Survey: Silent Errors in Scientific Code

This post originally appeared on the Software Carpentry website.

Posted on behalf of Daniel Hook and Diane Kelly:

We are members of a software research group from Queen's University who are investigating ideas and tools to assist with the development of scientific software. We are starting a project focused on finding silent or hidden errors in scientific code. (Silent errors are errors that don't result in a crash, an error message or an other obvious indicator of a problem.) To create a catalogue of common silent errors, we would like to hear debugging "war stories" from computational scientists. Using these stories we hope to provide improved code testing techniques specifically for scientists.

To conduct our study, we are looking for scientific software debugging stories: just a few lines explaining what the problem was and how you managed to solve it. You can contribute to our study by sending us a story and/or by passing this email along to colleagues who might be able to help with a story.

Please navigate to http://www.cs.queensu.ca/~hook/err_intro.html for more details if you are interested in contributing an error story to this study.

Thank you for your interest.

Sincerely,
Daniel Hook and Diane Kelly