The article is more concerned with "open development" than with "scientific software". Those interested in more specifically science-oriented advice might like the following, which is referenced in the OP: http://arxiv.org/abs/1210.0530. For example, this one mentions version control and testing against data.