Today I set out to enhance the social network visualization I run to interpret digital ham radio traffic I monitor.
This had been implemented as a shell script that did its heavy lifting with perl one liners. I thought this would be elaboration of the existing script but I could not thing through the required perl code while in the text editor.
I turned to wiki and wrote a page as if I were establishing smalltalk objects. The result is a description that I am confident will work when I elaborate it as code.
Eight hours of unintended propagation leaves an epidemic in progress.
(Bad news. It didn't work as intended. I was confused when I turned to "paper" design and there convinced myself that I had understood propagation. Now I have an epidemic of tags that has no meaningful connection with the contact that initiated it.)
We identify some stations as tagged and once tagged all stations they confirm will hence be marked as tagged too.
This side trip to conceptually establish a design is an example of me responding to unexpected complexity in my weekend tinkering.
(Good news. After three evenings of watching the "midas effect" misinterpret the available data I realized my temporary storage between runs was unnecessary. The correct code turned out even simpler once I applied the logic to all available data using a newly coded function.)