Reply To: Questions regarding DPF.

[Augustine]

Scientists do not work with consensus. Scientists do research, create a hypothesis, test to verify their hypothesis, work to eliminate alternate explanations to their hypothesis, and then they publish their data and methodology so that others can look at what they have done and reproduce it. This last step is quite important as the goal of science is to find and disseminate new knowledge to other people.

Looking at the CRU code snippets I can say in my professional opinion that they have the quality of a second semester undergraduate computer programmer. If I wrote code that was that sloppy I would be fired and probably my boss would perform a post mortem on our hiring practices. The code is a hodge podge of various subroutines with next to no organization and irrelevant documentation. The data sets that the code operate on seem to be in a similar poor condition. What this means is that if asked, the researchers at CRU probably could not reproduce their findings as the code is a mess and they don’t seem to know which data set they should run against.

I think that this debacle is a real black eye for science. I also think that this shows the real need for transparency when researchers are working on problems of this scope (problems of the “lets rearrange our economies” scope). Researchers should have to publicly explain their methodologies (publicly, not through peer review), researchers should use commercial software libraries or open libraries (things that are known to work), and researchers should make their own code publicly available. Real scientists and researchers should not fear skeptics- one of the advantages of science is that it survives in the face of skeptics.