I recently posted about three phases of prototyping: (1) the quick rough sketch you implement in an afternoon, (2) the more completely featured rough sketch that takes a few days, then (3) the robust polished prototype that you may work on for several weeks.
In response, J. Peterson sensibly commented that there is a fourth phase — creating a commercially robust version that can be monetized (followed by a fifth: making money).
Now that I have a working version of my current third phase prototype, I appreciate anew the vast gulf between phases three and four. The decisions I made to create this third phase prototype were very different from the decisions I would have made to create a commercial product.
Because my interest is research, everything I create is lab equipment. All of my effort goes into making things as flexible and easy to modify as possible. Pretty much no effort goes into making something that could “survive in the wild”, in J. Peterson’s words.
Asking the question “What should this be like?” (which is the fundamental question I ask) is very different from asking the question “How can I make something that will be used by millions of people?”
My responsibility as a researcher is to answer the first question. If I am successful, then any commercial developer worth his/her salt will first carefully examine my results, extract useful principles, then throw out all of my code, and build something to commercial specs, from the ground up.