Games Programmer, Anime fan, General nerd.

 

Super awkward but trying to improve~ n_n

 

(Have been advised to add: All views my own)


  • Proceeds to be dogmatic about being anti-that organizational method.

I understand wanting to break way from that one man's works in software, he's a tool and very full of himself. That said, organizing code files into folders by generic categories (interface, data types, components) instead of by feature makes code discovery and reuse that much easier in a C or C++ context - totally independent of his writings on the subject.

I want to be like "I want these data types for the DisplayManager" - then type #include "Data/DisplayManagerTypes.h" and be done with it! Not to have to hunt down the feature-specific folder, which may be less obvious than that example.

In better languages with more sensible module definitions and symbol auto-discovery tools, sure you can break away from that. But in C and C++, unfortunately, the convention serves its purpose. (Until C++20 modules, which isn't well-supported yet. And even then, it's only base-level support without standard library support until 23, which is a ways away from support. Not to mention the lack of tooling right now)


You must log in to comment.