Comments on program specification and testing

  • Authors:
  • Roger House

  • Affiliations:
  • 11450 Bodega Highway, Sebastopol, Calif.

  • Venue:
  • Communications of the ACM
  • Year:
  • 1980

Quantified Score

Hi-index 48.22

Visualization

Abstract

The data processing discipline is in a deplorable state, says Roger House. The author reviews an earlier study, in which programmers were largely unsuccessful in finding known errors in a test program. House shows that the fault lies not with the subjects but with the generally poor procedures underlying the writing of much contemporary software, including the test program in question. He analyzes the example, offers several specific critiques, and enumerates good software-writing practices.