[ 15/6/20 ]
Shaun Mootoo has most of it.
Most problems have solutions that can be approximated at least well enough to survive.
David Snowden does a great job of simplifying infinite complexity down to four classes – simple, complicated, complex and chaotic.
Simple and (to a reasonable degree) complicated problems can be solved using classical problem solving methods.
Solving anything in the class of complex systems demands an iterative approach, because whatever you do to the system, it will respond in unpredictable ways that you will need to adapt to. So plans have to change and adapt constantly.
The only thing one can do when in a chaotic system is attempt to get out of it. A random path is often the shortest.
As long as one is prepared to accept both of those last two classes of solutions as required, then anything is soluble – in a sense.