For hobbyists that's enough, for engineers often okay (I find myself in that situation) but for scientists "good enough" means nothing.

Optical metrology relies on accurate equations how a physical object maps to the image plane so in that case analytical solutions are necessary for subpixel accuracy.

I'm worried about how often kids these days discount precise mathematical models for all use cases. Sure, you get there most of the time but ignore foundational math and physics at your own peril.