Something in the “how could they do this” category

January 1, 2007

I read this a few minutes ago, and I am still shaking my head. I think the first thing I learned as a computational scientist was to develop a well known, familiar test set, something to test any program I used or wrote with. It is almost essential to have a familiar baseline, one that you know well. That’s the bare minimum. The error with the MsbA structure is, as Kaspar Locher puts it, a monumenental blunder. It’s sloppy science, and unfortunately, and maybe I am being stereotypical here, I feel too many grad students and post-docs are too eager to get results that they cut corners. Software-related shortcuts are probably top of the list. I hope this is a lesson to those who choose to treat software as something different or magical. It is an experimental tool, just like an instrument. I can’t do this issue the kind of justice that Sandra does, so I won’t try. Other than self-discipline, I wonder what might prevent something like this from happening more often.

Update: The Omics World (Hari knows a lot more about crystallography than I do) has a post that is worth reading