TL;DR :: Roman numerals were the challenge at hand. The method of the example is common to constraint searches beyond the example. Constraint searches exist in drug research and many other fields.
Roman numerals were chosen for a few really simple reasons. That was the golf challenge at hand. It was likely the golf challenge at hand because it's so simple and widely understood. Finding a random function which meets the demands is exactly the point. The demands, in this example are related to Roman numerals.
Part of computational pharmacology is the practice of using combinatorics to find substances that happen to achieve the desired molecular shape or the desired effect in simulations. Another is a combinatoric search for combinations of drugs that work together without negative interactions. Other parts of the field include algorithms to model relationships, data mining, learning algorithms, and network analysis.
Pick any field which has a large search space for just the right combination of properties in an as yet undiscovered item. Write a program which tries and makes a preliminary fitness determination for each possibility. Have that program spit out a short list of candidates for further investment of testing and development. That's the type of program the series of articles is about.
In this specific case, the fitness is a maximum length, a handful of inputs, and a handful of outputs that map correctly to those inputs. That it's Roman numerals, proteins, enzymes, metallic alloys for the skin of a jet, the ideal plant food for Granny Smith apple trees, or a new nanostructure for the anode of your battery design is really of little consequence to the method. The point of an example is that it is a concrete thing that is completed and shown rather than an abstract idea. Examples are often made of a simple case rather than the most complex case possible.
| 