22T16:38:06Z 2014-04-22T16:38:06Z 2014 2014-04-22 2014-05 Electronic Thesis or Dissertation http://hdl.handle.net/2429/46547 eng http://creativecommons.org/licenses/by-nc-nd/2.5/ca/ Attribution-NonCommercial-NoDerivs 2.5 Canada University of British Columbia
collection NDLTD
language English
sources NDLTD
description Software developers are often faced with the task of comparing two or more versions of software. Typical usages of software comparison utilities include: a code-review prior to checkin, tracking down a recently introduced regression, and searching for code-clones in the source code (for future refactoring). However, most traditional source code comparison tools typically use simple text-to-text comparison (with some simple rule-based comparisons for comments), which has the drawback of showing superfluous differences during comparison. Many projects, for a variety of business reasons, ship products and software development kits (SDKs) using multiple programing languages. It is desirable to compare amongst languages in order to detect potential errors and understand the meaningful differences between the two codebases. In some cases, fixes may be implemented in one language, but not in the other. In this paper, we create a tool called the Software Difference Analyzer Tool (SDAT), a tool capable of comparing Java and CSharp code, to address some of the unique problems associated with cross-language comparison. Automated testing demonstrated SDAT reduces the number of reported differences by up to 40%. User testing has shown a 37% increase in speed and 28% increase in accuracy.
author Biehn, Rolf
spellingShingle Biehn, Rolf
Structural comparison of source code between multiple programming languages
author_facet Biehn, Rolf
author_sort Biehn, Rolf
title Structural comparison of source code between multiple programming languages
title_short Structural comparison of source code between multiple programming languages
title_full Structural comparison of source code between multiple programming languages
title_fullStr Structural comparison of source code between multiple programming languages
title_full_unstemmed Structural comparison of source code between multiple programming languages
title_sort structural comparison of source code between multiple programming languages
publisher University of British Columbia
publishDate 2014
url http://hdl.handle.net/2429/46547
work_keys_str_mv AT biehnrolf structuralcomparisonofsourcecodebetweenmultipleprogramminglanguages
_version_ 1716665983708430336