I do a lot of fairly complex data processing that benefits from any abstraction I can get.
E.g. “Take these 50,000 data points, identify and apply the best set of filters for each sections, then produce the best curve fit for each section.” It’s massively complex and (hilariously) produces multiple equally massive data sets while working – acceleration and velocity plots, for example.
My workflow is :
Step one: write down a bunch of nonsense, massive block comments of old code in case the new code is worse, and cludges in one massive file.
Step two: rewrite the whole thing: breaking it down into classes so it’s readable and maintainable …
Step 3: come back later and wonder where the hell anything is. Go to step 1
Step 4: realize the shit was simpler and better without the classes.
I do a lot of fairly complex data processing that benefits from any abstraction I can get.
E.g. “Take these 50,000 data points, identify and apply the best set of filters for each sections, then produce the best curve fit for each section.” It’s massively complex and (hilariously) produces multiple equally massive data sets while working – acceleration and velocity plots, for example.