A Technique for Hiding Proprietary Details While Providing Sufficient Information for Researchers or Do You Recognize This Well-Known Algorithm?
Files
TR Number
Date
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
A major problem facing software engineering researchers is the difficulty of performing validation experiments requiring "real-world" data. This data may be composed of the requirements document, design descriptions, source code,test plans, and/or error histories. Most major software organizations object to giving this usually proprietary data to researchers. This objection is understandable since these organizations do not desire their proprietary source code or error history to be made public. This paper describes a technique to translate source code into a meta-language which will maintain the proprietary nature of the original source code by hiding the algorithms and data structures employed in the source code. This technique, however, preserves sufficient data for validation experiments to be performed by some software engineering researchers, particularly those in the software quality metrics community.