10:30 AM-12:30 PM
Monticello
The most important quality of any software is that it produces correct results for all valid inputs. Since we are currently unable to prove the majority of scientific software correct, we are forced to resort to testing as a means of increasing our confidence in its performance. There is strong evidence that good testing strategies can dramatically reduce the number of reported errors in released software, and recent work has revealed a strong correlation between a small number of software metrics and the occurrence of errors in Fortran software. A number of software tools are available to assist with testing, but appear to be rarely used by the scientific community.
Organizers: Tim R. Hopkins