FAST: a framework for automating statistics-based testing

  • Authors:

  • Huey-Der Chu;John E. Dobson;I.-Chiang Liu

  • Affiliations:

  • Centre for Software Reliability, Department of Computing Science, University of Newcastle upon Tyne Newcastle upon Tyne UK;Centre for Software Reliability, Department of Computing Science, University of Newcastle upon Tyne Newcastle upon Tyne UK;Department of Information Management, National Defense Management College Chungho 23500 Taiwan

  • Venue:
  • Software Quality Control
  • Year:
  • 1997

Quantified Score

Hi-index 0.00

Visualization

Abstract

To achieve software quality, testing is an essential component in all software development. It involves the execution of a deterministic software system with test data and a comparison of the results with the expected output, which must satisfy the users‘ requirements. This accounts for over 25% of the cost of a software development. Therefore, automation has considerable potential. The quality programming introduced by Cho can automatically generate data for testing, based on a so-called ‘SIAD tree‘ which is used to represent the hierarchical and ‘network‘ relation between input elements and also incorporates rules into the tree for using the inputs. However, it lacks a clear framework which would show how automated testing can be achieved. To address this problem, we present a Framework for Automating Statistics-based Testing (FAST), which is an extension of the testing concept in quality programming to achieve automated testing. In FAST, we propose a SOAD tree, which is similar to the structure of the SIAD tree, to describe the syntactic structure of the product unit and its defectiveness. Based on this tool, the inspection of test results can be automatically achieved by lexical and syntax analysis. The implementation of automated software testing or Command File Interpreter (CFI) software which incorporates the framework is also described.