We had talked about doing a readings call about software engineering and science, and I wanted to suggest a few readings. I don’t *think* that we were going to to it this week – since we need to talk about metadata friction with the group, but we could do it this week if it made sense.

First, I wanted to recommend Carver’s report of the Software Engineering for Computational Science and Engineering workshop, which is a good overview of thought about the topic within the software engineering community.

Easterbrook and Johns studied software engineering of climate science, concluding more optimistically that smart people (e.g., the scientists) will figure out good engineering practices. Steve Easterbrook, at the University of Toronto, is on my committee.

Finally, Segal has an ethnographic study which concludes that more lightweight process would be helpful for scientists. She details some tensions between scientists and technologists. (She also has a recent CSCW article of interest about this.)

If short on time, I would suggest reading in the order above. I look forward to hearing our discussion of this, and of course am happy to lead it.

Saw this interesting article about a coding bug at Scripps with significant impacts on the field – possibly setting the science back by several years. I think it’s interesting because scientists often don’t have great software engineering practices, such as systematically testing their code. And code is not published for review either. I’m not sure that publishing the code necessarily would have revealed the problem, but it might have! (or could it have made the problem worse, because everyone would borrow the problematic code rather than writing their own that might catch the problem?)

http://genomebiology.com/2007/8/2/103