I don't see it this way—that learning to code is in some way less elegant and less conceptual than learning computer science. That "it's not science" anymore. R

I don't see it this way—that learning to code is in some way less elegant and less conceptual than learning computer science. That "it's not science" anymore. Rather, your idea of "learning to code" is just operating at a higher level of abstraction from your idea of "learning computer science." When you start with primitives like array.sort(), you're implicitly bringing the organization of a program or system as a whole to the forefront, while abstracting way the mechanics of that sorting process. This is a good place to start understanding individual computer programs and systems. When you start with primitives at the logical and mathematical level (i.e. theory), you're bringing the organization of lower-level concepts like data and control structures to the forefront. This is a good place to start when understanding the components that go into higher-level primitives like "sorting." And you can go up and down in terms of the level of abstraction you want to view computer systems. Go higher, and you'll think more about the way multiple computer systems integrate (this is precisely what frameworks, APIs, and SDKs tackle). Go lower, and you'll think more about the way hardware interacts with machines (which, ironically, brings us back to electrical "engineering"—not science!). To me, it's all just layers of abstraction, and the layer you choose should just be the one you find natural, interesting, and maybe useful. It's not a strict hierarchy; there are branches and perhaps cycles. But our understanding of the systems that run industry today, are only possible upon a complex system of abstractions that goes all the way down to the physics of the universe. There are patterns and antipatterns to understand when "learning to code" (at the systems level), just like there are patterns and antipatterns to understand when "learning computer science" (at the theoretical level), but they just operate at different levels. The term "computer science" implies that it is science, and not engineering. I think that the main reason theory is seen more as an "academic" pursuit as opposed to more abstracted understanding of computer systems is precisely because it has been so thoroughly abstracted, that trying to operate on a purely theoretical basis is just too inefficient to be particularly "useful" in an engineering setting, by some metric. It's not because of innate beauty or complexity; it's just where that arbitrary line is drawn.