Joseph,
I totally agree that real world genetics are a schmear of too many genes, gene complexes, interacting regulators and who-knows-what to be modeled accurately by ANY simple theory.
Most of the real world is like that: hugely more complex than anything you can hope to do in a laboratory.
My take is that "Mendellian genetics" are from the typical perspective of "Science".
Simplify what you're talking about SO much that you can talk about and experiment with something so specific that the results make sense, and you can write a paragraph in a text book that makes sense. And most of all: simple enough that you can construct a controlled experiment that is so repeatable that you can MEASURE something. It's not Science unless you can produce a table of repeatable numbers.
And they care a lot whether it really is like what's going on, under the hood. "Truth", or at least accuracy about some specific aspect of something.
If it is relevant and inclusive enough to useful in the real world, that's an unexpected bonus.
You and Darwin have more of an engineer's perspective:
"How do things work in the real world?"
"What can I do, and how can I think about it, to get useful things done?"
"I don't need to pretend to UNDERSTAND things, I just need to DO things."
"Workable approximations are fine."
"I'm not hunting for provable Cosmic Truth, I just want it to work."