Step 2 — do the rest of the fucking analysis
Most research is difficult. Not only because we are asking difficult questions, but because asking these questions require tools of increasing sophistication. A lot can be challenging — identifying which tools to use, understanding what they do, and learning how to use them. I come from an experimental biology background, and this applies equally to both experimental and computational questions.
So when starting a new research project, one can feel like one is trying to draw an owl using the above tutorial. This is because we tend to learn about methods by reading papers, and the Methods section of any given paper is often, to put it mildly, terse. To pursue the How to draw an owl analogy, a Methods section could read
We draw the owl on 60.2 gsm white paper of the A4 dimension (210mm by 297mm), using 3H and 6B graphite pencils (Derwent, Cumbria, UK). We did so by looking at owls, and drawing what we saw on paper. This protocol yielded one drawn owl.
After reading this, I would be able to understand how Figure 1 — schematic depiction of the owl was produced, but it would be extremely difficult for me to draw an owl for myself. Methods section, I find, are poor as ensuring replicability of the results.
Not that they should be. I would be perfectly comfortable with methods section that describe what was done if we had an additional resource: how-to papers. These papers would be an incredible resource for the community, because they would present the technical details of how to actually do a particular thing.
For example, reading a mathematical introduction to Gibbs sampling is instructive. But it is a different exercise from reading through an explanation of how to implement a Gibbs sampler. Both have a lot of value, because mastering the methods one uses requires to understand how they work, but also how to perform them.
Sadly, the likelihood of seeing these papers published depends on a key factor: would they be cited? Probably not (or not as much as other research papers). But the success of PLOS Comp Biol 10 simple rules collection demonstrates that this is a resource that the community values.
