To do better science we need a better model of what science is. Hypothesis→Experiment→Knowledge is too linear and simplistic. In our work as scientists and coaches of scientists we have observed that the real journey of science looks like a random walk between these seven nodes:
Choose: decide what to work on; taste helps.
Observe: see something in the world, be curious about it.
Describe: formalize what you see so that someone else can know it when they see it
Distill: reduce your description to only what matters for the thing you’re interested in
Predict: Know what will happen before it happens
Control: Make specific things happen when you choose
Extend: Make new things happen when you choose
A single person or team might move from node to node to complete a project or members of a connected community might specialize to cover all nodes. The order of node traversal can be surprising. Animal breeders like Robert Bakewell could Control heredity better than they could Describe it and Describe it better than they could Distill it. We hypothesize that being an effective scientist requires skill at some or all of the nodes AND knowing when and how to transition work between nodes.
Readers with certain backgrounds might usefully think of this as a state machine with 7 states and transitions allowed between all states. We also hypothesize that this is the right framework for building an automated scientist.
Important topic - I'm working on addressing a similar problem space. Definitely looking forward to reading through your posts to see where you've gone with this arena.