Planning and estimation in scientific
software projects
For a software project viewed as a scientific exploration, it becomes apparent that what matters at a given time is the set of proven hypotheses, not whether the exploration has proceeded according to some original plan. The purpose of a project becomes not meeting some set goal, but exploring a field (the problem context), trying to answer questions, solving problems, and so on.
No one would even think of asking how long a specific scientific exploration would take to complete. Rather, it’s about estimating what value a solution to some problem would have, and funding the research project based on a speculation that the questions can be answered. Perhaps not all questions will be answered, but some of them probably will, and maybe new questions will be discovered and answered to the benefit of the funders.