Reproducibility in Scientific Computing: Episode 5 of inSCIght, the Scientific Computing Podcast

If the results can't be reproduced, it's not science: http://inscight.org/2011/03/23/episode_5/