It's not huge. I probably know all the more important books. Not that many. In chess I stopped counting when I hit 800 but I've probably got around 1500 or more books. I can take a novice and raise him to decent club standard within a few years with a structured training program. I've no idea how this would be done for
C++ coding.
It's true that it would be difficult to find much more than a hundred meaningful books on
C++. But as you noted,
C++ is smaller than chess (which is sort of odd since
C++ is an infinite language).
C++ is just a way to express algorithms. Or put another way,
C++ is just a tool for computer science. I will say that computer science is far bigger than chess and the literature is far larger. Computer science embodies a big chunk human thought: information theory, Kolmogorov complexity, computer graphics, fractal mathematics, machine learning, statistics and much more. In many areas what computer science does not encompass, it enables. Chess is old, but modern computation is a tool for the human mind. And, in fact, computer science has triumphed over chess. You would know better than I, but as I recall the top chess programs can beat the best human players.
As to a structured training program in
C++: I think that I'd have someone write an optimizing compiler in
C++ (for another language like Java). Compilers have huge sets of data structure and are large complex programs. After you've done a compiler in
C++ you would have a fair amount of mastery. Of course mastery fades. I use
C++ episodically (I'm back to using it now for some physics simulations). But I've been using Java for years so my
C++ gets rusty.
I pretty much suck at chess, so between
C++ and chess I'm better at
C++.