Just finished reading Programming as Theory Building by Peter Naur. It's a good read, putting into some high falutin' words thoughts I've also had as a programmer.

A point struck me as particularly interesting though. An argument that program modification is, long term, more costly than a rewrite to create a program that encompasses the real world changes that demanded modification. At first, I thought, of course a programmer would argue a rewrite is better, that's just cliche. But I got to thinking of a shop I was employed at briefly. Part of the onboarding was an explanation that the cost of program maintenance is potentially exponential, especially if the foundation of the program isn't solid. If that's true, it follows at some point in a program's lifespan, a rewrite is inevitably going to be less costly.

I guess the cliche is a cliche for a reason.